Intercom Pricing 2026: What You'll Actually Pay (Including Fin AI Add-Ons)

Damien Mulhall
Co-Founder, Strategy and Content

Published · Updated

41 min read
pricing guide intercom helpdesk software
Intercom pricing 2026 breakdown: Hay.chat mascot Bale as the confused math lady meme, surrounded by floating Fin AI cost calculations

Updated 21 October 2026: Intercom Fin AI agent pricing in 2026 is $0.99 per outcome, which is what Intercom used to bill as a resolution (Fin pricing page, as of October 2026). On 12 May 2026, Intercom renamed the company Fin, after its AI agent. The helpdesk software is still called Intercom. Salesforce completed its acquisition of Fin on 10 September 2026, so pricing may change again. Prices below reflect the current pricing pages.

Intercom pricing in 2026 starts at "$29/month" as of November 2026. You'll find that number right on their pricing page, and technically it's correct. It's also the price almost nobody ends up paying.

The $29 Essential plan is really for solo founders doing their own support. One inbox. No workflow automation. No lite seats for your developer who just needs to peek at technical tickets. The moment you hire a second support person and need to route conversations between them, you've outgrown it.

Then there's Fin AI Agent, and this is where the maths gets interesting. Fin charges $0.99 per outcome, which mostly means per resolution. Sounds cheap for a single ticket. But resolve 2,000 conversations with AI each month and you're looking at $1,980 just for the AI, before platform fees. That's nearly $24,000 a year. Most buyers expect AI to shrink their support costs. Fin often grows them.

If you're new to Intercom, or you're wondering why your forecast keeps climbing, this is the maths the sales call tends to skip.

Before You Start: What Nobody Tells First-Time Buyers#

A few things experienced Intercom buyers know but rarely bother explaining, before we get to the tiers.

The Trial Defaults to Everything#

A long buffet table piled with cakes, fruit and covered dishes, with a single small plate holding one item at the far end, showing how an Intercom trial gives you every feature and the plan you buy gives you far less
The trial lays out the whole buffet. Your plan is the side plate.

When you start an Intercom trial, you get access to every feature across all three tiers. Plus Proactive Support Plus. Plus Fin AI Agent. Plus Copilot. You'll build workflows using Advanced-tier automation, configure reports that require Expert, and integrate channels with features you'll lose on day 15.

This is by design.

So decide your realistic tier before the trial starts, and test inside it. Otherwise you'll spend your first month after buying rebuilding everything with less. We've seen teams burn two full weeks on this.

The features you lose matter more than the tier names. Drop to Essential and you lose the workflow automation builder (you're stuck with templates). Multiple team inboxes (everyone shares one queue). Round-robin assignment (you hand out conversations manually). The private help centre (public only). And every lite seat. So if you built routing rules in the trial, split billing from technical, or gave your product team read access... all of that just stops working when you downgrade.

The Workspace Architecture Decision You'll Make Once#

In week one, Intercom asks you to set up your workspace. Team inboxes. Routing rules. Permissions. Get these wrong and they're painful to unwind later.

Most first-time buyers create too few inboxes ("we'll just use one for now") or structure them around current team org charts. Six months later, when you need to route billing questions separately from product bugs, you're rebuilding routing logic while handling live conversations. Not fun.

Build inboxes around conversation types, not teams. "Billing", "Technical Support" and "Account Management" survive reorgs. "Sarah's Team" and "EMEA Support" break the day Sarah leaves or you redraw the regions. Thirty minutes of inbox planning saves thirty hours of migration pain later. Genuinely.

Full Seats vs. Lite Seats: The Distinction That Saves Thousands#

Intercom charges full price for "full seats", the teammates who actually handle customer conversations. But plenty of people only need a look in. Supervisors reviewing escalations. Product managers tracking feedback. Developers checking technical context.

"Lite seats" can view conversations they're linked to or @mentioned in, and add private internal notes. They can't respond to customers, reassign tickets or change conversation fields. Advanced plans include 20 lite seats. Expert includes 50. Essential offers none.

A 12-person company might only need 6 full seats at $85 each and can use lite seats for the rest. That's $510 saved monthly versus paying full price for everyone. Over $6,000 a year.

The catch we keep seeing: lite seats slowly turn into full seats. The product manager who can't tag conversations for feature requests. The supervisor who can't reassign a stuck ticket. The exec who can't reply to a VIP directly. Each one becomes a case for upgrading. So budget for 20-30% of your lite seats becoming full seats in year one. The savings are real. So is the creep.

