Salesforce Closes Fin and Puts Agents Into the Seat Price

Applications
Five days before Dreamforce, Salesforce closed the Fin purchase agreed at about $3.6 billion, finished Contentful, named an Enterprise AI Harness, and folded Slack, Tableau Next, and Agentforce credits into the seat price.
By Shashi Bellamkonda · September 10, 2026
$3.6B
Fin agreement value, June 15 (Salesforce)
76%
Fin claimed average resolution rate
$195 / $395 / $550
Core, Advanced, Max list per user month
FY28
unified Harness experience planned start
Customers I talk to already know this platform. They also already built workarounds for the pieces Salesforce did not sell. This week those pieces showed up as two closed deals, a named control layer, and a new seat map.

Salesforce customers I speak with are conflicted. They can run sales, service, and marketing on the platform without a map. They also spent years wiring content systems, chat agents, and analytics around gaps the catalog left open. Open interfaces plus two closed acquisitions now pull some of that work back onto the same bill.

I am not in San Francisco this week. Dreamforce runs September 15 to 17. Analyst friends still treat it as the largest vendor conference on the calendar, and Salesforce still comes up in almost every stack conversation I have. The product notes arrived before the keynote, so the desk is enough.

Fin closed on September 10 and stays in AI Labs

Salesforce said today it completed the purchase of Fin, the company formerly named Intercom. The June 15 agreement put the price at approximately $3.6 billion, subject to customary adjustments. The close note repeats a customer base of more than 30,000 companies and a claimed 76 percent average end-to-end resolution rate across live chat, email, WhatsApp, short message service texts, voice, and Slack. Fin stays in Salesforce AI Labs and keeps serving those accounts. Salesforce presents Fin as the fast-to-deploy agent that can sit on help desks a company already runs, next to Agentforce as the customizable path (Salesforce, 2026).

I wrote the agreement on June 15 as Salesforce buying the customer service model that was beating its own artificial intelligence on resolution claims. The close does not change that reading. You now decide which path a given desk should take.

Contentful closed September 1 as the Headless 360 content layer

An editor’s note on Salesforce’s June 1 announcement records the Contentful close on September 1. The agreement described a composable content platform used by more than 4,800 brands. Salesforce did not publish a close price. Dealroom, citing The Information at announcement, put the range near $1 billion to $1.5 billion, under Contentful’s 2021 Series F valuation. The stated job is a native content layer for Headless 360, so agents can assemble web, email, and commerce copy from customer data instead of a separate web stack (Salesforce, 2026; Dealroom, 2026).

Marketing and digital teams kept Contentful, or a rival, because Salesforce did not own a serious headless content system. Ownership changed on September 1. The integration work is still ahead.

The Harness names six capabilities. The unified screen is planned for early fiscal 2028.

On September 10 Salesforce published the Trusted Enterprise AI Harness. The six named pieces are Trusted Context, Trusted Agency, Trusted Action, Trusted Governance, Trusted Security, and Trusted Models. A new AI Control Plane is meant to register agents, set identity and policy, watch behavior, and track cost across Salesforce software and third-party agents. Rohan Kumar, president and chief platform and engineering officer, said differentiation would come from proprietary customer context turned into action, rather than from which model a company picks this quarter (Salesforce, 2026).

Many of the foundation technologies exist now. New capabilities and the unified experience are planned to begin rolling out in early fiscal 2028. Salesforce fiscal years start on February 1, so that window opens in calendar 2027. Pricing and upgrade paths for the Harness itself are marked for later announcements. Headless access is listed through Model Context Protocol servers, application programming interfaces, Skills, and plug-ins, including Claude, Slack, Microsoft Teams, and Agentforce.

That list is the openness customers asked for. It also gives the identity and policy team more surfaces to govern.

Core, Advanced, and Max put Agentforce on the invoice you already pay

On September 3 Salesforce introduced Core, Advanced, and Max editions for Agentforce Sales, Agentforce Service, and Agentforce Industries. The company presents them as replacements for the old Enterprise, Unlimited, and Agentforce 1 packaging. Every tier lists Slack, Tableau Next analytics, Premier Success, data security controls, and role-level Agentforce features. Flex Credits land at 500,000 on Core, 1 million on Advanced, and 2.75 million on Max. List prices are $195, $395, and $550 per user per month. Anirban Ghoshal at CIO, working from the Salesforce post, put the prior Enterprise list at $175 and Unlimited at $350. Max holds the Agentforce 1 price. Agentforce 1 customers can move to Max at no extra seat cost. Salesforce says Max carries about 60 percent more packaged value than Agentforce 1. Existing contracts on legacy editions stay at current price until they change (Salesforce, 2026; Ghoshal, 2026).

Agentforce used to sit on a separate line. Slack and Tableau often sat on other lines. The new map puts collaboration, analytics, credits, and agents on the same seat.

Open interfaces and two closed deals pull work back onto one bill. The new editions put the agent on that bill.

What I would watch after the keynote

Does Fin keep its 76 percent resolution claim on Salesforce-owned traffic, with a method a customer can audit, or does the number stay a vendor average from the close note?

When does Contentful stop being a separate console and start writing into the same customer record Agentforce reads?

Which Harness controls a customer can turn on in fiscal 2027, and which wait for the early fiscal 2028 experience?

CIO/CTO Viability Question

Pull last year’s Salesforce invoice and list every tool your teams bought because the platform did not include it: content, chat resolution, analytics, extra agent credits. Ask which of those lines the new editions and the two closes actually retire, and which still need a contract outside Salesforce after Dreamforce.

Sources

Salesforce. "Salesforce Introduces the Trusted Enterprise AI Harness." Salesforce News, 10 Sept. 2026, www.salesforce.com/news/stories/enterprise-ai-harness/.

Salesforce. "Salesforce Completes Acquisition of Fin." Salesforce News, 10 Sept. 2026, www.salesforce.com/news/press-releases/2026/09/10/salesforce-completes-acquisition-of-fin/.

Salesforce. "Salesforce Signs Definitive Agreement to Acquire Fin." Salesforce News, 15 June 2026, www.salesforce.com/news/press-releases/2026/06/15/salesforce-signs-definitive-agreement-to-acquire-fin/.

Salesforce. "Salesforce Signs Definitive Agreement to Acquire Contentful." Salesforce News, 1 June 2026, www.salesforce.com/news/stories/salesforce-signs-definitive-agreement-to-acquire-contentful/. Editor’s note on the page records the September 1, 2026 close.

Salesforce. "New Salesforce Editions Bundle Everything Businesses Need for Agentic Transformation." Salesforce News, 3 Sept. 2026, www.salesforce.com/news/stories/salesforce-simplifies-editions-2026/.

Ghoshal, Anirban. "Salesforce Offers More Agentforce Credits to Drive Adoption." CIO, 4 Sept. 2026, www.cio.com/article/4218804/salesforce-offers-more-agentforce-credits-to-drive-adoption.html.

Dealroom. "Salesforce Acquires Contentful, the Berlin-Born Headless CMS." Dealroom, June 2026, app.dealroom.co/news/note/salesforce-acquires-contentful-the-berlin-born-headless-cms.

Bellamkonda, Shashi. "Salesforce Buys the Customer Service Model That Was Beating Its Own AI." shashi.co, 15 June 2026, www.shashi.co/2026/06/salesforce-buys-customer-service-model.html.
Disclaimer: This blog reflects my personal views only. Content does not represent the views of my employer, Info-Tech Research Group. AI tools may have been used for brevity, structure, or research support. Please independently verify any information before relying on it.