Quick answer: If you want an intercom alternative for small business that answers from your own pages and starts on a free plan with a monthly message limit, we built AI BotKit for that job. Stay on Intercom, now called Fin, if you run a shared ticket inbox or need your bot to act inside customer accounts.
Intercom now goes by Fin, and Salesforce completed its acquisition on September 10. Its pricing still has two layers: seats for your team, plus $0.99 each time the AI agent delivers an outcome. That makes this a fair moment to ask what you actually pay for.
An AI chatbot trained on your own content answers visitor questions from the pages, posts, PDFs and links you give it, instead of from general model knowledge. We built an AI chatbot trained on your own site for readers who mainly need that part, and this post is honest about who does not.
Intercom is now Fin and owned by Salesforce. What changes for a small team?
Intercom completed its rebrand to Fin in May 2026, named after its AI agent. TechCrunch covered the June 15 deal announcement, and Salesforce published the completion notice.
What we know
Salesforce says it wants to use Fin’s team and technology to improve Agentforce, its AI agent platform. The pricing page still shows seats plus a per-outcome fee for the AI agent. We found no announced price change tied to the deal.
What nobody knows yet
The roadmap and future pricing under Salesforce are not public, so we won’t guess. Readers on the Hacker News thread about the deal split on it. One argued Intercom’s chat bubble and CRM can be bought for much less elsewhere, and that the AI agent differentiates less each year.
Others argued the opposite: that most businesses can’t build and tune a support agent themselves. Nothing about the deal forces you to move, but it is a natural time to price your options.
What does Intercom actually cost a team of three?
Intercom bills in two layers. Seats pay for your human team, and the AI agent is billed separately at $0.99 per outcome. Intercom’s help center lists the Essential plan at $29 per seat per month on annual billing, but other listings show different figures, so check the pricing page before you budget.
What counts as an outcome
An outcome is not only a customer saying the answer worked. Third-party pricing guides say it can also count when the customer goes quiet for 24 hours after Fin’s reply. One commenter on the same Hacker News thread called that a soft measure, so read Intercom’s outcome definitions before you forecast.
A worked example
Say your three-person team handles 40 chats a week, about 170 a month. If Fin delivers an outcome on 100 of them, that is $99 in Fin charges before seats. This is our arithmetic for illustration, not a quote.
A commenter on that thread put a small store at roughly $500 a month for 500 outcomes. The issue is not the rate itself but paying $0.99 each for questions your site already answers.
Startup discounts
Intercom runs an Early Stage program with discounts, and its help center mentions free Fin outcomes in year one for eligible startups. Check the eligibility rules before you rule Fin out on cost.
Where does Fin earn its price?
What reviewers praise
Fin earns its price when support is a team job. G2’s Fin page shows a 4.5 rating from 3,925 reviews, and 57.5% of reviewers are small businesses. Reviewers praise fast resolution, easy setup and training, and helpdesk integrations. See the full G2 page.
The company says Fin resolves 76% of support queries without a human, as Yahoo Finance reported. That is a vendor claim, not an independent figure.
Where reviewers push back
The recurring G2 complaint is inaccurate or incomplete answers to complex, multi-step questions. G2 also scores Fin’s ease of setup at 8.5 out of 10, below the 9.0 category average.
Price comes up in user comments on AlternativeTo too, where Intercom gets called expensive but feature-rich.
When the extra cost makes sense
One reader on the Hacker News thread, who says they have run Fin rollouts, named two strengths. Fin can act on account data, and its tooling shows unresolved questions and suggests help-article fixes. A simple custom build, they argued, falls over once actions are needed.
If your bot must refund, update or look up accounts, Fin’s extra cost buys something real.
What does AI BotKit do, and what is it not trying to replace?
At AI BotKit, we train a chatbot on your pages, posts, PDFs and external URLs, and it answers visitor questions from that content. Setup needs no code: a WordPress plugin, or an embed snippet for Shopify, Squarespace, Wix and custom HTML. Our documentation walks through the first setup.
How it answers
Grounding is the point and also the limit. A bot trained on your pages is only as good as those pages, so thin or out-of-date pages give thin or out-of-date answers. Our comparison with generic ChatGPT widgets explains the difference.
In AI BotKit, we capture leads and hand a chat to you when a question needs a person. Chats and leads hold personal details, so check where conversations are stored, who can read them and what you tell visitors. We can’t make compliance promises for your business; our privacy policy is the place to start.
What it doesn’t replace
We built our chatbot around answering questions and capturing leads, not around running a team’s shared ticket inbox. Intercom lists a shared inbox and ticketing on all three plans, so if your day runs on that, Intercom is the better fit. We are built for answering and lead capture, not for running a ticket queue.
If your bot must look up an order or change an account, test that exact task first. We have Playbooks for procedures like qualifying leads, bookings, quotes and order chasing, plus Actions, but we can’t promise parity with Fin’s procedures.
Handoff needs a plan too. A WordPress.org user on a different chatbot plugin asked what happens when a human is requested and nobody replies. Decide who takes over, and when, before you go live.
AI BotKit vs Intercom (Fin): how do they compare side by side?