Channel Costs: The Usage-Based Surprise#

Your plan covers the platform. It doesn't cover usage on some channels:

  • SMS: In the US, $0.06 per outbound segment for the first 500, dropping in tiers to $0.01 at high volume. Rates vary by region

  • WhatsApp Business: Charged per message, not per conversation. In the US, $0.10 per message for the first 500, falling to $0.07 at higher volume

  • Email campaigns: $0.00025-$0.045 per bulk email sent. Inbox emails (responses to customers) are free

  • Phone: Pay-as-you-go for phone numbers, call minutes and recording

WhatsApp campaigns cost more than teams expect. Say you run a Black Friday campaign that sends proactive WhatsApp messages to 5,000 customers. At Intercom's per-message rates, that's roughly $400-440 just to send, before anyone replies. Teams that budgeted on the old per-conversation rates (a few cents each) get a shock. So segment hard. Send only to customers likely to act. And test a small batch before the full send.

Implementation Timeline: Reality Check#

Basic setup really is quick. You can be handling conversations within hours. Being properly ready takes longer:

  • Basic setup (Messenger, inbox, simple routing): 1-3 days

  • Full configuration (workflows, automations, help centre, macros): 2-4 weeks

  • Fin AI training (knowledge base optimisation, testing, iteration): 2-4 weeks

  • Agent training and adoption: 2-4 weeks (expect a 20-30% productivity dip during the transition)

  • Performance optimisation: Ongoing, typically 2-3 months before things stabilise

Fin AI training is where most rollouts stall. Fin is only as good as your help centre, and most help centres are written for humans browsing, not AI retrieving. Getting it ready can mean 40-60 hours of rewriting. Answers moved into the first paragraph, not the fourth. Pronouns like "it" and "this" swapped out, because they lose meaning once they're pulled out of context. Trigger phrases added ("to get a refund", "refund process", "request refund") that match how customers actually ask. Standalone answers that don't need three other articles read first. It's not light editing. It's rebuilding content for a different reader.

Intercom Pricing Tiers Explained#

Intercom has three core plans. Prices are per seat, and monthly billing costs 5-34% more than annual, depending on the plan.

Plan Monthly Annual Lite Seats
Essential $39/seat $29/seat None
Advanced $99/seat $85/seat 20 included
Expert $139/seat $132/seat 50 included

Essential ($29-39/seat)#

Built for solo operators and very early startups. You get the basics. A shared inbox. Intercom Messenger. A basic help centre. Prebuilt reports. Access to Fin AI (charged separately). What you don't get: lite seats, the workflow automation builder, multiple team inboxes or round-robin assignment.

Essential works when one person handles all support. It breaks with your second hire. You can't auto-assign conversations, can't create separate queues, can't give your developer read-only access to technical tickets. At that point the $56 a seat you save over Advanced is a false economy.

"Prebuilt reports" on Essential means Intercom's standard dashboards and nothing else. No custom reports. No filtering by custom attributes. No granular exports. So when leadership asks "what's our first-response time for enterprise customers?", you can't answer on Essential. You'd have to tag and track it by hand. Or upgrade.

Advanced ($85-99/seat)#

Most growing companies land here. Advanced adds workflow automation for rule-based routing, multiple team inboxes, round-robin assignment, a private multilingual help centre, and those 20 lite seats for the rest of the business.

The jump from Essential is steep, roughly 3x the price. But you get a lot for it. If you're serious about scaling support, this is usually the minimum tier that works.

The workflow builder is the real reason to be on Advanced. But nobody warns you about the learning curve. Workflows are powerful, and fragile. A workflow that routes "billing" to your billing queue will also grab the customer asking about a "billing address for their shipping label". "Cancel" catches both "cancel my subscription" (retention-critical) and "cancel that last message" (trivial). Plan for 2-3 rounds on every workflow as real traffic throws up edge cases. Test on historical conversations before going live. And never, ever ship complex routing changes on a Friday.

Expert ($132-139/seat)#

Enterprise requirements live here. SSO and identity management for security compliance. HIPAA-eligible infrastructure for healthcare. SLA management with breach alerts. Multibrand messenger for companies with distinct product lines. And 50 lite seats.

The jump from Advanced is smaller (roughly 55%), and a lot of what you're paying for is compliance rather than capability. If you don't need the security and compliance stuff, Advanced probably does the job.

