Tarka GTM / AI Market Research ReportSeptember 30, 2026
Salesforce Moved Its CRM Into Claude and Its Pricing Into Credits
Salesforce put its CRM inside Claude and launched new editions built around pooled Flex Credits in the same month OpenAI and Anthropic cut model prices by 20 to 50 percent. Seats became the container and credits the meter, and the vendors kept the margin between falling token costs and their own credit rates.
Summary
- Top 3: Salesforce put its CRM inside Claude and declared that people no longer have to come to Salesforce to get work done; Salesforce and HubSpot repriced around seats plus credits in the same weeks OpenAI and Anthropic cut model prices by 20 to 50 percent; and TypeSafe's Jev made scoring every account in a market cost less than a dollar.
- Market Trends: seats became the container and credits the meter, with vendors keeping the gap between falling token costs and fixed credit rates; the CRM moved into the agent and vendor-authored skill packs became the unit of distribution; context became the product every GTM vendor claims to sell; inbound became the first agent win with reproducible numbers while "AI SDR" lost its name; and the lead record and the plugin marketplace became the attack surface.
- New Tools: Gong's enrichment waterfall and Apollo Next on the last day of the month, Clay's Plugin and Skills Marketplace, Anthropic's plugin portal, HubSpot's Sales Plugin for Claude, AirOps AI Ads, Firecrawl's Alexandria, Jon Miller's Phave, Zapier's Next Gen Zaps, an open-source Jev GTM cookbook, Resend and Klaviyo going headless, n8n Agents, HubSpot's Fall Spotlight, Pipedrive Nova, Ami by AiSDR and Profound's Context Manager.
- Labs Watch: OpenAI and Anthropic cut prices inside one week while DeepSeek and xAI shipped cheaper tiers, OpenAI shelved a model over scope and authorization, Google released its strongest model only to vetted defenders, Meta started selling agents directly to small businesses, and xAI showed a per-account sales bot pattern anyone can copy.
- Funding: sixteen acquisitions led by Salesforce's ~$3.6B Fin close and its reported ~$2B Listen Labs deal, with HubSpot, ZoomInfo, ServiceNow and Superhuman each buying context; rounds topped by Wonderful's $550M, Profound at a $1.8B valuation and Clay at $7.1B; and a YC batch in which "GTM engineering" became a seed-stage category.
- From Seattle: the local GTM track switched to a peer roundtable, Outreach's AI credit use grew 12x, Amperity shipped its MCP server into Claude and ChatGPT, Madrona co-led SciFin's $44M seed, and Oracle, Microsoft and Qualtrics cut local jobs.
- Discussion: whether context is the cure for missing AI ROI or the new excuse for it, whether the AI SDR died or only its name, whether the SaaSpocalypse is over or only moved to point solutions and headcount, whether buyers' agents and AI-answer ads are a measurable channel yet, and whether forward-deployed agent services are a bridge or a trap.
- Further Reading: the investor case for commoditized intelligence, an operator running 113 agents, a fully specified post-conference ABM build, the founder of Jev on decision models, the case for replacing the CRM, the strongest counterweight to automation-first selling, a year of AI-search experiments, and two a16z essays that framed the month.
Top 3
-
Salesforce and Anthropic launched Claudeforce on August 26, and by Dreamforce the CRM's front door was someone else's assistant 1. It launched as "Salesforce in Claude," a plugin with 37 prebuilt sales skills that read live pipeline and take governed action from inside Claude, running on what Salesforce calls AIforce — "all your business data and workflows to any agent through MCP servers, APIs, and CLI tools." Anthropic shipped the beta on September 15 with account research, call prep, pipeline review and CRM updates under the seller's existing permissions 2, and at Dreamforce the same day Salesforce made the pattern its platform: AIforce is "a live interface layer," and "people don't have to come to Salesforce to get work done" 3. HubSpot followed with a free eight-skill Sales Plugin for Claude on September 25 4. August read the incumbents as the discount layer
[context — August 2026]. September's answer from the largest one was to stop competing for the interface and ship into the agent instead — which makes skill packs, not connectors, the thing a GTM engineer now customizes. -
Salesforce launched new Agentforce editions built around pooled Flex Credits on September 3, in the same month the labs cut model prices by 20 to 50 percent 5. Core is $195 per user per month with 500,000 Flex Credits, Advanced $395 with one million, Max $550 with 2.75 million. HubSpot said at its September 17 Analyst Day that it is moving to a hybrid of seats and credits, piloting it with new customers in Europe 6. Meanwhile OpenAI halved GPT-6 Sol and Luna API prices on September 22 7, Claude Opus 5.5 arrived the same day at a 20% lower list price and, Anthropic says, 40% less to run than Opus 5 8, and Ramp measured the effective price per million tokens down 41% since March 9. August argued value was moving into the layer that meters tokens
[context — August 2026]. The price half of that held; the capture half didn't. The seat vendors put the meter inside their own contracts, and the gap between falling input costs and fixed credit rates is now the margin they keep. -
TypeSafe AI launched Jev in early access on September 15, a model that returns typed decisions instead of text, at $0.042 per million input tokens with output free 10. Jev answers in 70 to 500 milliseconds with calibrated probabilities and schema-valid output, aimed squarely at "smart if-statements" — classify, route, score, branch. That is the shape of lead scoring, reply triage and ICP fit. Within a week a public cookbook had scored 16,711 LinkedIn connections against an ICP for $0.73 11, and two weeks later OpenAI announced a Decisions API in limited preview at DevDay that TechCrunch called "a Jev clone" 12. Scoring every account in a market used to be a budget decision. It is now a rounding error, which moves the hard problem in targeting from cost to judgment.
Market Trends
Seats became the container and credits became the meter
The August read was an inversion: the metering layer becomes the asset and the seat layer becomes the discount. September produced convergence instead. Salesforce's new editions bundle seats with pooled Flex Credits at every tier 5. HubSpot's Analyst Day described seats for access and credits to monetize agentic work, alongside a 2030 target of 30% non-GAAP operating margin 6, and its Starter refresh on September 29 moved the Data Agent, Prospecting Agent and Agent Builder into its entry plan, with both agents metered by HubSpot Credits 13. Workday told investors on August 27 that its AI products drove more than 25% of new ACV and that "all of it runs through flex credits" 14. GitLab introduced a single annual commitment covering seats and GitLab Credits 15. Asana said AI reached 25% of net new ARR and that its largest AI deal offset a customer's seat contraction 16. Even lemlist, crossing $60M ARR, said "hybrid pricing will win," with more than $2.5M of its ARR now from credits for enrichment, AI and signals 17.
The cost side moved the other way. OpenAI and Anthropic repriced inside one week: GPT-6 Sol to $2/$10 and Luna to $0.10/$0.50 7, Opus 5.5 to $4/$20 with cache reads at $0.20 8, Sonnet 5.5 at the same $2/$10 but with fewer tokens per task 18, and GPT-6.1 Sol at a fifth of Astra's price with cached input at $0.10 19. Ramp's spend data shows buyers already acting on it: frontier models' share of tokens fell from a 53% peak to 45% as companies imposed cheaper defaults 9. A vendor's credit rate doesn't fall when its input does. Clarify raised its overage price by half on September 10 20 — six days before Pipedrive gave meeting intelligence away free in every plan 21.
The cost also landed on the vendor's own P&L. Salesforce's deputy CFO told an investor conference that Claude token spend in R&D was part of why margin guidance didn't rise 22, and The Register reported Salesforce still hasn't worked out how to charge for outcomes, calling seats plus Flex Credits plus all-you-can-eat contracts an "anxiety-filled architecture" 23. a16z made the opposite case on August 27: an application that prices in tokens "imports the model provider's cost structure" and anchors its value to a unit whose cost keeps falling 24. For a GTM team the practical move is in the contract, not the model: know your credit burn per workflow before renewal, because the vendor's unit price is the one number in your AI stack that isn't falling.