Here is how the two compare on the points small teams ask about most. Intercom’s column reflects its public pricing and G2 data, and ours reflects what we ship today.
| Intercom (now Fin) | Ours: AI BotKit | |
|---|---|---|
| Built around | A customer service platform: helpdesk plus the Fin AI agent | An AI chatbot trained on your own content |
| Pricing model | Seats for your team, plus $0.99 per Fin outcome | Free plan with a monthly message limit; paid plans on our pricing page |
| Answers from | Your help content, with connected systems for actions | Your pages, posts, PDFs and external URLs |
| Setup | G2 ease-of-setup score of 8.5 out of 10 (category average 9.0) | No code: WordPress plugin or embed snippet |
| Works on | Web chat, email, WhatsApp, SMS, phone and Slack | WordPress, plus Shopify, Squarespace, Wix and custom HTML via embed |
| Human handoff | Yes, into its helpdesk | Yes, a chat can be handed to you |
| Team inbox and tickets | Included on every plan | Not our focus |
| Account actions | Yes, through configured procedures | Playbooks and Actions exist; test your own task first |
| Lead capture | Yes, Fin also handles sales conversations | Yes |
If the rows that matter to you are inbox and account actions, Intercom wins. If they are cost, answers grounded in your own content and quick setup, ours is worth a trial. Neither column is better for every business.
Should you switch, stay, or run both? Find your situation
Match the tool to the job: tickets and account actions point to Fin, while answering and lead capture point to a lighter chatbot.
You use Intercom as a full helpdesk
Keep the helpdesk. Test our chatbot on a copy of your site or one product area, and compare the questions it answers with what Fin resolves. If it handles the easy ones, you may pay for fewer outcomes.
You only use its chat widget and bot
This is the strongest case for a lighter tool. If you pay for seats mainly to keep a chat widget running, ours can start on a free plan with a monthly message limit and train on pages you already have. Check first that those pages answer your top questions.
You’re comparing for the first time
Start with the job, not the brand: answering FAQs, capturing leads or managing tickets. If you run an online store, our Tidio comparison covers another route. Write down your top 20 customer questions before you compare anything.
You build chat for client sites
Each client has different pages and different support needs, so test per site. Our chatbot for agencies guide covers lead generation on your own agency site.
A 6-question check: is a lighter chatbot enough for you?
Answer yes or no to each question about your own support.
- Do most chats ask the same 20 or so questions your website already answers?
- Does your team assign tickets between people from a shared inbox?
- Does the bot need to change something in a customer’s account, such as a refund?
- Is most of your Fin or Intercom bill driven by easy questions?
- Are your pricing, shipping, returns and FAQ pages current enough to train on?
- Is a person ready to take over when the bot hands off a chat?
What your answers suggest:
- Yes to 1, 4, 5 and 6, and no to 2 and 3: a lighter chatbot is likely enough. Run a two-week trial next to what you have.
- Yes to 2 or 3: stay on Intercom or Fin, or keep the helpdesk and test a chatbot separately for the easy questions.
- No to 5: fix your pages first. Any trained chatbot will repeat what they say.
- Mixed answers: trial first and decide with your own numbers.
How can you test an alternative without breaking your support?
A two-week trial tells you more than any comparison, including this one.
- List your 20 most common questions from email, chats and calls.
- Check that your pages answer them, and update the ones that don’t.
- Train the chatbot on those pages and add the PDFs you send customers.
- Test all 20 questions yourself, then ask a colleague to try odd phrasings.
- Set when a chat should reach a person, and who gets it.
- Run it for two weeks and note which questions it missed.
Judge any chatbot on your own 20 questions, not on a demo. Our list of questions to ask before choosing a chatbot helps you judge any option, ours included.
Frequently asked questions
Is there a cheaper alternative to Intercom for small businesses?
It depends on what you replace. If you only need a website chatbot, a lighter tool such as ours starts on a free plan with a monthly message limit. If you need a helpdesk with tickets and a team inbox, compare helpdesks on your real seat count and chat volume instead.
What counts as a Fin outcome?
Intercom bills $0.99 per outcome, and its pricing page lists resolutions and completed procedures, including handoffs. Third-party guides say a resolution can be assumed when the customer goes quiet for 24 hours. Check Intercom’s current definitions before you forecast a bill.
Can I keep Fin and drop Intercom’s helpdesk?
Yes. Intercom says you can buy Fin without an Intercom plan if you use a helpdesk such as HubSpot, Freshdesk or Salesforce. Its help center adds that Fin for Sales and Fin for Ecommerce are not available that way, so you need an Intercom plan for those.
Will an AI chatbot make up answers?
Answering from your own pages lowers the risk but does not remove it. If a page is thin or out of date, the answer will be too. Even Fin draws accuracy complaints on complex questions in G2 reviews, so test your own questions and keep a human handoff.
Does our chatbot work outside WordPress?
Yes. We offer a WordPress plugin and an embed snippet that works on Shopify, Squarespace, Wix and custom HTML sites.
If most of your Intercom or Fin bill goes to questions your website already answers, run a two-week trial and count what changes. We built AI BotKit for exactly that test, and you can start on our free plan with no card needed. If your day runs on tickets and account actions, stay where you are.