"Multibrand" sounds like a nice-to-have. Until you acquire a company or launch a second product. Without Expert, you can't run separate messengers with different branding from one Intercom workspace. You'd need entirely separate Intercom accounts. Separate billing. Separate reporting. Separate agent training. No single view of customers who use both products. If M&A or a second product is anywhere on your 18-month roadmap, factor Expert pricing in now. I've seen teams scramble on this and it's never pretty.

Intercom Fin AI Agent Pricing in 2026: The $0.99 Question#

Fin AI Agent costs $0.99 per outcome as of October 2026, according to Fin's pricing page. An outcome is mostly what Intercom used to call a resolution, so per resolution the price is still $0.99. Used standalone, there's a 50-outcome monthly minimum. Inside the Intercom helpdesk, seats from $29 a month come on top.

Fin AI Agent is Intercom's headline feature, an AI that resolves customer conversations without a human. Plenty of AI chatbots just match keywords to help articles. Fin holds a real multi-turn conversation. It keeps track of context ("what about the blue one?" after you've been discussing products), copes with ambiguity without dead-ending, and mostly knows when to hand over instead of confidently making something up.

Fin's self-reported average is a 76% resolution rate across more than 12,000 customers, though it swings a lot by use case. Simple FAQ-heavy operations might see 85%+. Complex B2B support with technical edge cases might sit closer to 35%.

The catch is the pricing model. $0.99 per outcome.

Why Fin's AI Quality Commands a Premium#

Before you judge the $0.99, it's worth knowing what's behind it. Intercom has been handling customer conversations since 2011, and Fin's models are trained on hundreds of millions of real support interactions. What customers actually ask. How they word their frustration. Which replies fix things and which ones just get a follow-up. That data shows up in the product, and it's hard to copy.

Where it shows up most: edge cases. When a customer writes "I guess that works," cheaper AI often counts that as resolved. Fin recognises the passive-aggressive tone and offers to connect them with a human. When a customer asks "can I return this?" then immediately follows with "actually, what's the warranty situation?" Fin maintains both threads. Simpler AI tends to lose the first question.

Whether the quality difference justifies the price depends entirely on your specific conversations. More on that below.

What Counts as an Outcome (And What Doesn't)#

Since March 2026, Fin bills per outcome rather than per resolution. Resolutions are the main one, but Procedure handoffs (where a Procedure ends by passing the customer to a human), disqualifications and self-serve routings also cost $0.99 each, and qualified leads cost $9.99. Only one outcome is billed per conversation.

A resolution is counted when, following Fin's last answer, the customer either confirms satisfaction or exits without requesting further help. The important details:

  • If the customer asks follow-up questions or requests a human, it's NOT a resolution (no charge), unless a Procedure ends in that handoff, which is billed as an outcome

  • You're charged once per conversation, even if Fin resolves multiple questions within it

  • If a customer returns to a "resolved" conversation and needs more help, the resolution gets reversed and refunded

  • Fin responding to a greeting doesn't count as a resolution unless it actually answers a question

Fin answers. What happens next decides the bill.
The customer confirms it helped, or leaves without asking for more Billed $0.99
The customer asks a follow-up or asks for a human Not billed
A Procedure ends by handing the customer to a human Billed $0.99
The customer comes back to a resolved conversation for more help Refunded
One outcome is billed per conversation. Greetings alone don't count. Qualified leads cost $9.99.

The distinction between "confirmed" and "assumed" resolution matters way more than most buyers realise. A confirmed resolution means the customer explicitly said the answer helped. An assumed resolution means the customer left without asking for more help. But that also includes the frustrated customer who gave up and called your phone line instead. Typical split: 30-40% confirmed, 60-70% assumed. Your 60% resolution rate might include 20-25% of customers who simply abandoned the chat. Pull the breakdown in your Fin reporting and examine the assumed resolutions manually. How many are genuine successes versus silent failures?

The Maths at Scale#

Per-outcome pricing sounds reasonable for one ticket. Zoom out and it looks different:

Monthly Outcomes Monthly Cost Annual Cost
100 $99 $1,188
500 $495 $5,940
1,000 $990 $11,880
2,000 $1,980 $23,760
5,000 $4,950 $59,400

For context, a mid-sized e-commerce company with 4,000 conversations a month and a 50% AI resolution rate would pay $1,980 a month for Fin alone. That's $23,760 a year. More than some companies spend on their entire support stack.

