What the Salesforce–Fin (Intercom) deal means for small businesses

If you run a small business on Intercom, two things happened this year that you should read before your next renewal. In May 2026 Intercom renamed itself Fin, after its AI agent. A month later it agreed to be acquired by Salesforce. Neither changes your bill today. Both change who the product is for.
Short version: Intercom was already priced for funded startups and enterprises, with seats from $29 to $132 per person per month and Fin billed at $0.99 per "outcome". Under Salesforce that direction will not reverse. If your support is a website chat, a shared inbox and a handful of tickets, this is the moment to look at tools built for that job at a flat price. Here is how SupportLime compares with Intercom.
What actually changed
Intercom spent the last two years turning Fin, its AI agent, into the centre of the company. The rename made that official: the product you buy is the AI agent, and the inbox, help centre and messaging are what surround it.
The Salesforce agreement was announced about a month later. Salesforce sells to large enterprises, and its history with acquired products (Slack, Tableau, MuleSoft) is to integrate them into its platform and price them for those customers. That is not a criticism; it is the business model. It just means the small-business tier is unlikely to get more attention than it does now.
How Intercom bills a small team today
Three meters, and the combination is what hurts.
Seats. Essential is $29 per seat per month on annual billing, Advanced $85, Expert $132. Every teammate who touches the inbox needs one.
Fin outcomes. $0.99 for every conversation Fin resolves. An outcome includes "assumed" resolutions, which is the polite term for a visitor who stopped replying. One Hacker News commenter put the arithmetic bluntly: a ticket that takes a human a minute to answer now costs $0.99, which works out at more than $60 an hour. When Fin is used standalone there is a 50-outcome monthly minimum.
Everything else. Product tours, email campaigns, phone and the app marketplace are add-ons or higher tiers. A small team pays for the platform whether it uses those or not.
For a two-person shop with modest volume, the seats alone are $58 to $264 a month before a single AI answer. That is the bill that the acquisition will not make smaller.
The per-outcome problem in one paragraph
Per-outcome billing aligns the vendor's revenue with the number of conversations marked resolved, not with your customer's experience. If the AI answers a question and the visitor leaves, that is an outcome. If it answers wrongly and the visitor leaves in disgust, that is also an outcome. The vendor has no incentive to hand off early, and every incentive to count generously. You cannot budget for it, because you do not control how many conversations the AI decides it has resolved.
Flat allowances of AI replies are the opposite: you know the number, a handoff costs nothing, and a reply that did not help costs the same as one that did, so the vendor has no reason to pad the count.
What to do if you are on Intercom
Check your renewal date and your meters. Pull three months of invoices and separate seats, Fin outcomes and add-ons. Most small teams find that seats are the bulk and Fin is growing.
List what you actually use. For most small businesses it is the website messenger, the inbox, maybe the help centre. Tours, campaigns and phone are usually either unused or replaceable with a tool you already pay for.
Trial the alternative with your old tool still running. Point a flat-price assistant at your public help centre, test it on your top twenty questions, install it on a staging page or a low-traffic page, and compare the answers. Keep Intercom until you are satisfied; the point is not to switch fast, it is to switch once.
Export what matters. Conversations and contacts can be exported from Intercom. Help-centre articles are public pages, so any crawler-based assistant can read them directly.
What a small team needs instead
The job is the same as it was before any acquisition: answer the repetitive questions from your own content, hand the rest to a human with the visitor's email attached, keep a list of what still needs doing. The things to insist on:
- One meter you can predict, ideally AI replies, with human replies never counted.
- No per-seat pricing, so the whole team can be in the inbox.
- An honest handoff that triggers on low confidence and on "talk to a person", not a bot that never lets go.
- Sources under every answer, so you can see where the AI got it from and fix the page.
- Tickets and notes without a separate help desk subscription.
SupportLime was built around that list. The Free plan covers one site and 100 AI replies a month with the inbox and tickets included; Starter is $29 a month with 1,500 replies and unlimited teammates. Nothing is billed per seat or per resolution, and the assistant pauses rather than guessing when it is out of answers. If you are weighing it against Fin specifically, the Fin alternative page has the side-by-side.
The honest caveat
If you use Intercom's tours, campaigns, phone and marketplace integrations and have the budget, stay. It is a good product for the customers it is built for, and Salesforce will keep investing in that direction. The point of this post is narrower: if you were only ever using the chat and the inbox, you have been paying for the rest, and the pricing is about to get less attention, not more.