Chatbase vs Intercom Fin vs Anserra: which AI chatbot for customer support?

Three names come up almost every time someone shops for an AI customer-support chatbot in 2026: Chatbase, Intercom Fin, and — increasingly — Anserra. They solve overlapping problems from different angles. This is the honest version of the comparison.

I’ll cover pricing, retrieval, embedding, brand control, and the parts vendors don’t write on their landing pages.

The one-line version

• Chatbase — cheapest per chatbot, weakest source attribution, easiest to spin up. Good for docs-as-a-service and marketing sites.
• Intercom Fin — best if your team already lives in Intercom’s inbox. Priced per resolution, which scales with success and rewards teams that already automate ticketing.
• Anserra — grounded answers with mandatory citations, one workspace across many chatbots, honest about limits. Best when auditability matters (regulated verticals, agencies with multiple clients, teams that resell white-labelled bots).

Pricing (as of Q3 2026)

Numbers rounded. Vendors change prices — check on the day.

Retrieval quality

The bit vendors don’t compare. All three do embedding-based search. Only some do hybrid retrieval, and it matters — customers paste exact part numbers, error codes and account IDs into support widgets, and pure semantic search blurs those strings into vague neighbours.

• Chatbase — semantic only in 2026. Excellent on paraphrase, weak on entity match.
• Intercom Fin — semantic + keyword blend; details opaque.
• Anserra — semantic + full-text + trigram, fused by reciprocal rank. Configurable minimum relevance threshold so weak matches are refused rather than guessed at.

Source attribution

Every answer to “Do you offer refunds?” needs to be defensible. Vendors handle this differently:

• Chatbase can show a source, but it’s not shown by default in many widget themes.
• Intercom Fin shows the article it drew from; whether visitors see this is theme-dependent.
• Anserra ships citations on by default, and every answer that isn’t backed by a source is refused rather than fabricated.

If you’re in healthcare, finance, legal or education, this alone is likely to decide the choice for you.

Embedding and brand control

• Chatbase: one script tag, iframe or React component. White-labelling on paid plans.
• Intercom Fin: lives inside the Intercom Messenger. If your brand needs the widget to not look like Intercom, this is a hard sell.
• Anserra: script tag, React, Next.js, WordPress, Shopify, Webflow, and a full JS API. Domain-locking so the embed key can’t be lifted onto competitors’ sites.

Agencies and multi-brand teams

If you resell chatbots to clients, one workspace across many chatbots is the difference between a scalable business and a firefighting business.

• Chatbase: separate workspaces per chatbot; billing per chatbot.
• Fin: not designed for agencies.
• Anserra: one workspace, many chatbots, five roles, per-chatbot API keys. Built for it.

What each vendor hides

I’ll say it plainly:

• Chatbase — the free tier is thin. The moment traffic is real, you upgrade.
• Intercom Fin — the per-resolution model looks cheap in a pilot and scales fast. Model it against your actual ticket volume before signing.
• Anserra — the built-in model in the free plan is a deterministic mock. Real replies need OpenAI/Anthropic/Google/OpenRouter keys. We say so.

Which should you actually pick?

• You already run Intercom and pay for it → Fin. The integration savings dominate.
• You need a docs bot with citations and don’t care about deep integrations → Chatbase or Anserra. Anserra if honesty of answers matters more than time-to-launch.
• You resell white-label chatbots, or work in a regulated vertical → Anserra. This is where source attribution and workspace-level RBAC pay off.

Try Anserra free — no credit card, about five minutes to your first live chatbot. If it’s not the right fit, at least you’ve calibrated what to compare the others against.