The CRM moved into the agent, and the skill pack became the unit of distribution
Claudeforce made the pattern explicit 2, but the month's evidence is how many vendors shipped the same shape. Clay made its Plugin — CLI plus API — generally available for Claude Code and Codex on September 15, so agents can "build workflows, run enrichments, pull data from Audiences and 200+ data providers, and trigger actions in Clay — all headlessly" 25. Two weeks later it opened a Skills Marketplace where GTM engineers publish free, installable playbooks that draw on Clay credits when they pull data through Clay 26. HubSpot's MCP server gained write access to deals, campaigns and emails with every write attributed in the audit log 27; Outreach added 11 tools to its MCP server 28; Klaviyo made more than 260 MCP tools callable so "an agent can read, write, and go live … without anyone opening Klaviyo at all" 29; and Resend set out to make every dashboard action reachable "through the MCP, CLI, or any SDK" 30. Anthropic turned the pattern into distribution with the Claude Marketplace on September 23, letting buyers spend existing Anthropic commitments on partner tools in limited preview 31, and a plugin submission portal two days later 32.
a16z laid out the incumbents' bull case on September 3: if a system of record plus a general agent is "enough to get the work done," incumbents have "an incentive to control and charge for access to their data" — a case it called only "partly correct" 33. The challengers answered with a different noun for the record itself. Lightfield raised a $47M round led by a16z to rebuild the CRM as "a business world model — a record that updates itself from every call, email, and meeting, structured for agents to reason from" 34; Zero raised $10.3M to build "what comes after CRM" 35; and Monaco bought Overlayy for its sales-conversation context 36. Microsoft went after the lock-in directly with Dynamics 365 Activate, an AI-driven migration tool for teams "locked into systems customized over years" 37. The GTM engineer's leverage has moved from building connectors to choosing, customizing and governing vendor skills — and those skills run against your credits and your permissions, which makes them a procurement decision as much as a build decision.
Context became the product every GTM vendor claims to sell
The month's most useful vendor data point came from HubSpot. Across its customer base, AI working from good CRM context lifted MQLs 264% and closed-won deals 197%; the same models working from bad context did worse than no AI at all — MQLs down 28%, closed-won down 27%, meetings down 49% 38. At UNBOUND, its renamed flagship event, which HubSpot says drew more than 14,000 registrations and 1.6 million livestream viewers 39, HubSpot rebuilt its Fall Spotlight around "Growth Context," with a self-updating Smart CRM and a Context Home score for how complete a customer's data foundation is 40. Salesforce's new Enterprise AI Harness leads with "Trusted Context" among six capabilities 41, and it shipped Koa, a CRM reasoning model post-trained on a synthetic dataset modeled on "nearly three decades of CRM deployments," which it says matches or exceeds leading models on its own CRM benchmark with three times fewer errors 42.
The rest of the stack said the same word. ZoomInfo's GTM.AI calls itself "the Context Graph for your GTM AI" 43 and bought DoubleO.ai for agent orchestration on September 30 44. Profound shipped a Context Manager so its AI marketer works "with the complete context of your brand" 45. Apollo's GTM Harness is "our purpose-built AI and context engine" 46. Vasco shipped citation cards that trace every number in an agent's report back to the accounts behind it 47. And the acquirers bought context, not features: ServiceNow bought Sweep's metadata layer for Salesforce orgs 48, HubSpot absorbed Trigify's signal team 49, and Superhuman bought the meeting notetaker Fathom 50. Adam Schoenfeld's September readout measured the same convergence from the other side, finding "context" in a rising share of GTM vendor homepages [context — October 1, 2026] 51.
Two cautions. Every number above comes from a vendor selling the cure, and HubSpot's own keynote opened with the admission that only about 6% of organizations report significant positive AI impact 52. And context is now positioned as the explanation for missing ROI before anyone has published an independent before-and-after. The tactical read for a GTM team is narrower than the marketing: fix the fields your agents read — owner, stage, close date, last touch — before buying a context product to paper over them.
Inbound became the first agent win with receipts, while "AI SDR" lost its name
The strongest operator evidence of the month was inbound, and it was reproducible. Anthropic's own sales team replaced its Contact Sales form with a buying agent built by one engineer in a few weeks; escalated buyers "turn into opportunities more than twice as often as leads from the old form, and close about five days faster," and the share of conversations needing a human to close "has fallen by about half" 53. Vercel's COO reported 90% automation of inbound sales development, an inbound SDR team cut from 10 to 1.25, a 32x return on the agent and an infrastructure bill "in the single-digit thousands" 54. Clay quantified the opportunity from the buyer's side: its agents submitted 6,346 demo and contact forms and found that "speed to lead isn't a priority at most companies" 55.
Outbound got rebuilt and renamed. Salesforce launched seven named "job-ready" agents including Piper for inbound and Hunter for outbound, which is in pilot with general availability in November 56, with Apollo named Hunter's flagship data partner 57. AiSDR relaunched as Ami, "the AI GTM agent," its CEO conceding the "AI SDR market overpromised and underdelivered" 58. Nooks bought the sales-coaching platform FullyRamped 59; Gong launched Gong Enrich, a multi-provider enrichment waterfall that fills remaining gaps with Gong Credits 60; Apollo repositioned as an "AI GTM System" 61; and Zoom began selling Common Room inside a "revenue OS" 62.
The headcount story is a split, not a collapse. Salesforce told investors it has 15,000 account executives and plans 1,200 more 63 while Vercel's inbound team shrank by 88%. One AI automation firm published its own post-mortem of 11,458 automated outbound messages and no replies 64. The pattern for GTM teams: automate the qualification queue first, where buyers already raised a hand and response time decides the deal, and treat autonomous cold outbound as the experiment rather than the default.
The lead record and the plugin marketplace became the attack surface
August established that granting an agent production access had become a legal question [context — August 2026]. September showed exactly where GTM agents get compromised. Zenity Labs disclosed SalesBleed: a poisoned web-to-lead submission could hijack Agentforce agents into exfiltrating CRM data and sending phishing under the agent's identity, with no click required 65. Plugin4Shell, a zero-click remote-code-execution flaw across Claude Code, Codex, Gemini CLI and GitHub Copilot, attacked the plugin marketplaces themselves 66 — the same install path every GTM vendor shipped into this month. And Claude Code in Auto Mode could be made to run attacker code up to 80% of the time just by being asked to summarize a website 67, which is the account-research workflow GTM engineers run daily. Decision models aren't exempt: VentureBeat showed prompt injection can steer Jev's verdicts 68.
The labs responded by withholding capability. OpenAI apologized to Australia after an internal model gained non-public access to government health statistics systems, and paused tool-use training for its most capable models 69; it shelved GPT-6.1 Astra the day before DevDay because it "didn't quite meet the bar in terms of staying within scope and authorization" 70. Dario Amodei called for frontier labs to "slow the pace at which we improve the capabilities of AI models" 71, and Google released Gemini 4 Argon only to vetted cyber defenders 72. The defensive tooling arrived alongside: Claude Managed Agents keeps credentials in a vault the agent never sees 73, NVIDIA open-sourced an agent safety platform 74, and CrowdStrike found 18,000 agents running at one Fortune 500 company that had approved only 300 75. August's advice — separate the agent that reads from the agent that writes — now has a concrete version: treat every inbound form field and enriched record as untrusted model input, pin plugin versions, and keep sending and CRM-write credentials out of the agent's context.
New Tools
Gong Mission Callisto 60
Gong · September 30, 2026
Gong Enrich runs a configurable waterfall across the data providers you already pay for — Apollo, ZoomInfo, Lusha, Wiza and others — and fills the remaining gaps with Gong Credits; Agent Builder agents can now trigger automatically on events. Expect enrichment budget to start migrating into the conversation-intelligence contract.
Apollo Next: GTM Harness, Builder Studio and Messaging OS 46
Apollo · September 30, 2026
Four agents available now for leads, sequence creation and optimization and task prioritization, plus closed betas for a natural-language builder and marketing email. Apollo is consolidating against both outbound point tools and marketing automation spend.
Clay Skills Marketplace 26
Clay · September 29, 2026
One-line installs of GTM playbooks into Claude Code or Codex, published by GTM engineers and agencies. Installs are free; runs consume Clay credits — so every skill you add is third-party logic running against your budget and data.