When Per-Resolution Pricing Works in Your Favour#

Per-resolution isn't inherently bad. It's the right model when:

  • Your human ticket cost is high. If your fully-loaded cost per human-handled ticket (salary, benefits, tools, management, office) is over $15-20, then $0.99 per AI resolution is genuinely good value. Enterprise support teams with specialised agents often see $25-40 per ticket. Fin at $0.99 saves real money there.

  • AI handles a minority of your volume. If only 20-30% of your tickets are AI-resolvable, per-resolution means you're only paying for value actually delivered. A flat-fee model would have you subsidising capabilities you barely use.

  • Conversation quality affects revenue. For enterprise accounts or high-value B2B relationships, a botched AI conversation damages the relationship. A frustrated customer who gets generic AI replies might churn, and that costs thousands. Fin spotting frustration and handing over in time protects revenue directly.

When Per-Resolution Pricing Hurts#

The model becomes a problem when:

  • You run lean operations with low human ticket costs. If your agents efficiently handle tickets at $5-8 effective cost, Fin at $0.99 isn't saving money. It's adding a cost layer for marginal time savings.

  • You have high AI-resolvable volume. If 70%+ of your tickets are routine queries AI could handle, per-resolution quickly becomes your largest line item. At 5,000 monthly resolutions, you're at $60K/year for AI alone.

  • Your tickets need actions, not answers. If most tickets are "process my refund" or "change my subscription" rather than "how does X work?", Fin can handle them through Procedures and data connectors, but your team builds and maintains those connections, and each one still costs $0.99.

Action-Taking AI: A Different Approach#

Fin started as an answering engine. It surfaces help articles, explains policies and points customers to information, and it's really good at conversation: it keeps context and copes well with ambiguity. It can take actions now too, through Procedures that call your systems via APIs, data connectors and MCP, but you set those connections up yourself.

Action-first AI works the other way round, with the actions ready-made. Instead of telling a customer their refund status, it processes the refund. Instead of explaining how to update account details, it updates them. The AI connects directly to business systems (Shopify, WooCommerce, CRMs) and executes the resolution rather than describing it.

This matters for pricing because it changes what "resolution" actually means. With Fin, resolution often means the customer received accurate information and left satisfied (or at least didn't ask for a human). With action-taking AI, resolution means the transaction completed. Refund issued. Order tracked. Subscription changed.

Hay.chat is one of these. Its Starter plan includes 2,000 resolutions for €135/month flat. The same 2,000 on Fin would cost $1,980. So you're paying a bundled rate for things getting done, rather than paying per conversation for conversational polish.

The trade-off is real, and I'd rather be upfront about it. When a customer combines a complaint with sarcasm and a buried question ("Oh great, so my order is delayed AGAIN. I suppose you'll tell me it's 'on the way' like last time. Any chance I can actually get a refund or is that too much to ask?"), Fin parses the actual request and picks up on the frustration. Simpler action-taking AI might miss the subtext. Whether that matters depends on your ticket mix.

Setting Usage Limits#

A circuit breaker panel with one switch flipped off and a small spark settling, representing a hard usage limit that switches Fin off before AI costs run away
A hard cap is your circuit breaker. Fin switches off, and the bill stops climbing.

Intercom lets you set usage alerts and hard limits for Fin in Settings > Subscription > Usage, on the Fin outcomes card. You can get notifications at specific thresholds and set a hard cap that switches Fin off for the rest of the billing period once reached.

Set a hard limit on day one. Even if it's generous. Without a limit, a traffic spike (product launch, viral moment, service outage) can generate thousands of Fin resolutions in hours. One buyer reported a $3,200 surprise bill from a single day when their product went viral on Reddit and Fin "helpfully" resolved 3,200 "how do I sign up" conversations before anyone noticed. Set the limit at 150% of your expected monthly volume so it catches anomalies without blocking legitimate usage.

Fin Standalone: Using Fin Without Intercom#

Fin AI Agent is also sold separately from the Intercom helpdesk, for teams on Zendesk, Salesforce, HubSpot, Freshworks, Front, Gorgias and other helpdesks. Pricing is $0.99 per outcome with a minimum of 50 outcomes a month ($49.50), and no seat costs. Fin is now owned by Salesforce, which is worth factoring into any multi-year decision.

Fin Standalone makes sense in two cases. Fin's conversations are clearly better than your current platform's AI, especially for complex multi-turn chats where context and tone matter. Or you want to test Fin before committing to a full Intercom migration.

