Intercom Fin vs Text: per-outcome or bundled AI pricing

TL;DR: Fin bills you per outcome, uncapped, on top of Intercom seats, and the rate stays $0.99 however far you scale. Text bundles resolutions into the plan, and the allowance grows much faster than the price — 10 on Essential, 200 on Growth, 2,000+ on Enterprise — so your effective cost per resolution falls as volume and plan tier rise. Top-ups past the allowance are a flat $0.99 either way. Which means the sticker price decides less than your storefront does. Fin for ecommerce is Shopify-only and requires Intercom, off Shopify you fall back to Fin for Service plus connector work, while Text runs on WooCommerce, BigCommerce and Adobe Commerce natively.

Fin wins on voice, on published analytics methodology, and on being the most-reviewed AI agent in the category. Text wins on storefront breadth, on one subscription instead of two, and on a per-resolution cost that gets cheaper as you grow. Fin is also now owned by Salesforce, which completed the acquisition in September 2026 — worth a clause in your contract.

Underneath, both vendors use the verb "resolve" with two different meanings. For Fin, a resolved conversation is one that ended. For Text, it's one that ended with a cart. Every criterion below inherits that split.

What changed in 2026: Intercom rebranded to Fin, Text went on the offensive

If you're comparing notes against an older article, a lot has moved on both sides.

Fin: a new name and a new owner

Intercom rebranded the whole company to Fin in May 2026, taking the name of its AI agent. The Intercom inbox still exists as the helpdesk product, but the company is now Fin.

Then, on June 15 2026, Salesforce signed a definitive agreement to acquire Fin for approximately $3.6 billion, and the deal closed on September 10 2026. Fin brought more than 30,000 companies with it, and now operates as part of Salesforce's AI Labs, with Salesforce framing the Fin AI Agent as the fast-to-value option beside Agentforce.

Text: selling agents, and no change of control

Text shipped the features this comparison turns on: integrations with ecommerce platforms, live chat, help desk, AI Agent, and AI Custom Skills, alongside a brand overhaul with a red visual identity and a rebuilt website. The framing was explicit. "Text is pivoting from a defensive to an offensive stance," said its VP of Brand.

Scale is comparable to Fin's. Text reports 30,000+ businesses against Fin's 30,000+ companies. It has been public on the Warsaw Stock Exchange since 2014.

Intercom Fin vs Text: the short version

The criteria this comparison uses

Nine criteria: core purpose, model architecture, automation depth, pricing model, billable unit, platform requirements, channels and languages, integrations, and reporting. Between them they decide the customer experience your store can actually deliver, and what your support team does all day.

Text is ours, which you should weigh accordingly. A comparison published by one of the two vendors only earns its keep if it tells you when to pick the other one, and there's a section below that does exactly that.

All figures below come from the two vendors' own public pages, retrieved 7 September 2026, and pricing in this category moves — check before you sign. Every performance figure is first-party unless labeled otherwise, so treat it as directional and measure your own customer conversations instead. Both vendors are selling an AI agent to the same support teams, and both describe the customer experience they produce in their own terms.

How Fin works: an AI powered resolution engine

The Fin AI Agent

Fin is a resolution engine. Fin AI reports a 76% average resolution rate across 12,000+ customers, built on multi-stage answer validation and a proprietary model. The Fin AI Agent answers questions from your help center articles and connected customer data. It answers questions a human would otherwise pick up, then bills when no more help is requested. For your support team, the appeal is volume absorbed before a human sees it, and an AI agent handles thousands of concurrent conversations simultaneously, which is the whole structural argument at Black Friday scale.

Fin's own numbers go further than the average: many customers see over 85%, and Fin reports 2 million weekly resolutions and growing. Independent write-ups are more conservative, with one noting that real-world case studies land between 42% and 50% of common inquiries resolved autonomously. Fin leverages advanced large language models underneath, so the gap between those figures is about your content, not the model.

