|
GTM stack intelligence, enriched.
|
|
Salesforce agreed to buy Fin, the company formerly known as Intercom, for roughly $3.6 billion, putting a packaged customer-service agent inside Agentforce. A day later HubSpot shipped Revenue Hub, pulling quoting, contracts, billing, and payments into the CRM where the customer data already sits.
Underneath those two, the smaller vendors kept moving in a different direction, opening themselves up rather than walling themselves in. Apollo is the clearest example this week, and it’s the deep dive. My read on what these moves have in common is at the bottom.
|
|
|
Salesforce — Signs Agreement to Acquire Fin, Formerly Intercom, for ~$3.6 Billion
Salesforce announced on June 15th that it has signed a definitive agreement to acquire Fin, formerly Intercom, for approximately $3.6 billion. Fin’s core product is an AI Agent that resolves customer queries across live chat, email, WhatsApp, SMS, phone, and Slack. Per the announcement, the agent runs on Fin’s proprietary model, Apex, which the company describes as purpose-built for customer support, and Salesforce cites examples of it resolving on average 76% of support volume end-to-end. The deal brings more than 30,000 customers and is expected to close in the fourth quarter of Salesforce’s fiscal 2027.
The acquisition is interesting from two angles; acquiring a specialized model versus building, and the number of acquisitions Salesforce is completing. Agentforce is the customizable, build-it-yourself platform. Fin is a packaged agent you can point at a support queue and turn on. Salesforce is buying the fast-to-value path; a specialized and proprietary model which is a different kind of asset than the frontier models its agents rent from OpenAI and Anthropic. The model capabilities from frontier labs have raised questions about where value will accrue moving forward. The application layer is no longer able to rely on moats, except where there is a specialized model trained on proprietary data. Salesforce is betting that Fin fits the bill.
The Fin acquisition is just one in a series of deals announced by Salesforce over the past year including Momentum, Contentful, m3ter, and Informatica. This level of M&A activity may help you spin up capabilities faster than building in-house, but it introduces execution risk. How does Salesforce thoughtfully integrate these acquisitions naturally into the platform? More interestingly, how do these acquisitions play within the headless CRM world that Salesforce has sought to develop? If the CRM becomes headless, Salesforce hasn’t yet put forward a compelling answer to where the orchestration layer sits across the core platform and its acquisition capabilities.
Source: Salesforce Newsroom — Fin acquisition, June 15, 2026
|
|
|
HubSpot — Revenue Hub Goes Live: Quote-to-Cash Moves Into the CRM
HubSpot launched Revenue Hub on June 16th, available the same day. It’s the evolution of Commerce Hub, and it brings quoting, contracts, billing, and payments into HubSpot rather than leaving payment as a separate end step in a separate tool. Reps build quotes from a deal record using the Breeze Assistant, buyers receive an interactive quote they can review, sign, and pay in one place, and a Closing Agent answers buyer questions without a rep stepping in. Billing runs off the same record and payments collect automatically through HubSpot Payments or a connected Stripe account.
The pitch from HubSpot’s CPTO is that revenue data “completes Growth Context,” the company’s framing for the business knowledge its AI runs on. One of the hardest aspects of building a revenue model in a usage-based business is connecting pricing strategy, quote generation, revenue data, and renewals. When these motions have integrated workflows and data sources the business can generate the revenue flywheel. Revenue Hub allows customers to easily understand and query their contracts 24/7. Customer success, Finance, and AI agents all have visibility into the commercial terms and live revenue data for each customer contract. If implemented correctly, Revenue Hub can be a meaningful step towards integrating those motions and creating the flywheel.
Many RevOps teams spend significant effort building custom CPQ capabilities and tying those workflows together across systems. If you run a separate CPQ or billing tool alongside HubSpot today, this is a consolidation question to put on the table at your next renewal.
Source: HubSpot Company News — Revenue Hub, June 16, 2026
|
|
A closer look at one smaller stack vendor each issue.
|
Apollo — Turning the Prospecting Engine Into Something Agents Can Call
Apollo sits in the crowded prospecting category, the data-plus-sequencing layer where it competes with ZoomInfo on contact data, Clay on enrichment workflows, and Outreach on execution. To win, Apollo is making the same bet as Unify, which we covered in last week’s issue: the company is making its engine callable by AI agents rather than only usable inside its own interface.