The operational trade-off: you're now running two vendor relationships and two billing cycles. When something breaks at the integration boundary, neither vendor owns the problem. Factor in that support overhead, not just the licensing maths.

Add-Ons That Add Up#

Then there are the add-ons. Several of them can push the bill a lot higher.

Intercom's add-ons in one place.
Add-onPriceWhat you get
AI Copilot$29/seat/month annual, $35 monthlyUnlimited AI help for human agents (10 conversations a month come with every full seat)
Proactive Support Plus$99/month, 500 messages includedProduct Tours, Surveys, In-app Posts, mobile push
ProFrom $100/month, 6,000 credits included, then $0.01 eachOperator, an AI agent for your support ops team, plus conversation analysis
Premier OnboardingOne-time fee, contact salesAn onboarding specialist for complex setups
Premier SupportMonthly fee, contact salesFaster responses, priority escalations, video chat

AI Copilot ($29-35/seat/month)#

Copilot is Intercom's AI assistant for human agents. Fin handles conversations on its own. Copilot sits next to your agents instead. It surfaces help articles as they type, suggests replies from the conversation, summarises long threads and drafts responses for them to check.

Every full seat gets limited Copilot. Ten conversations a month. Unlimited use costs $29/seat/month billed annually ($35 monthly). For a 10-agent team that's an extra $290 a month, or $3,480 a year.

Copilot adoption follows a predictable curve that catches managers off guard. Month one: agents ignore it because they "know the answers already." Month two: a few agents (usually the newer ones) start using it for complex edge cases. Month three: those agents visibly handle 15-20% more tickets at equal quality. Month four: other agents notice and adoption spreads. Budget for three months before measuring ROI. And track adoption by agent. Your senior agents often resist longest because they've memorised responses, while newer agents benefit most because they haven't.

Proactive Support Plus ($99/month)#

This add-on adds outbound features. Product Tours for in-app walkthroughs. Surveys. In-app Posts for announcements. Mobile push notifications. The flat $99/month includes 500 messages, and anything over that is charged by usage.

If you're using Intercom purely for reactive support, you might not need this. If you're building proactive onboarding flows or feature adoption campaigns, it becomes essential pretty quickly.

Product Tours sound optional until you look at your ticket data. That "how do I do X" question appearing 200 times monthly? A well-timed product tour triggered on first feature access eliminates roughly 80% of those tickets. But nobody budgets for Proactive Support Plus until after they spot the pattern. Run this analysis before signing: what percentage of your tickets could be prevented by in-app guidance at the right moment? If it's over 15%, include Proactive Support Plus in your initial contract. Adding it later tends to cost more.

Pro ($100+/month)#

Intercom's newer Pro add-on starts at $100/month. It bundles Operator, an AI agent for your support operations team, with conversation analysis and usage credits (6,000 included, extra credits at $0.01 each). It's aimed at teams running Fin at scale who want help monitoring and improving it.

Premier Services#

  • Premier Onboarding: One-time fee for an onboarding specialist. Pricing varies by complexity (contact sales for a quote). Useful for setups with multiple inboxes, complex routing or a big data migration.

  • Premier Support: Monthly fee for faster response times, priority escalations and video chat access. Contact sales for pricing. Usually worth it if Intercom going down costs you revenue directly.

Real-World Cost Scenarios#

Pricing tables only get you so far. These are what costs look like in practice, assuming annual billing.

The four scenarios side by side, per month on annual billing.
1: Startup2: Scale-up3: Mid-market4: AI-heavy
Team and plan3 agents, Essential8 agents, Advanced25 agents, Expert8 agents, Advanced
Platform$87$680$3,300$680
Fin resolutions301504002,000
Fin cost$29.70$148.50$396$1,980
Messaging~$20~$150~$800~$150
Monthly total~$137~$979~$4,496~$2,810
Annual total~$1,644~$11,748~$53,952~$33,720
Fin's share of the bill22%15%9%70%

Scenario 1: Early-Stage Startup#

Profile: 3 agents, Essential plan, light AI usage, minimal outbound messaging

  • Platform: $87/month ($29 × 3 seats)

  • Fin AI (30 resolutions): $29.70

  • Minimal SMS (a few hundred messages): ~$20

  • Total: ~$137/month (~$1,644/year)

A team this size may qualify for the Early Stage programme (see Negotiation Tactics), which cuts year one sharply and includes free Fin outcomes.