For complex queries, Fin executes a configured sequence of steps it calls a Procedure, which automates complex processes with multiple actions such as refunds and order cancellations. A Procedure that ends by passing the conversation to a human is a billable procedure handoff. Fin also handles lead qualification, priced as a premium outcome.

Two features matter more for ecommerce than the marketing leads with. Data connectors pull your customer data in so answers are personalized rather than generic, and complex tasks can touch real records. And Fin Vision reads images — "screenshots, invoices, and error messages" — which in retail means a shopper can photograph the damaged item instead of describing it.

Fin also provides rich media in its replies, with text, links, forms and applications rather than plain messages. It claims set up in under an hour on any helpdesk and works across content in Intercom, Zendesk, and Salesforce.

Two integration counts get quoted as one, so separate them. Fin's own library lists 51 connectors across helpdesks, APIs, MCP and AI platforms. The 450+ apps and integrations figure belongs to Intercom's app store. The first is what Fin can reach. The second is what the helpdesk underneath it can.

Fin for ecommerce is Shopify-only

This is the criterion that ends the evaluation for many merchants, and it rarely appears in comparison tables.

Fin's own FAQ answers it without hedging. Asked whether Fin for ecommerce only works with Shopify, Fin says that it is purpose-built for Shopify and is only available to Shopify-powered businesses. It is not available for WooCommerce, Magento, BigCommerce, or custom-built storefronts. It also requires Intercom and is not available as a standalone product.

The same answer gives the alternative, and that part matters: if you're on another platform, Fin for Service works with any helpdesk, including Intercom, Zendesk, and Salesforce, and connects to any platform with an API via no-code data connectors.

So state the constraint precisely. A WooCommerce merchant can't buy Fin for ecommerce, but can run Fin for Service and wire it to the store anyway — Shopify, Stripe and Linear all appear among Fin's data connectors, and Fin offers to deploy on any helpdesk over API with no migration. What you give up is the packaged commerce product: catalog-aware recommendations, browse-to-checkout, live in minutes. Not Fin itself.

Where it applies, Fin for ecommerce is strong: it syncs pricing, availability and variant information from the Shopify catalog and handles returns, exchanges and order updates in one customer conversation. Those commerce features are the reason Shopify merchants shortlist Fin in the first place.

How Text works

The AI Agent

Text is built around the opposite goal. Its headline metrics are $1.5M in revenue after 8 months and +25% average order value, alongside an 80% AI resolution rate.

The Text AI Agent issues discount codes, recommends products, handles order status, offers extended warranties, qualifies leads and detects abuse. All configurable features rather than fixed behavior — customer support and selling inside one conversation. Since the May 2026 release you create these as AI Custom Skills with plain when/then conditions, so the rules read like a support macro.

Text integrates with Shopify, WooCommerce, BigCommerce, Adobe Commerce, Webflow, Swell, Medusa and WordPress, and needs no separate helpdesk subscription because the helpdesk is part of the whole platform.

Proactive support

The mechanism matters more than the metric. A reactive AI agent, whoever built it, waits until the shopper asks, by which point a sizing doubt has usually already become an abandoned cart.

Text reads customer intent in real time: pages they visit, products they compare, moments they hesitate, then opens the chat itself. Its reporting answers what the customer conversation was worth — revenue closed, carts recovered, leads captured — rather than tickets closed.

What else Intercom offers beyond support

Worth noting if you're weighing the whole platform rather than the AI agent alone.

Intercom offers more than a helpdesk. Product tours, in-app messages and outbound messaging sit in the same product, so Intercom conversations can be onboarding and activation as well as customer support. If your customer operations team already runs product tours or in app messages, that's real consolidation you'd be giving up by moving.

Text is narrower by design: an AI powered live chat, ticketing and AI agent stack aimed at ecommerce, not a lifecycle-messaging suite. It's an AI chatbot and a help desk that answers questions and sells, not a product-adoption tool. Which way that cuts depends on whether onboarding is part of your job.