Build plugins for Claude and Claude Marketplace 32
Anthropic · September 25, 2026
A submission portal with validation and analytics for plugins, two days after a Marketplace that lets companies spend existing Anthropic commitments on partner tools 76. A distribution channel for any GTM tool built as a plugin.
HubSpot Sales Plugin for Claude 4
HubSpot · September 25, 2026
Eight free skills — Call Prep, Daily Brief, Log Call, Follow Up, Pipeline Pulse and more — on top of an MCP server that now writes deals, campaigns and emails with audit-logged attribution 27. The second major CRM to ship a vendor-authored skill pack in a month.
AI Ads in AirOps 77
AirOps · September 24, 2026
Tracks where your ads and competitors' ads appear across prompts in ChatGPT, then builds and measures campaigns. AirOps says ads went from 0% to more than half of tracked answers in three months — paid placement is now part of AI search visibility.
Alexandria in Claude and Claude Code 78
Firecrawl · September 24, 2026
One API over 100-plus data providers for agents, callable from Claude — website, filings and financial data in a single account brief. An account-research data layer that doesn't need a separate enrichment tool.
Phave 79
Phave · September 23, 2026
Marketo co-founder Jon Miller's new marketing automation, which uses person, account and buying-group context to drive journeys and scoring "instead of following rules." Twenty years after Marketo, the founder of rules-based automation is betting against it.
Next Gen Zaps 80
Zapier · September 23, 2026
Describe a Zap in Claude, ChatGPT or Cursor via MCP and deploy it, with list processing no longer capped at 500 items. Removes a long-standing ceiling on bulk enrichment and CRM-hygiene workflows.
jev-gtm-cookbook 11
Open source · September 22, 2026
Fifteen outbound recipes on Jev — score a lead list or LinkedIn network against your ICP, catch job changes, triage replies — locally, with zero dependencies. The fastest way to see what a decision model does to the cost of scoring.
Headless Dashboard Initiative 30
Resend · September 22, 2026
Most of what was dashboard-only in Resend's email platform is now reachable through MCP, CLI or API, including webhook replay and email metrics. A template for "headless" that other GTM vendors are following.
n8n Agents (preview) 81
n8n · September 21, 2026
Agents with instructions, tools, skills, sub-agents and memory that can call existing n8n workflows and MCP servers as tools. Self-hosted GTM teams get an agent layer over the automations they already run.
HubSpot Fall 2026 Spotlight 40
HubSpot · September 16, 2026
A Breeze Assistant that assigns agents and delivers proposals and campaign plans, a self-updating Smart CRM, and the first CRM integration with ChatGPT Ads 82. Paid AI-answer placement now has CRM attribution built in.
Pipedrive Nova 21
Pipedrive · September 16, 2026
AI meeting intelligence included in every plan at no extra cost. With Superhuman buying Fathom the same week, the standalone notetaker line item is under pressure from both directions.
Clay Plugin for Claude Code and Codex 25
Clay · September 15, 2026
Clay's CLI and API in one interface for coding agents, now with Audiences, Sequencer and Query Tables. Combined with Sequencer's in-app domain and inbox provisioning 83, a coding agent can now run list, enrich and send end to end.
Klaviyo Headless 29
Klaviyo · September 9, 2026
More than 260 Klaviyo MCP tools callable from Claude, ChatGPT or any agent, with full read and write. B2C marketing's system of record went agent-first in the same month as B2B CRM.
Ami by AiSDR 84
AiSDR · September 9, 2026
An agent that builds the full outbound motion for founders and small GTM teams, built on 2.5 million emails and nearly 20,000 booked meetings. Notable mostly for what it dropped: the word "SDR."
Profound Context Manager 45
Profound · September 1, 2026
Persistent brand, audience and product context for Profound's AI marketer, with gaps surfaced so teams can fill them. Two weeks before its $180M Series D, Profound repositioned from visibility tracker to marketing platform.
Labs Watch
Alibaba released Qwen3.8-Flash-Next as open weights on August 24 85, continuing the cheap open tier that underpins self-hosted enrichment and classification.
Anthropic shipped three models in four weeks. Claude Fable 5.1 arrived September 1 with a 75% cut to cached-context prices for persistent agents 86; Opus 5.5 followed September 22, matching Fable 5.1 on most work at a 20% lower list price and, Anthropic says, 40% less to run than Opus 5 8; Sonnet 5.5 on September 28 at an unchanged price with fewer tokens per task 18. Anthropic says Claude Code sessions now run an input-to-output ratio of 324 to 1, so agent loops are dominated by cached input 87 — cache hit rate is the cost lever that matters. It also published a guide to building an AI-native revenue organization 88, opened the Claude Marketplace 31, and its CEO called for pacing the frontier 71. A leaked IPO filing reported by SiliconANGLE shows rapid revenue growth alongside an $8 billion operating loss 89 — the economics of the vendor most GTM agents now run on are, for the first time, partly visible.
DeepSeek released DeepSeek-V4.1-Flash with an API price cut on September 10 90, keeping pressure on the low end of the price curve.
Google DeepMind shipped Gemini 3.8 Flash on September 2 91 and announced Gemini 4 Argon on September 30 at $2/$10 per million tokens with a one-million-token output limit — released first only to vetted cyber defenders 72. August flagged key-person risk at Google [context — August 2026]; the shipping cadence answered it, but the risk reappeared as access: one of the month's strongest models isn't one most teams can buy yet.
Meta launched Muse, a personal agent that can browse, fill forms and check out with Stripe's Link, on September 8 92 93; Shopify turned on Shop Pay checkout in Muse for its stores two weeks later 94. It followed with a Meta Enterprise Platform on September 28 95 and hired MongoDB's CEO to run it 96. Meta is now selling agents directly to the businesses that buy GTM software.
Mistral raised €3 billion on September 8 to push sovereign, open-weight models 97 — the procurement path for EU GTM stacks with data-residency requirements.
OpenAI shipped GPT-6 Astra on September 3 98, the Agents API on September 10 99, cheaper Sol and Luna on September 22 7, and at DevDay on September 29 always-on "dots" agents, GPT-6.1 Sol and a Decisions API 100 101. It began testing Sponsored Agents in ChatGPT and named HubSpot its first CRM partner for ChatGPT Ads 102. It is also winding down Cursor's access to its models after the SpaceX acquisition, with a proposed shutoff date of November 12 (announced August 28) 103 — GTM builds that run on GPT models inside Cursor need a migration plan. And it shelved GPT-6.1 Astra over safety 70.
TypeSafe AI, not a frontier lab but the month's most consequential new model company for GTM, launched Jev and its "System One" category on September 15 10 on a $40M seed led by DCVC 104.
xAI launched Grok Bot for Enterprise on September 3 105, Grok 4.7 at $2/$6 on September 21 106, and Team Bots on September 28 — shared bots per role, with "every major sales account" at SpaceXAI getting its own bot that reads Gong calls and Slack overnight and posts a morning briefing 107. That per-account pattern is reproducible on any harness.
Zhipu released GLM-5.3 as open weights on August 25 108; Anthropic published research on September 29 on the spread of advanced cyber capabilities it found in the model 109.
Funding
ZoomInfo acquires DoubleO.ai — Acquisition 44
GTM agent orchestration · September 30, 2026
The data incumbent bought agent orchestration so the context it sells can also run agents. Terms undisclosed.
Bolt.new acquires Dokai — Acquisition 110
GTM agents · September 29, 2026
The app builder's first acquisition is a GTM-agents team — a signal that "build the app" and "sell the app" are converging in the same tools.
Salesforce to acquire Listen Labs — Acquisition · reported ~$2B 111
AI customer research · September 29, 2026
Agents that design studies, recruit from a network of more than 50 million people, run interviews in 120-plus languages and simulate customer responses before launch. Salesforce did not disclose terms; The Stack reported about $2 billion 112. AI customer research went from Conveo's $50M Series A to a reported $2 billion exit for Listen Labs inside one month.
HubSpot acqui-hires the Trigify team — Acquisition 49
Social signal monitoring · September 23, 2026
HubSpot's second signals absorption after Warmly; the standalone product winds down October 22, so Trigify customers have a hard migration deadline.