Scenario 2: Growing Scale-Up#

Profile: 8 agents, Advanced plan, moderate AI usage, regular outbound campaigns

  • Platform: $680/month ($85 × 8 seats)

  • Fin AI (150 resolutions): $148.50

  • Outbound messaging (SMS/email): ~$150

  • Total: ~$979/month (~$11,748/year)

Scenario 3: Mid-Market Company#

Profile: 25 agents, Expert plan, moderate AI usage, significant outbound volume

  • Platform: $3,300/month ($132 × 25 seats)

  • Fin AI (400 resolutions): $396

  • Outbound messaging: ~$800

  • Total: ~$4,496/month (~$53,952/year)

Scenario 4: AI-Heavy Operation#

Profile: Same 8-agent scale-up from Scenario 2, but with 2,000 monthly Fin resolutions

  • Platform: $680/month ($85 × 8 seats)

  • Fin AI (2,000 resolutions): $1,980

  • Outbound messaging: ~$150

  • Total: ~$2,810/month (~$33,720/year)

This one's worth sitting with. Same team as Scenario 2, but with higher AI volume. Fin costs now exceed their platform costs. AI went from 15% of the bill to 70%.

Same team, more AI: Fin goes from 15% of the bill to 70% Stacked bar chart of monthly cost for an 8-agent team on Intercom Advanced. With 150 Fin resolutions the bill is about $979: $680 platform seats, $150 messaging, $148.50 Fin. With 2,000 Fin resolutions it is about $2,810: $680 platform seats, $150 messaging, $1,980 Fin. Same team, more AI: Fin goes from 15% to 70% Platform seats Messaging Fin AI outcomes Scenario 2: 8 agents, 150 Fin resolutions Platform seats: $680 a month $680 Messaging: $150 a month Fin AI outcomes: $148.50 a month $979 Scenario 4: same team, 2,000 Fin resolutions Platform seats: $680 a month $680 Messaging: $150 a month Fin AI outcomes: $1,980 a month $1,980 $2,810
Monthly cost on annual billing. The platform bill stays at $680. Fin goes from $148.50 to $1,980.

Your scenario will drift. Costs in month one rarely match month twelve. Track three numbers every month: total conversations, Fin resolution rate and cost per resolution. Then build a simple model. If conversations grow 10% a month and the Fin rate holds steady, costs grow 10% a month too. And this is the counterintuitive bit. If you improve your knowledge base and the Fin rate jumps from 40% to 60%, costs jump 50% even with flat conversation volume. Model six months forward. That's where the surprises show up.

Cheaper Intercom Fin Alternatives for a Small Ecommerce Store#

If per-resolution pricing at $0.99 doesn't fit your economics, there are alternatives with different cost structures and capability trade-offs. None of them are perfect (nothing is), but depending on your ticket mix, one might make a lot more sense. We've put ten of them side by side, with prices checked on each vendor's site, in our Intercom Fin alternatives guide.

Per-Resolution AI on Top of Seats: Zendesk AI Agents#

Zendesk includes AI agents on every plan, with 5 automated resolutions per agent each month (10 on Professional). Beyond that, each costs $1.50 committed or $2.00 pay-as-you-go, on top of seat fees. Its $50/agent/month add-on is Copilot, which helps human agents rather than talking to customers. So Zendesk isn't a flat-fee escape from per-resolution pricing: at 2,000 AI resolutions a month, a 10-agent Professional team pays about $2,850 in resolution charges on top of $1,150 in seats. Our Zendesk pricing guide has the full breakdown.

The trade-off: you're buying into the whole Zendesk ecosystem. Its platform complexity. Its pricing tiers. Its roadmap. If you already use Zendesk, adding the AI is easy. If not, you're looking at a full platform migration, not just an AI swap. If cost is the main reason you're looking, our guide to free Zendesk alternatives shows what each free helpdesk plan actually includes.

Conversation-Based Pricing: Tidio Lyro#

Tidio prices AI based on conversation volume rather than resolution outcomes. Lyro is sold in packs, from 50 conversations a month ($39, or $32.50 billed annually) up to 1,000, and above that you talk to sales. You know your maximum cost regardless of resolution rate.

Tidio's strength is e-commerce. Deep Shopify integration, a pre-built order status template, and setup that takes hours rather than weeks. The AI handles single-turn transactional queries well ("where's my order") but doesn't match Fin's sophistication for multi-turn conversations or nuanced edge cases. A customer who changes their question mid-conversation or combines multiple issues in one message can exceed what Tidio handles comfortably.