Pricing: $0.99 per outcome vs per seat

What Fin bills for

Four events trigger a charge.

Fin's own research team draws the distinction plainly: confirmed resolutions are more valuable than assumed ones because they come with clear, reliable signals, and warns that if a bad experience is misread as a good one, Fin might abandon the user prematurely, leaving them without help.

You're charged once per conversation however many customer questions it contains, which is genuinely generous on long threads. Intercom seats run $19 per seat per month at entry, rising to $85 and $132 on higher tiers, where Lite seats come bundled. Add-ons sit on top of that: Copilot for your support team, Proactive Support Plus, and the Pro add-on are all separate lines, and it's the stacking of add-ons rather than the headline $0.99 that surprises people.

What Text bills for

The allowances are where the model differs from Fin's. Moving Essential to Growth costs 4.2× more per seat but includes 20× more resolutions, so the effective cost per included resolution drops from roughly $1.90 to about $0.40. Enterprise bundles 2,000+ resolutions and 50 AI Agents. Your cost per resolved conversation goes down as your AI Agent handles a larger share of cases, unlike a meter, which charges the same rate at every volume.

Two honest caveats. Top-ups don't tier down: exceed your allowance and it's a flat 50-resolution refill at $49.50, $0.99 each, on both paid plans. And the falling effective rate is a function of the bundle, not a volume discount you negotiate. The way to reach it is a higher plan, which also means paying for seats you may not need.

Text also meters agent-side AI writing help. 5 reply suggestions monthly on Essential, 50 on Growth — so a large team on the entry plan hits that ceiling first.

Using Fin with an existing helpdesk

You don't have to move to Intercom to run Fin. Fin for Platforms connects to Intercom, Salesforce, HubSpot and Freshworks, with Zendesk, so Fin works alongside your existing helpdesk and human agents rather than replacing them.

Two caveats. Running Fin against a non-Intercom helpdesk carries a 50-outcome monthly minimum. And the packaged commerce behavior, the Shopify catalog sync, stays Shopify-and-Intercom only.

Text takes the other route: one platform, shared inbox included, no underlying helpdesk to license. Either way, check that a human handoff arrives with full context attached, because an agent who has to ask the customer to start again undoes whatever the AI saved.

Channels: Fin Voice and the phone gap

Fin covers voice, live chat, email, Slack and social across its platform. Text runs website live chat, email, Facebook Messenger, Whatsapp and SMS into one team inbox, all in one view. Phone is the real difference.

Fin Voice answers customer calls and resolves issues end-to-end without IVR menus, reporting a 24.5% lift in AI resolution and 0.43 seconds off time-to-first-audio. It runs in 30 languages, connects to your systems to verify identities, process refunds and book appointments, and reaches your lines through Twilio, Genesys, RingCentral, Five9, 8x8, Aircall and Mitel. Fin Voice is offered on a custom pricing plan. It's currently available to select customers working directly with our sales team.

If phone is a core channel for your customer support, that gap decides the comparison. Text's live chat and email coverage is deep, but its phone coverage is absent. On languages the two are level: Fin supports 45, Text 46 — though Fin is explicit that it works in those languages where you have the corresponding support content, so an English-only help center makes that count a capability rather than a result.

SMS, and why the channel is not the intelligence

Both platforms send SMS, and it's worth separating the channel from what runs on it.

Plain text messaging is a basic communication channel for short messages, effective for quick updates and personal communication, which in retail means shipping notifications and delivery windows. It's limited to plain text, so none of the rich media, links, forms or applications an AI agent can put in a chat reply survive the trip, and it requires strict adherence to opt-in compliance rules. In the US that means express written consent, easy opt-out, and quiet hours under the TCPA.

The channel's reputation rests on reach. SMS is widely quoted at a near-98% open rate, and it is genuinely high, but read it the way you should read a resolution rate, because the figure measures device-level message expansion rather than active reading.