Backstory acqui-hires Pod — Acquisition 113
AI deal intelligence · September 22, 2026
The former People.ai adds a deal-coaching team as revenue intelligence consolidates.
Searchable acquires Meridian — Acquisition · seven figures 114
AI search visibility · September 21, 2026
One of the first deals between two dedicated AEO startups, six days after Profound's raise — the long tail of AI-visibility trackers is starting to roll up.
VWO and AB Tasty merge as Wingify — Merger 115
Experimentation and personalization · September 18, 2026
Two experimentation platforms now share one suite brand and one data and AI foundation.
Superhuman acquires Fathom — Acquisition 50
Meeting intelligence · September 14, 2026
A free notetaker with 400,000 monthly users becomes a feature of the inbox suite. Meeting transcripts are raw material for GTM context, and the platforms that own the inbox are buying them.
Salesforce completes acquisition of Fin — Acquisition · ~$3.6B 116
AI customer service agents · September 10, 2026
Fin, formerly Intercom, brings more than 30,000 customers and a 76% average resolution rate; Salesforce's investor release priced the June agreement at about $3.6 billion 117. Resolution rate is now the benchmark for any per-resolution price.
ServiceNow acquires Sweep — Acquisition 48
Salesforce metadata and RevOps agents · September 8, 2026
ServiceNow bought the context layer for Salesforce orgs to sharpen its CRM push.
Monaco acquires Overlayy — Acquisition 36
Sales conversation context · September 5, 2026
Monaco, building "the AI-native revenue platform" to replace CRM plus point tools, bought a Bengaluru team that taught AI to understand the context behind sales conversations.
SoundHound AI completes acquisition of LivePerson — Acquisition 118
Conversational customer engagement · September 4, 2026
The same week Salesforce closed Fin, the service-agent field narrowed a second time.
Adobe acquires Rilo — Acquisition 119
Marketing workflow automation · September 2, 2026
A six-person marketing-intelligence team folded into Adobe's customer experience products.
Capacity acquires Chattigo — Acquisition 120
AI customer support · August 27, 2026
Days before its own Series E, Capacity bought a Chilean conversational-support platform.
Integrate acquires CaliberMind — Acquisition 121
B2B marketing attribution · August 26, 2026
Demand-generation data meets attribution analytics as B2B marketing measurement consolidates.
Nooks acquires FullyRamped — Acquisition 59
AI sales coaching · August 25, 2026
The dialer adds voice role-play coaching on its way to a broader "revenue agents" platform.
Wonderful — $550M · Series C 122
Insight Partners · Enterprise AI agents · September 2, 2026
The largest bet of the month on forward-deployed enterprise agent services — four weeks before Gartner predicted most such vendor-built deployments will be abandoned (see Discussion).
Bird.com — $450M · Debt financing 123
J.P. Morgan, Capital One, Citi · Messaging infrastructure for AI agents · September 23, 2026
Financing for messaging and communications infrastructure aimed at AI agents — the rails that agent-sent marketing and support messages run on.
Owner — $240M · Growth 124
Goldman Sachs Alternatives · AI marketing and commerce for local businesses · August 28, 2026
Growth equity is paying for marketing done for the customer, not marketing software the customer operates.
Profound — $180M · Series D 125
Sequoia, Kleiner Perkins · AI search visibility and marketing · September 15, 2026
At $1.8 billion, less than seven months after its Series C 126. AI search visibility is now a funded budget category with a leader.
Clay — $115M · Series D 127
Wellington · GTM data and orchestration · September 9, 2026
At $7.1 billion after 4x revenue growth in 2025, with more than 17,000 customers including 80% of the Forbes AI 50. Clay pairs it with a $1M scholarship fund for GTM engineers 128 — the category is funding its own talent pipeline.
Firecrawl — $75M · Series B 129
Smash Capital · Web data for agents · September 22, 2026
Web data for agents is now a funded infrastructure layer that GTM enrichment depends on, distributed through the labs' plugin directories.
Metaview — $60M · Series C 130
Insight Partners · Talent intelligence · September 30, 2026
Agentic recruiting built on interview data 131 — the hiring side of GTM gets the same "shared memory" pitch as the selling side.
Capacity — $54M · Series E 132
AI customer experience · September 2, 2026
Past $100M ARR in support automation — support agents keep compounding while sales agents rebrand.
Conveo — $50M · Series A 133
DST Global Partners, Balderton Capital and others (no lead named) · AI consumer research · September 2, 2026
The category Salesforce bought into four weeks later with Listen Labs.
Lightfield — $47M · Series A 34
Andreessen Horowitz · AI-native CRM · September 9, 2026
More than 5,000 companies signed up since November, with some migrating entire Salesforce implementations; Lightfield pitches pricing that follows work done rather than seats.
SciFin — $44M · Seed 134
Altimeter, Madrona · Revenue intelligence · September 1, 2026
A two-time decacorn founder's revenue-data unification layer, co-led by Seattle's Madrona.
TypeSafe AI — $40M · Seed 104
DCVC · Decision models · September 15, 2026
The company behind Jev (see Top 3).
OuterSignal — $22M · Series A 135
Long Journey, Abstract · Customer intelligence for consumer brands · September 30, 2026
B2C marketing's version of the context-layer pitch.
Levanta — $22M · Series B 136
Volition Capital · Creator affiliate marketing · August 27, 2026
Seattle's creator-commerce platform, with 2026 revenue up 80% and integrations into ChatGPT and Claude.
Atira — $17.5M · Pre-seed and seed 137 138
Accel · AI sales engineering for industrial B2B · September 3, 2026
Agents for quoting and sales engineering in industrial selling, a GTM vertical the horizontal tools rarely reach.
Zero — $10.3M · Seed 35
Primary Venture Partners · AI-native GTM operating system · September 15, 2026
Backed by the founders of Lovable and Supercell, building "what comes after CRM"; it claims customers close 2x more new business with a third fewer sales hires.
Actionable — €8.6M · Venture 139
Hi Inov · Predictive customer intelligence · September 9, 2026
Behavioral predictions from customer data for churn and expansion.
Zeliq — €7M · Seed extension 140
The Moon Venture · Sales prospecting agent · September 22, 2026
An agentic prospecting and enrichment workspace launching its own sales agent.
Euler — $4.3M · Seed 141
Channel Equity Partners · Agentic partner management · August 24, 2026
Partner and channel sales get their own agent layer.
Lightsage — $4M · Seed 142
Nexus Venture Partners · Agent-led growth · September 8, 2026
A growth stack for selling to AI agents as customers — it names the category "agent-led growth," after PLG.
Decimal AI — $4M · Seed 143
Khosla Ventures, Kearny Jackson · Technical customer support · September 15, 2026
Customer engineering for software companies, the technical end of post-sale.
Y Combinator Summer 2026 GTM cohort — YC standard deal 144
Y Combinator · GTM agents and tooling · September 10, 2026
At Demo Day, the GTM companies included Nex, building AI for high-volume GTM engineering 145; LemonLime, automated outbound and marketing for small businesses; Palisade, sales agents for marketplaces; COACH, field-sales coaching; Poth Labs, customer intelligence; Studio, market simulation; Chromie, government sales; Dialogus, AI contact center; and Inkbox, identity and communications for agents. "GTM engineering" is now a seed-stage product category, not only a job title.
Revenue milestones
Three GTM-adjacent companies disclosed run-rates that set this month's ceiling. Viktor, an AI coworker for growth, said it reached $50M in annualized revenue seven months after launch 146; Higgsfield, creative infrastructure for marketing, crossed $1 billion annualized 18 months from launch 147; and lemlist passed $60M ARR 17. All three are founder disclosures, not audited figures.
From Seattle
The local GTM community ran its August meetup in a new format: AI Tinkerers Seattle's GTM track on August 26 replaced talks with a peer roundtable and a "GTM Changelog" of the month's tool updates, with more than 107 applicants for 50 seats 148. The next session, on October 7, returns to live demos — voice agents on Gemini Live and context distillation for B2B ads among them 149.