On June 3rd, Apollo launched a connector for Perplexity, letting users run core GTM workflows in natural language without leaving the assistant: searching for prospects, enriching people and companies, creating or updating records, building sequences and enrolling contacts, and analyzing team performance. Apollo’s MCP now spans Claude, ChatGPT, and Perplexity, and the June update added guided workflows and guardrails for high-impact actions like sequence enrollment, bulk enrichment, and bulk list updates. Plus a new sequence skill that generates a sequence from a prompt, an ICP, or company context. This is all bolstered by the existing functionality allowing teams to pull Gong call recordings directly into Apollo so AI summaries work across the full call history.
Most tools in this category assume they are the place you work: you log into the app, run your searches, and the data lives there. Apollo is following Unify and betting a real share of its users would rather treat it as a service their agent calls, with guardrails so an agent can’t bulk-enroll a thousand contacts by accident. That reframes the buying question for a RevOps team. Do you want prospecting and enrichment to be infrastructure your agents reach into, or another destination app your reps tab over to? We’re seeing more and more vendors bet on the former, and Enriched’s own bet is that number will continue to grow.
Sources: Apollo Release Notes 2026 · Apollo press release, June 3, 2026
|
|
The rest of the week, in brief:
|
|
|
|
Salesforce Summer ’26 Release Ships With an Inbound Qualification Agent: Salesforce’s Summer ’26 release became available June 15th. The piece most relevant to GTM teams is the new Customer Engagement Agent, which holds two-way conversations across websites and email to qualify buyers around the clock and hand warm leads to sales. The release also adds Multi-Agent Orchestration and a Tableau MCP integration that lets agents query Tableau’s analytics engine directly. If you own inbound routing, this is a vendor-native alternative to the chatbot-plus-SDR handoff you may be paying a separate tool for.
Source: Salesforce Newsroom — Summer ’26 release, June 15, 2026
|
|
Microsoft Puts GPT-5.5 Into Copilot Chat: Microsoft made GPT-5.5 Instant and GPT-5.5 Thinking available in Copilot Chat, the assistant layer that increasingly underpins Copilot for Sales. The practical point for RevOps is that the model behind a tool many reps already use shifted underneath them, with the Thinking variant aimed at multi-step reasoning. Worth a note in your AI vendor review for the same reason last issue’s model story was: you rarely get to choose, or even see, which model your vendor routes you to.
Source: Microsoft 365 Copilot release notes, June 2026
|
|
|
👁️ Noticed
Salesforce is paying roughly $3.6 billion for an agent whose selling point is a proprietary model that, per Salesforce’s own announcement, outperforms the frontier models Salesforce also pays to run. Owning the model is apparently worth more than renting the best one.
|
|
|
|
|
The platforms have stopped building the agentic stack feature by feature and started absorbing it. For a year the megavendors shipped agents the way they ship everything, one release at a time. This week Salesforce wrote a $3.6 billion check for a finished customer-service agent and a model to run it, and HubSpot welded quote-to-cash directly into the CRM so billing, contracts, and payments live next to the customer record.
What ties them together is Enriched’s first lens, data gravity. Both moves pull a function that used to live in a separate tool into the platform that already holds customer data because the agent is only as good as the context it can see. HubSpot is explicit about this, its whole pitch is that revenue data “completes” the context its AI runs on. Salesforce is making the same bet from the service side. This matters because every function you let migrate into the platform is a function whose data now lives where the vendor’s platform, not you, controls the exits. It brings us straight to Enriched’s sixth lens, avoiding vendor lock-in, and it’s the one GTM teams are trained to be suspicious of, because the tool that holds your integrations holds your leverage at renewal.
Apollo is in this issue next to two companies a hundred times its size. The smaller vendors are reading the same board and playing the opposite move. While the platforms consolidate functions inward, Apollo spent June making its engine callable from the outside, by Claude, by ChatGPT, by Perplexity, with guardrails so an agent can operate it safely. Salesforce’s own Agent Fabric is a governance layer for exactly the multi-vendor agent sprawl that an open approach produces. So the stack is forking. One path is a few platforms that own the data and the agents and sell you the convenience of having it all in one place. The other is an open layer where your data stays yours and your agents call best-of-breed tools as services. Most teams will end up with some of both, which is precisely why the choice has to be deliberate rather than accumulated.
This means it is time for an ownership audit. Take your top ten tools and sort each one into a column: destination, a place your people log in and your data lives, or service, something your stack and your agents can call and your data passes through. When the agent is the primary user of your stack, who controls the context, you or your vendor’s platform? The teams that can answer that have a headstart on choosing their architecture. Teams that can’t might have their architecture chosen for them one acquisition at a time.
See you next week. — Andrew
|
|
|
|
|
|