The part that matters here: traditional text messaging relies on human agents or basic automated scripts for responses, and lacks built-in intelligence unless connected to external systems. Pointing an AI agent at the channel is what changes it from a broadcast tool into a conversation. Both Fin and Text do that, neither gets credit for SMS itself.

Knowledge base and data connectors

Both AI agents learn from the same raw material: your help center articles plus whatever customer data you connect through data connectors. Neither asks you to build a knowledge base from scratch.

Fin lets teams configure workflows, knowledge, data connections, guardrails and tone without vendor dependence, and reports resolution rate improving 1% every month. Text builds its knowledge base from your website, docs, FAQs and policies, and holds one communication style across every language rather than a bot per market.

The binding constraint on accurate answers is your content, not the model. One independent test measured roughly 38% resolution over 60 days purely because sparse help content caps what any AI agent can resolve. Write down the values that decide edge cases such as return window, restocking fee, which SKUs are final sale, or both tools will guess plausibly and wrongly.

What it means for your support team

Buying an AI agent is a change-management project, not just a license.

With Fin, your support team keeps working in the Intercom inbox or its existing helpdesk while Fin handles the front line. Agents get Fin's answers as context, and the team's job shifts toward the complex questions the AI escalates. Because Fin bills per outcome, the team also acquires a habit nobody enjoys: watching a meter. A support team that resolves more conversations costs more, which is a strange incentive to hand people.

With Text, the shared inbox, the live chat widget and the AI Agent are one product, so there's a single set of features to learn and one place where customer conversations land. The trade is a smaller ecosystem and a shorter track record.

Either way, decide who owns the AI Agent before you buy. The best predictor of a good customer experience twelve months in isn't which vendor you picked but whether one named person on the support team is responsible for the help center content, the escalation rules, and reading what customers actually asked last week.

Analytics: what your transcripts tell you

Fin publishes more about its analytics than we do, and that's worth saying plainly on our own blog.

Support conversations are hard to analyze: Fin's research team describes them as "short, noisy, and informal", full of typos and emojis, which is why classical topic modeling fails on them. Their published method clusters message embeddings with HDBSCAN, then uses a generative model to write "clear, human-readable labels", reporting assignment accuracy above 85%.

What that buys a merchant is theme discovery you didn't ask for — emerging topics such as early bugs surfacing before they reach your reviews. Fin's conversation analytics puts the same machinery behind contact drivers, sentiment and frustration signals, and resolution quality, turning qualitative conversations into structured, actionable data.

Fin also scores the customer experience directly. Its Customer Experience (CX) Score measures the quality of every conversation with 5x more coverage than CSAT, which is a genuinely better instrument than a survey nobody fills in — and it sits alongside resolution rate and involvement rate on the same performance dashboard. Fin learns from every interaction to continuously improve, which is the mechanism behind that 1% monthly figure.

Text isn't absent here, but its own reports documentation is built around operational metrics rather than topic-level analysis: AI resolution rate, CSAT split into automated, manual, and assisted, and response and resolution time. There's no topic-clustering report and no tagging-as-analysis feature documented there, and no named instrument for conversation quality to set against CX Score. Sales reporting sits on a separate track: the pricing page lists an AI Agent sales report and sales reports tied to chats as plan features, which is Text's closest answer in ambition to CX Score, but it's revenue and conversion reporting, not topic or sentiment analysis.

The reporting also points the same raw material at a different question: what the conversation was worth. That's the number a CFO asks for. So the honest split is that Fin tells you more about why customers are writing in, and Text tells you more about what it earned you.

Free trial and how to test both

Both offer a 14-day free trial with no credit card required. Fin's includes unlimited outcomes for two weeks.

Pull twenty real customer conversations from your queue and run the same twenty through both. Include a return outside policy, a sizing question needing the catalog, an angry customer, and one in a second language: the customer questions your team dreads, not the easy ones. Then measure four things:

• Confirmed resolutions, not assumed. Count only where the customer said it was solved.

• Cost per confirmed resolution. Real spend divided by that count.