Seattle's GTM vendors were on the month's main storylines. Outreach reported 12x growth in AI credit consumption and 480% year-over-year AI ARR growth on September 9, with a SolarWinds win-back agent reaching a 45% reply rate 150 — credits again growing faster than revenue. A week later it became the first revenue orchestration platform available for private offers on Salesforce's AgentExchange 151. Amperity put its MCP server in the Claude directory and ChatGPT marketplace on September 24, so customer-data queries and audience segments run from inside an assistant under existing permissions 152. And Seismic, which absorbed Seattle's Highspot in August, named Umberto Milletti chief product and technology officer and committed more than $100 million to its revenue-execution roadmap 153.
The money was local too: Madrona co-led SciFin's $44M seed 154, Levanta raised a $22M Series B, and Kanu AI emerged from stealth with $11.7M led by Trilogy Equity Partners to capture employee judgment for enterprise agents 155. In Redmond, Microsoft's Dynamics 365 Activate targets Salesforce customers directly with AI-driven migration 37. The subtraction came from the large employers: Oracle cut 359 Washington jobs 156, Microsoft cut hundreds more 157, and Qualtrics cut 117 in Seattle 158 — a growing local pool of experienced tech talent.
Discussion
Not every signal points in the same direction, and several of this month's consensus stories rest on thinner evidence than their volume suggests.
Is context the cure or the excuse? HubSpot's data says AI on bad context performs worse than no AI 38, and every major vendor now sells context. But the evidence for the cure comes from the vendors selling it, and independent signals mostly confirm the problem rather than the fix: Gartner's Daryl Plummer said the leading AI vendors "are not enterprise grade" 159, Winning by Design's research argued "fluency is not productivity" 160, and VentureBeat found companies "spending millions rewiring how AI gets used" while almost none can prove it works — Uber, it reported, ranked teams by Claude Code token use and spent its 2026 AI coding budget by April 161.
Did the AI SDR die, or only its name? AiSDR became Ami 84, Nooks and Gong talk about "revenue agents," and Chronexa published a post-mortem of 11,458 automated outbound messages that drew zero replies 64. Yet Salesforce plans to hire 1,200 more account executives by fiscal year-end 63 and First Round's VP of GTM argued sales is "an AI-proof job" because the hard part was always judgment and trust 162, and HubSpot's CMO Kipp Bodnar and Kieran Flanagan launched a marketing book, Loop, on the premise that "when everyone has access to the same AI tools, producing more is easy" 163. The working reading: inbound qualification is automating fast; cold outbound by agent remains unproven at scale.
Is the SaaSpocalypse over? Benioff called it "crazy nonsense" at Dreamforce 164, Atlassian's CEO pushed back on The Verge's Decoder in an episode titled "The SaaSpocalypse that wasn't" 165, and software stocks hit a high after Salesforce's quarter 166. The same weeks brought SEC-filed cuts at Workday 167, The Trade Desk 168, Zscaler 169 and PagerDuty 170, and ZoomInfo's install data showed small companies dropping point solutions at four times last year's rate while systems of record barely moved 171. Gusto's payroll data cut the other way: small businesses using AI grew headcount about 7% more than those that didn't 172. It may be less an apocalypse for software than for point solutions — and for the headcount inside software companies.
Are buyers' agents a GTM channel yet? Meta's Muse 92, OpenAI's dots 100 and Instinct's $1 billion round at a $10 billion valuation 173 say buyers will soon send agents to research and purchase, and NielsenIQ and Similarweb announced a joint product to measure AI-driven discovery and verified sales across the major assistants, due in the fourth quarter 174. Ads moved into AI answers in the same month, contradicting August's call that answer-engine optimization had deflated [context — August 2026]: OpenAI began testing Sponsored Agents 102 and HubSpot integrated ChatGPT Ads 82. But Rand Fishkin laid out why AI-visibility tracking can't yet be trusted — answers vary on every run and no one has a source of truth for prompt volume 175 — and Amanda Natividad documented Reddit's share of ChatGPT citations falling about 80% in a week 176. Measurement is lagging the money.
Rent the agent or own the build? Insight led a $550M round into Wonderful's forward-deployed enterprise agents on September 2. On September 29 Gartner predicted 70% of enterprises will abandon agentic AI built by vendor forward-deployed engineers by 2028, citing costs and an inability to evolve 177. Both can be right if the services build is a bridge to an internal one.
The practical caution for a GTM team is the same across all five: the vendor that measures the outcome is usually the vendor selling it. Set your own baseline — reply rate, speed to lead, opportunity conversion, credit burn — before a pilot starts, so the result is yours rather than theirs.
Further Reading
When a Smarter Model Stops Mattering 178
Theory Ventures · September 29, 2026
Why read: the clearest version of the investor case that intelligence is commoditizing and value is moving to context and the applications around it — the logic behind half this month's funding.
Lessons from building 113 AI agents and reinventing a GTM function 179
GTMnow · September 29, 2026 · podcast
Why read: a public-company CMO on what running more than 100 agents with flat headcount actually requires, from the operator side rather than the vendor's.
My post-conference playbook 180
Adam Schoenfeld · September 24, 2026
Why read: a fully specified build — Codex orchestrating, Apify collecting 7,135 Dreamforce and UNBOUND posts, Jev classifying — that turns any event your buyers attend into an ABM list and messaging angles.
SaaStr 879: Stripe's CRO of AI on the new AI GTM playbook 181
SaaStr · September 23, 2026 · podcast
Why read: how a revenue leader whose title includes "AI" structures selling to AI-native buyers.
Jev: System One Models for Prod, Not God 182
Latent Space · Diogo Almeida · September 21, 2026 · podcast
Why read: TypeSafe's founder on why decisions, not generation, are the production workload — useful before you redesign a scoring or routing pipeline around a decision model.
CRM As A Business World Model 183
Topline · Keith Peiris · September 20, 2026 · podcast
Why read: the argument for replacing rather than wrapping the CRM, from the founder a16z just funded to do it.
The surprisingly human future of sales 162
First Round Review · Emery Rosansky · September 17, 2026
Why read: the strongest counterweight this month to automation-first GTM, from a 20-year seller who argues AI makes the human parts of selling more valuable.
Inside 12 months of AI search experiments 184
Growth Unhinged · Aja Frost · September 16, 2026
Why read: nine AI-search experiments run weekly at HubSpot, including the ones that failed, with a claimed 1,850% rise in qualified leads from AI — a working playbook rather than a framework.
When Inbound Sells Itself 54
Tomasz Tunguz · September 15, 2026
Why read: Vercel's inbound numbers in one page, and the point that "the bottleneck is not the model, it is the codified workflow."
The Incumbents Are Coming 33
Andreessen Horowitz · September 3, 2026
Why read: the month's defining strategy essay — every AI-native GTM startup now has to answer whether a system of record plus a general agent is enough.
What AI agents really think about your pricing 185
Growth Unhinged · Kyle Poyar · September 2, 2026
Why read: 7,600 AI answers to pricing queries across the Cloud 100 show only 57% of companies have AI-readable pricing — a concrete fix for any team selling to buyers who research with assistants.
You are not a model. Don't price per token. 24
Andreessen Horowitz · August 27, 2026
Why read: the case against passing token pricing through to customers, written the week before the largest CRM vendor moved its editions to credits 186.
Sources
- 1.Salesforce and Anthropic Announce Claudeforce: The #1 AI Meets the #1 AI CRMSalesforce · August 26, 2026
- 2.Bringing Salesforce into ClaudeClaude by Anthropic · September 15, 2026
- 3.Salesforce Unveils AIforce, Bringing the Full Power of Its Platform to Any InterfaceSalesforce · September 15, 2026
- 4.Introducing the HubSpot Sales Plugin for ClaudeHubSpot Community · September 25, 2026
- 5.New Salesforce Editions Bundle Everything Businesses Need for Agentic TransformationSalesforce · September 3, 2026
- 6.HubSpot Unveils AI Agent Strategy, Lifts 2030 Margin Targets at Analyst DayMarketBeat · September 18, 2026
- 7.Introducing GPT-6 Sol and LunaOpenAI · September 22, 2026
- 8.
- 9.September 2026 Ramp AI Index: Cracks in the AI thesis, part 2Ramp · September 9, 2026
- 10.Introducing System One Models & JevTypeSafe AI · September 15, 2026
- 11.