Best fit: Shopify stores with high volume of repetitive, single-question queries and limited need for conversational nuance.

Action-Taking AI at Lower Price Points: Hay.chat#

Hay.chat comes at it from the other end. Fin is strongest at understanding questions and writing good answers. Hay.chat connects to your business systems and carries out the action: processing a refund through Shopify, tracking an order, updating a contact record in HubSpot. When a customer asks for a refund that fits your policy, the agent prepares it and sends it to your team, because refunds wait for your approval by default.

Here's 2,000 AI resolutions a month on each, as of November 2026:

Tool What each one bills you for 2,000 a month
Fin (Intercom) Each outcome, plus Intercom seats $1,980 plus seats
Zendesk AI agents Each automated resolution past the per-agent allowance, plus agent seats $2,850 for 10 Professional agents at $1.50 each, plus seats
Hay.chat Starter A flat monthly plan with 2,000 resolutions included, no seats €135

Hay.chat counts a resolution when the AI resolved the conversation by following your rules, and any action it sent for your approval counts as one, whatever you decide. Chats where the shopper got no real answer, left unhappy or went to your team unresolved are free.

The full side-by-side is in Hay.chat vs Fin, and Hay.chat pricing has the other plans. We run the same sums across more tools in AI customer service pricing compared.

What Hay.chat does well: transactional tickets where the customer wants something done. "Process my refund" gets prepared and sent to your team to approve. "Update my address" gets updated. It's also source-available, so you can read the code. Handy when procurement starts asking about AI transparency.

What Hay.chat doesn't do: Hay.chat is a newer entrant without Intercom's 15 years of refinement or its track record at enterprise scale. The 450+ app integrations that Intercom offers aren't there (though it does connect to Shopify, WooCommerce, Magento, Wix, Stripe and HubSpot). Complex conversational flows where customers combine complaints, questions, and sarcasm in ambiguous ways are Fin's strength, not Hay.chat's primary focus. If you're choosing between them, sample your tickets. Count how many are transactional actions versus nuanced conversations. That ratio basically makes the decision for you.

Setting up Hay.chat requires connecting it to your business systems. Hay.chat operates as a layer on existing infrastructure rather than a complete platform. For human agents, it hands off inside its own dashboard, or inside Chatwoot when it runs there.

Best fit: e-commerce and SaaS companies where most tickets are transactional (refunds, order updates, address changes) rather than tricky problem-solving. Teams already on Shopify, WooCommerce or similar, where ready-made actions pay off straight away. And anyone feeling real cost pressure from Fin's $0.99 per outcome, now or at projected volume.

Platform vs. Layer: Which Approach Fits You?#

Intercom is a complete customer service platform. Ticketing, inbox, messenger, help centre, reporting, 450+ app integrations, AI. All in one system. Alternatives like Hay.chat and Fin Standalone operate as AI layers on top of your existing infrastructure.

All-in-one platform
AI (Fin)
Inbox and ticketing
Messenger and help centre
Reporting
450+ app integrations
One vendor, one contract. Example: Intercom.
AI layer on your existing stack
AI layerAnswers customers and takes actions
Your existing helpdeskInbox, ticketing and reporting for human agents
Your business systemsStore, billing and CRM the AI connects to
Best tool for each job. Examples: Hay.chat, Fin Standalone.

That breadth matters in practical ways. Your CRM probably already connects. Your analytics tools probably already connect. There's probably a pre-built workflow for your use case. With AI layers, you're building (or buying separately) each of those integration points.

Some companies prefer having one vendor accountable for everything. One contract, one support relationship, one place to look when something breaks. Other companies prefer picking the best tool for each function and assembling a custom stack. If you've worked in operations long enough, you probably already know which camp you're in.

Run this calculation: what percentage of your tickets actually need conversational AI quality? If 80% are transactional ("process my refund," "track my order," "cancel my subscription") and only 20% need nuanced conversation handling, you might be overpaying for sophistication you rarely use. Action-taking AI like Hay.chat at €135/month for 2,000 resolutions handles that 80% while costing about 92% less. Route the complex 20% to human agents. The maths often works better than paying premium rates across the board.

Negotiation Tactics#

Intercom's list prices aren't necessarily your prices.

Volume Discounts#

Meaningful discounts grow with seat count. Vendr's data puts typical savings around 18%, and buyers running competitive evaluations often get 15-30% off.