• Human handoff quality. How often the AI agent escalated at the right moment, in both directions.

• Revenue touched. Whether the conversation moved an order.

• Confidently wrong. How many it answered wrongly while sounding certain. Neither vendor's resolution rate will show you this, and it's the count that should decide it: a high resolution rate on wrong answers is worse than a lower one on right answers.

Accept a null result if you get one. If both resolve fourteen of twenty, the decision falls back to storefront, channels and pricing shape, a legitimate outcome, reached in two weeks instead of a year.

Where Fin is the better choice

A comparison from one of the two vendors is only worth reading if it tells you when to buy the competitor. Four situations where Fin is the right answer.

You're on Shopify and staying there

Fin for ecommerce syncs pricing, availability and variant data straight from the Shopify catalog. Nothing in Text's integration layer is that tightly coupled to one platform, and tight coupling is an advantage when it's your platform. If Shopify is your whole storefront, Fin's biggest limitation costs you nothing.

Phone is a real channel for you

Fin Voice resolves calls end-to-end without IVR menus, in 30 languages, through seven named telephony providers. Text has no voice product. This isn't a gap we can argue around — if your customers phone you, the comparison ends here.

You need to know what customers keep asking

Fin's published topic-modeling work and its CX Score are ahead of what we offer. Our own reporting is built around resolution rate, CSAT, and response times rather than topic-level analysis; Fin ships a research paper and a named conversation-quality score. If your problem is "I don't know what's driving my ticket volume" rather than "I can't prove support pays for itself," Fin answers the question you actually have.

Where Text pulls ahead

Your storefront is Shopify and more

Fin for ecommerce is not available for WooCommerce, Magento, BigCommerce, or custom-built storefronts. Text integrates with all of those natively, each with its own connector in a few clicks rather than a build.

You want one subscription, not two

Fin sits on a helpdesk you license separately. Intercom seats from $19, or your existing helpdesk plus a 50-outcome monthly minimum. Text ships the helpdesk, the live chat and the AI Agent as one product, so there's one bill, one login and one set of features to learn.

You'd rather not watch a meter

Fin's per resolution billing has no ceiling, which is why reviewers report spend at 2–3× their first estimate, and the $0.99 rate is the same at 200 resolutions as at 20,000. Text's are bundled, and the bundle scales faster than the price — 2,000+ on Enterprise — so the effective cost per resolution falls as you grow rather than holding flat. A Black Friday spike doesn't arrive as an invoice.

You're measuring revenue, not deflection

Text reports revenue closed, carts recovered and leads captured by default, and the AI Agent is built to recommend products and issue discount codes mid-conversation. If the case you're making internally is about revenue rather than cost per ticket, that's the shorter argument.

Frequently asked questions

Is Intercom now called Fin?

Yes. Intercom rebranded the company to Fin in May 2026, after its AI agent. The Intercom inbox continues as the helpdesk product, and Intercom pricing still governs seat costs, so you'll see both names in play.

What does Intercom Fin do?

Fin is an AI agent for customer support. It answers customer questions from your help center articles and connected data, takes actions on your systems, and hands off to human agents when a conversation needs judgment. Fin delivers across chat, email, phone and social.

How much does Fin AI cost?

$0.99 per outcome, plus Intercom seats from $19. Lead qualification is $9.99. There's no cap, so cost tracks conversation volume.

Is Intercom Fin good?

By review count, yes. It's the most reviewed AI agent on G2, and reviewers note strong accurate answers on routine customer requests. The recurring complaint is that Intercom pricing is hard to forecast once seats, outcomes and add-ons stack.

Can I use Fin with my existing helpdesk?

Yes, via Fin for Platforms, which supports Salesforce, HubSpot, Freshworks and Zendesk among others. A 50-outcome monthly minimum applies off Intercom.

Does Fin work with WooCommerce?

Fin for Service does, connected to your store through an API data connector. Fin for ecommerce does not. Fin states it's Shopify-only and requires Intercom.