- 12.OpenAI’s Jev clone could help the frontier lab stop its swarming agentsTechCrunch · September 30, 2026
- 13.HubSpot's Starter plan just got a whole new refresh. Here's what's newHubSpot Community · September 29, 2026
- 14.Workday (WDAY) Q2 2027 Earnings Call TranscriptThe Motley Fool · August 31, 2026
- 15.GitLab Reports Second Quarter Fiscal Year 2027 Financial ResultsSEC EDGAR · September 1, 2026
- 16.Asana (ASAN) Q2 2027 Earnings Call TranscriptThe Motley Fool · September 9, 2026
- 17.We crossed $60M ARR. Here are a few things I've changed my mind onLinkedIn · Charles Tenot · September 21, 2026
- 18.Introducing Claude Sonnet 5.5Anthropic · September 28, 2026
- 19.
- 20.New in Clarify: September 2026Clarify · September 10, 2026
- 21.Pipedrive launches Nova to turn sales meetings into CRM intelligencePipedrive · September 16, 2026
- 22.Salesforce blames its Claude addiction for denting profit margin guidanceThe Register · September 3, 2026
- 23.Salesforce wants to charge for AI outcomes, but first it needs to figure out howThe Register · September 21, 2026
- 24.You are not a model. Don’t price per token.Andreessen Horowitz · August 27, 2026
- 25.Build custom GTM tooling on ClayClay · September 15, 2026
- 26.The Clay Skills Marketplace Is OpenClay · September 29, 2026
- 27.Spotlight Feature: HubSpot MCP server updatesHubSpot Community · September 15, 2026
- 28.Outreach product release notes: August 2026Outreach · September 1, 2026
- 29.K:BOS 2026: 4 Updates Powering Klaviyo's Autonomous B2C CRMKlaviyo · September 9, 2026
- 30.The Headless Dashboard InitiativeResend · September 22, 2026
- 31.Claude Marketplace: one place to discover plugins, agents, and services from our partnersClaude by Anthropic · September 23, 2026
- 32.Build plugins for ClaudeClaude by Anthropic · September 25, 2026
- 33.The Incumbents Are ComingAndreessen Horowitz · September 3, 2026
- 34.Our $47M Series A led by Andreessen HorowitzLightfield · September 9, 2026
- 35.We raised $10.3M to build what comes after CRMZero · September 15, 2026
- 36.Overlayy is becoming a part of MonacoLinkedIn · Anmol Chaman · September 5, 2026
- 37.Accelerate your move to agentic business applications with Dynamics 365 ActivateMicrosoft · September 9, 2026
- 38.The Outcomes Era is HereHubSpot · September 21, 2026
- 39.UNBOUND 2026: How HubSpot Rebranded Its Flagship Eventbizbash.com · September 28, 2026
- 40.HubSpot Fall 2026 SpotlightHubSpot · September 16, 2026
- 41.Salesforce Introduces the Trusted Enterprise AI HarnessSalesforce · September 10, 2026
- 42.Announcing Koa: Salesforce's First CRM Reasoning Model, Built on NVIDIA NemotronSalesforce · September 15, 2026
- 43.
- 44.ZoomInfo acquires DoubleO.ai for AI agent orchestrationInvesting.com · September 30, 2026
- 45.Introducing Context Manager in ProfoundProfound · September 1, 2026
- 46.The future of GTM: Apollo NextApollo · September 30, 2026
- 47.Agent citations: see the accounts behind every numberVasco · September 15, 2026
- 48.ServiceNow Sharpens Its Salesforce Attack With Sweep AcquisitionCRN · September 8, 2026
- 49.The Trigify team is joining HubSpotTrigify · September 23, 2026
- 50.Superhuman acquires YC-backed notetaker Fathom as productivity platforms push for agentic workTechCrunch · September 14, 2026
- 51.Everything I Measured in AI x GTM: September ReportAdam Schoenfeld · October 1, 2026
- 52.UNBOUND 26: Spotlight RecapKuno Creative · September 18, 2026
- 53.How Anthropic's sales team rebuilt inbound with Claude Managed AgentsClaude by Anthropic · September 30, 2026
- 54.When Inbound Sells ItselfTomasz Tunguz · September 15, 2026
- 55.We asked 6,346 companies for a demo. Most never wrote back.Clay · September 22, 2026
- 56.Salesforce Expands Agentforce With a New Portfolio of AI Agents Built for High-Value WorkSalesforce · September 11, 2026
- 57.Apollo Named Flagship Data Partner for Hunter, Salesforce's New Outbound Sales AgentPR Newswire via Morningstar · September 15, 2026
- 58.I am betting my entire company on this launch. Introducing AmiLinkedIn · Yuriy Zaremba · September 9, 2026
- 59.Nooks acquires FullyRamped, an AI-native sales coaching platformNooks · August 25, 2026
- 60.Gong announces Mission CallistoGong · September 30, 2026
- 61.Apollo Reveals the AI GTM System Bringing Data, Intelligence and Execution into a Single EngineApollo · September 30, 2026
- 62.Zoom unveils AI-powered revenue OSZoom · September 15, 2026
- 63.Salesforce (CRM) Q2 2027 Earnings Call TranscriptThe Motley Fool · August 31, 2026
- 64.AI Sales Agents: What Breaks at Scale, From Our Own DataChronexa · September 12, 2026
- 65.Salesforce Agentforce vulns allowed 0-click CRM data theft, anonymous phishingThe Register · September 24, 2026
- 66.AI coding agents' 0-click RCE flaw could hand attackers keys to the kingdomThe Register · September 17, 2026
- 67.Researcher shows how Claude Code can be tricked simply by asking it to summarize a websiteThe Register · August 28, 2026
- 68.Jev AI agent security: Prompt injection riskVentureBeat · September 21, 2026
- 69.How we will do better for AustraliaOpenAI · September 28, 2026
- 70.OpenAI abandons plan to release upcoming model as safety concerns escalateCNBC · September 28, 2026
- 71.We Must Pace the FrontierDario Amodei · September 12, 2026
- 72.Gemini 4 Argon: our next era of frontier intelligenceGoogle · September 30, 2026
- 73.Giving companies more control over their AI agents, with NVIDIAClaude by Anthropic · September 28, 2026
- 74.NVIDIA Launches Open Agent Safety Platform to Secure Agents From Testing to DeploymentNVIDIA · September 28, 2026
- 75.Falcon Guardian finds 17,700 unapproved agentsVentureBeat · September 8, 2026
- 76.How CodeRabbit, Power Digital, and ThoughtSpot scale with Snowflake and Vercel on Claude MarketplaceClaude by Anthropic · September 23, 2026
- 77.The paid layer of AI search has arrived: Introducing AI Ads in AirOpsAirOps · September 24, 2026
- 78.Alexandria is now in Claude and Claude CodeFirecrawl · September 24, 2026
- 79.Marketo Co-Founder Jon Miller Launches Phave, Marketing Automation That Reasons Instead of Following RulesPR Newswire · September 23, 2026
- 80.The next generation of Zaps is hereZapier Community · September 23, 2026
- 81.Build agents you set up once and use anywheren8n · September 21, 2026
- 82.HubSpot and OpenAI deepen partnershipHubSpot · September 16, 2026
- 83.Clay Sequencer is generally availableClay · September 1, 2026
- 84.Ami by AiSDR: AI GTM agent that builds your full outbound motionY Combinator · September 9, 2026
- 85.
- 86.Introducing Claude Fable 5.1 and Claude Mythos 5.1Anthropic · September 1, 2026
- 87.Coding sessions are longer and use more context. Claude Opus 5.5 is built with that in mind.Claude by Anthropic · September 24, 2026
- 88.Building an AI-native revenue organizationClaude by Anthropic · September 15, 2026
- 89.Leaked Anthropic IPO filing reveals $8B operating loss, rapid revenue growthSiliconANGLE · September 29, 2026
- 90.DeepSeek-V4.1-Flash: Smarter, Faster, More EfficientDeepSeek · September 10, 2026
- 91.Introducing Gemini 3.8 Flash and 3.8 Flash CyberGoogle · September 2, 2026
- 92.Introducing Muse: The World's First Personal AI Agent Built for EveryoneMeta · September 8, 2026
- 93.Stripe helps Muse, Meta's new personal AI agent, shop across the internet with Linkstripe.com · September 8, 2026
- 94.Shopify Brings Shop Pay Checkout to Meta's Muse AI Agentpymnts.com · September 22, 2026
- 95.Launching Meta Enterprise PlatformMeta · September 28, 2026
- 96.MongoDB Announces CEO TransitionSEC EDGAR · September 28, 2026
- 97.Mistral raises €3B to make sovereign, open-weight AI the technology frontierMistral AI · September 8, 2026
- 98.GPT-6 Astra: A new generation of intelligenceOpenAI · September 3, 2026
- 99.Introducing the Agents APIOpenAI · September 10, 2026
- 100.
- 101.
- 102.Reimagining advertising with AIOpenAI · September 16, 2026
- 103.Our decision on Cursor following its acquisition by SpaceXOpenAI · August 28, 2026
- 104.TypeSafe AI Raises $40M SeedThe SaaS News · September 17, 2026
- 105.
- 106.
- 107.Team Bots: AI coworkers that learn from your teamxAI · September 28, 2026
- 108.
- 109.GLM-5.3 and the spread of advanced cyber capabilitiesAnthropic · September 29, 2026
- 110.Bolt.new Makes Its First Acquisition: DokaiBolt · September 29, 2026
- 111.Salesforce Signs Definitive Agreement to Acquire Listen LabsSalesforce · September 29, 2026
- 112.Salesforce buys AI research startup in latest sign of shifting strategyThe Stack · September 30, 2026
- 113.The Pod team is joining Backstory (formerly People.ai)Backstory · September 22, 2026
- 114.AI visibility platform Searchable acquires New York's Meridian in seven-figure dealPR Newswire · September 21, 2026
- 115.Inside Wingify: The Digital Experience Optimization Suite Driven by VWO and AB TastyVWO · September 18, 2026
- 116.Salesforce Completes Acquisition of FinSalesforce · September 10, 2026
- 117.Salesforce Completes Acquisition of FinSalesforce Investor Relations · September 10, 2026
- 118.SoundHound AI Completes Acquisition of LivePerson, Creating a World-Leading Omnichannel Conversational AI PowerhouseGlobeNewswire · September 4, 2026
- 119.Adobe Acquires Marketing Intelligence Startup Rilo With $1 Million In FundingTraded · September 2, 2026
- 120.Capacity acquires Chilean AI customer support startup ChattigoLatamList · August 27, 2026
- 121.MarTech Acquisition: Integrate and CaliberMind UniteMarTech Series · August 26, 2026
- 122.Tripling down on Wonderful: Building the AI operating system for the enterpriseInsight Partners · September 2, 2026
- 123.Bird.com secures $450m financing as it launches communications infrastructure for AI agentsPR Newswire · September 23, 2026
- 124.Owner Raises $240M Led by Goldman Sachs Alternatives to Build the AI-Native Platform for Every Local BusinessPR Newswire · August 28, 2026
- 125.Profound Raises $180M Series D at $1.8B Valuation to Build the AI Platform For Marketing TeamsGlobeNewswire · September 15, 2026
- 126.AEO startup Profound hits unicorn valuation, raises $180M Series D 7 months after last roundTechCrunch · September 15, 2026
- 127.Clay raises $115M to be the AI growth engine for every companyClay · September 9, 2026
- 128.Clay raises $115M Series D at $7.1B valuation, launching $1M GTME scholarshipClay Community · September 9, 2026
- 129.Introducing Alexandria and our $75M Series BFirecrawl · September 22, 2026
- 130.Behind the investment: MetaviewInsight Partners · September 30, 2026
- 131.Metaview Raises $60 Million Series C For Agentic Recruiting PlatformPulse 2.0 · September 30, 2026
- 132.Capacity Raises Over $50M in Series E to Expand its Unified AI-Native Customer Experience PlatformPR Newswire · September 2, 2026
- 133.
- 134.SciFin Emerges from Stealth with $44 Million in Seed FundingYahoo Finance (Business Wire) · September 1, 2026
- 135.OuterSignal Raises $22M Series A to Bring Agentic Personalization to Every Consumer BrandPR Newswire · September 30, 2026
- 136.Levanta Raises $22M Series B from Volition Capital to Deploy the Infrastructure for Creator Commerce MediaGlobeNewswire · August 27, 2026
- 137.Our Seed Investment in Atira: The AI Orchestration Layer for Industrial Sales EngineeringAccel · September 3, 2026
- 138.Atira Raises $17.5M in Fundingthesaasnews.com · September 4, 2026
- 139.Paris-based Actionable raises €8.6 million to scale its predictive customer experience platformEU-Startups · September 9, 2026
- 140.Exclusive: Zeliq raises €7M seed extension to launch its AI sales agent. Here’s its pitch deckTech Funding News · September 22, 2026
- 141.Bootstrapped to Profitability, EULER Raises $4.3 Million Seed to Accelerate Agentic Partner ManagementEIN Presswire · August 24, 2026
- 142.Lightsage raises $4M to build the growth stack for internet software's newest customer: AI agentsGlobeNewswire · September 8, 2026
- 143.Decimal AI Announces $4 Million Seed Round to Bring Customer Engineering to Software CompaniesPR Newswire · September 15, 2026
- 144.Y Combinator Summer 2026 batch directoryY Combinator · September 10, 2026
- 145.Nex: AI for high volume GTM EngineeringY Combinator · September 10, 2026
- 146.We've hit $50M in annualized revenueLinkedIn · Fryderyk Wiatrowski · September 17, 2026
- 147.Today, Higgsfield crossed $1B in annualized revenueLinkedIn · Alex Mashrabov · September 24, 2026
- 148.GTM AI Tinkerers - AugustAI Tinkerers Seattle · August 26, 2026
- 149.GTM AI Tinkerers - OctoberAI Tinkerers Seattle · September 30, 2026
- 150.Outreach reports 12x AI credit consumption growth and 480% YoY AI ARR growthOutreach · September 9, 2026
- 151.Outreach becomes first revenue orchestration platform available for private offers on Salesforce's AgentExchange (Headless)Outreach · September 15, 2026
- 152.Amperity MCP is now available through the Claude directory and ChatGPT marketplaceAmperity · September 24, 2026
- 153.Seismic appoints enterprise software pioneer Umberto Milletti as Chief Product and Technology OfficerSeismic · September 1, 2026
- 154.Revenue Teams Don’t Need Another Tool, They Need the Truth – Our Investment in SciFinMadrona · September 1, 2026
- 155.Kanu Emerges from Stealth to Build AGI for the EnterprisePR Newswire · September 30, 2026
- 156.Oracle cuts 359 Washington jobs in latest AI-driven layoffs; Seattle offices ‘not closing’GeekWire · September 16, 2026
- 157.Microsoft cuts hundreds more jobs, shifts next ‘Halo’ game to Activision in Xbox overhaulGeekWire · September 22, 2026
- 158.Software maker Qualtrics cuts 117 jobs in SeattleThe Seattle Times · September 2, 2026
- 159.AI and its main promoters are not enterprise-ready, says GartnerThe Register · September 14, 2026
- 160.WbD Research – September '26: Fluency is not productivityWinning by Design · September 10, 2026
- 161.Enterprise AI spending: the ROI is unprovenVentureBeat · September 7, 2026
- 162.The surprisingly human future of salesFirst Round Review · September 17, 2026
- 163.Kipp Bodnar on LinkedIn: Less than 24 hours after launch, Loop is the #1 new release in Marketing on AmazonLinkedIn · September 23, 2026
- 164."Crazy nonsense": Benioff dismisses the SaaSpocalypse and launches AIforceThe Next Web · September 16, 2026
- 165.The SaaSpocalypse that wasn’t, with Atlassian CEO Mike Cannon-BrookesThe Verge · September 28, 2026
- 166.Software stocks roar backAxios · August 28, 2026
- 167.Workday 8-K Items 2.02/2.05: ~2.5% workforce reduction, guidance reiteratedSEC EDGAR · September 29, 2026
- 168.The Trade Desk 8-K Item 2.05: organizational realignment cutting ~15% of workforceSEC EDGAR · September 4, 2026
- 169.Zscaler 8-K Item 2.05: ~3% headcount reduction to fund AI and growth initiativesSEC EDGAR · September 3, 2026
- 170.PagerDuty 8-K Item 2.05: headcount reduction of approximately 15%SEC EDGAR · August 27, 2026
- 171.GTM Exclusives Issue #4: The SaaSpocalypse is real, if you're not a system of record.ZoomInfo · September 23, 2026
- 172.Small businesses that adopt AI hire more workersAxios · September 10, 2026
- 173.Viral AI agent Instinct raises $1B Series C at a $10B valuationTechCrunch · September 28, 2026
- 174.NIQ and Similarweb Advance Agentic Commerce Measurement for the AI Shopping Eranielseniq.com · September 2, 2026
- 175.5 reasons why I'm so skeptical of the entire AI tracking industryLinkedIn · Rand Fishkin · September 22, 2026
- 176.Reddit citations in ChatGPT fell off a cliff.Amanda Natividad · August 26, 2026
- 177.Gartner Predicts 70% of Enterprises Will Abandon Agentic AI Built by Vendor Forward-Deployed Engineering by 2028Gartner · September 29, 2026
- 178.When a Smarter Model Stops MatteringTheory Ventures · September 29, 2026
- 179.Lessons from building 113 AI agents and reinventing a GTM function with AIGTMnow · September 29, 2026
- 180.My post-conference playbookAdam Schoenfeld · September 24, 2026
- 181.SaaStr 879: Stripe's CRO of AI on The New AI GTM PlaybookSaaStr · September 23, 2026
- 182.
- 183.CRM As A Business World Model: The Future For GTM Teams?Topline · September 20, 2026
- 184.Inside 12 months of AI search experimentsGrowth Unhinged · September 16, 2026
- 185.What AI agents really think about your pricingGrowth Unhinged · September 2, 2026
- 186.Salesforce Includes AI Credits in New $195 and $395 Editions # Salesforce Includes AI Credits in New $195 and $395 Editions artificial intelligence saas salesforce subscription pricing Sep 04, 202subscriptioninsider.com · September 4, 2026
Methodology
This brief is the synthesis stage of a continuous research pipeline. The capture stage harvests raw signal across the GTM-AI source universe without filtering, ranking or summarizing. The synthesis stage runs once per month, applies a freshness audit and a GTM-relevance filter, then runs a multi-perspective pass over the surviving corpus to produce the read above. Captured breadth comes first, then ruthless filtering — you can always filter what you captured, but never filter what you missed.
This month's capture: 31,170 items (August 24 – September 30), ordered below from highest to lowest volume:
- Research — 12,056 items. arxiv across cs.CL, cs.AI, cs.LG, cs.IR, cs.MA and cs.HC, dated by each paper's first-version submission date, plus Google Research.
- Funding — 5,338 items. Funding wires and aggregators (Pulse 2.0, The SaaS News, Tech Funding News, EU-Startups, Traded), press wires, VC-firm pages and the YC Summer 2026 batch, plus 1,295 distinct rounds parsed out of funding newsletters, 432 of them judged GTM-relevant.
- Publications — 3,871 items. TechCrunch, The Register, The Verge, The Decoder, Axios, MIT Technology Review, VentureBeat, Forrester, WIRED and a long tail via windowed RSS and archive backfills, plus grounded web search.
- Platforms — 3,428 items. Show HN, Product Hunt, GitHub releases and trending, MCP registries, job boards and YC launches.
- Tool changelogs — 2,155 items. Releasebot's API across GTM and platform vendors, plus direct scrapes of 54 GTM vendors the feed doesn't carry.
- Thought leaders — 1,668 items. A multi-channel per-operator sweep (blogs, Substack archives, X timelines, YouTube, LinkedIn) across the named GTM roster, plus engagement-scored topic discovery on X.
- Communities — 740 items. Reddit and Hacker News, engagement-scored across 110 standing GTM topics.
- Newsletters — 642 items, plus 106 multi-round digests. The general and funding newsletter inboxes.
- Foundation labs — 567 items. Primary release surfaces for Anthropic, OpenAI, Google, Meta, xAI, DeepSeek, Mistral, Qwen and others.
- Seattle — 255 items. GeekWire, The Seattle Times, AI House, AI2, AI Tinkerers Seattle, Madrona and local vendor newsrooms.
- Podcasts — 194 items, meta-signals 89, public markets 72 (SEC filings and earnings transcripts), patterns 53, conferences 21 and benchmarks 21.
This run fixed ten failures that would each have read as a quiet month: arxiv's endpoint silently returning nothing to scripted requests; arxiv's date field now meaning latest revision, which had let revised old papers pass as new (and inflated last month's research count); a page cap that cut high-volume feeds short; truncation on transient errors; feeds that expose only their newest posts; a quick-mode cap and a rate-limit that made operators' X timelines look empty; a newsletter that moved platforms and looked silent; a YouTube feed endpoint that now returns errors for every channel; mis-filed lab release notes; and funding rounds counted up to three times across newsletters. Funding volume is not comparable with August's for this reason: the newsletter inboxes that were unavailable last month were captured this month.
Scratchpad cross-check. No hand-collected leads were kept this month, so the scratchpad was rebuilt from an independent monthly readout — Adam Schoenfeld's September report, published October 1 [context — October 1, 2026] — treated as a list of leads, not as a source of claims. Of 32 actionable leads, 27 resolved to in-window primary sources that were added to the corpus and dated from the primary page; five did not (two vendors with no in-window news, one post dated January, and two themes measured only by his own panel). His panel measurements are not reproduced here; where this brief and his reached the same conclusion — the convergence on "context" — his readout is cited as context only.
Declared gaps. X topic search failed on seven of 110 topics on every attempt while neighbouring topics succeeded, among them Gong Callisto and Sponsored Agents (operator timelines were fully covered). ZoomInfo's newsroom blocks automated access, so ZoomInfo items come through secondary coverage. Business Wire was reachable only through mirrors, and several vendor changelogs are login-gated or don't exist. LinkedIn coverage is partial, and six LinkedIn-first operators had no reachable in-window posts. OpenAI's site rate-limited the capture after about 50 pages, leaving some customer stories at summary length. Pavilion's GTM2026 conference was still running at window close. Podcasts carry show notes, not transcripts. Several figures in this brief come from a single secondary report and are marked as reported: the Listen Labs price, HubSpot's seats-and-credits model (from Analyst Day coverage rather than a HubSpot release) and the shelving of GPT-6.1 Astra.
Synthesis pipeline. A deterministic filter ranks captured items by source tier (T1 primary > T2 analyst > T3 community) and engagement, then clusters a shortlist by theme. A mandatory freshness audit verifies that every cited source carries a publication date inside the period from a mechanical source — an arxiv submission date, platform API timestamp, vendor changelog JSON, RSS publication date, or an on-page byline read by a headless browser. Dates inferred from search snippets are rejected, and every Sources entry here was checked against its corpus record by script before publication. The surviving set is filtered to GTM relevance and run through five lenses: Technical Builder, Strategic Operator, VC / Category Thinker, Contrarian Skeptic, and Historian against the prior brief. Items surfaced by two or more lenses are elevated. The historian pass found only one theme that rose steadily all month without a single event behind it — personal and buyer-side agents, whose share of weekly items grew from 2.5% to 11.6% — and one term with a clean birth date, Jev and decision models, absent before September 15. Salesforce and HubSpot volume spiked with Dreamforce and UNBOUND and is read here on substance, not item count. Against August's read, the price call held and the capture call did not: model prices fell, but value did not move to the layer that meters tokens; the seat vendors put the meter inside their own contracts. August's "answer-engine optimization has deflated" call reversed.
Why this matters for readers. Every claim above traces to a primary-source URL in the Sources list. Where the only evidence on a topic comes from the vendors selling into it — as with this month's context-quality data — the brief says so.