The negotiation lever nobody mentions: competitive quotes. Before your Intercom call, get written quotes from Zendesk and one other alternative at your scale. You don't need to prefer them. You need the documentation. When Intercom's rep says they can't discount further, producing a competitive quote changes the conversation entirely. "I have a written offer from Zendesk at $X" signals you've done the work and have real options. Sales teams have discretionary discount authority they only use when losing the deal is the alternative.

Timing Your Negotiation#

Fin belongs to Salesforce now, and Salesforce's fiscal year ends on 31 January (per its annual filings with the SEC). That puts its Q4 in November, December and January, which is when reps are chasing annual numbers and usually have the most room to move. December is the sweet spot in our view: late enough to catch the year-end push, early enough that you're not competing with every other deal in the final week of January. The acquisition only closed in September 2026, though, so ask your rep straight out when their quarter ends. If Fin's sales team is still on an old calendar, their answer tells you where the leverage is.

For renewals, start negotiations 60-90 days before your contract expires. Not 30. The extended timeline signals you're seriously evaluating alternatives, not just going through the motions.

Your renewal is their retention problem. Intercom's sales team is measured on net revenue retention. Losing an existing customer hurts more than missing a new deal. If you really are considering leaving, communicate this at 90 days: "We've been evaluating alternatives and have identified a migration path. We need a reason to stay." That's different from "can we get a discount?" The first signals real risk. The second signals routine negotiation. Have your migration plan documented even if you hope you'll never use it.

Multi-Year Commitments#

Two- and three-year contracts typically save 10-25% versus annual pricing, and annual price increases (often 3-7%) can be capped at 2-3%. The trade-off: you're locked into the platform regardless of how your needs evolve.

Multi-year contracts should include three protective clauses. (1) Price cap: lock in a maximum annual increase at 3-5%. Without this, "locked in" can mean "locked in at whatever price they quote at renewal." (2) Seat flexibility: the right to reduce seats by 10-15% at each anniversary without penalty. Your team might shrink. (3) Exit clause: define the data export window and process if you don't renew. Surprises at contract end are expensive. Get all three in writing before signing.

Early Stage Programme#

For qualifying startups (up to $10M in funding, fewer than 15 employees, new to Intercom), the Early Stage programme takes 93% off in year one, 50% in year two and 25% in year three. You get 6 Advanced seats, 6 Copilot seats, 20 lite seats and Proactive Support Plus, plus up to 300 free Fin outcomes a month in year one (150 in year two, 75 in year three). Messaging usage is still charged.

How the Early Stage discount steps down. The bundle is 6 Advanced seats, 6 Copilot seats, 20 lite seats and Proactive Support Plus, with messaging charged on top.
YearDiscountFree Fin outcomes a monthRough cost
Year 193%Up to 300~$55/month
Year 250%150~$4,700/year
Year 325%75~$7,050/year
Year 4NoneNone~$9,400/year

The Early Stage discount shrinks every year, so plan for each jump. Year one starts from about $55/month. Year two is half of list, around $4,700 for the bundle, and full price (about $9,400 a year) arrives in year four. Budget for every step, not just year one. And don't build workflows during year one that depend on Expert-tier features you can't afford at full price. The worst outcome is building your entire support operation on capabilities you'll lose when the discount expires.

The Decision Framework#

When Intercom Makes Sense#

  • You need a complete platform: messenger, inbox, help centre, AI, all in one system

  • Messenger-first customer engagement aligns with your product experience (SaaS, mobile apps)

  • AI conversation quality directly affects revenue, because you're serving enterprise accounts or high-value B2B relationships where failed AI creates real churn risk

  • Your cost-per-human-ticket exceeds $15-20, making $0.99 per outcome good value

  • You value vendor consolidation and want one platform, one vendor, one support relationship

  • Your existing tools likely already integrate with Intercom's 450+ apps

  • You have budget for a premium solution and resources to configure it properly

When to Look Elsewhere#

  • Per-resolution costs at $0.99 exceed your unit economics, particularly at high volume

  • Most of your tickets need actions executed (refunds processed, orders tracked, subscriptions changed) rather than questions answered. Action-taking AI like Hay.chat bundles 2,000 resolutions for €135/month flat

  • You already have a helpdesk you're happy with and just need AI automation as a layer

  • Your support volume is projected to grow substantially, making per-resolution costs uncomfortable at scale

  • Your tickets are primarily simple, single-turn queries that don't need conversational complexity

  • Your budget is tight

Questions to Ask Before Committing#

  • What's our current cost-per-ticket for human-handled conversations? (If it's under $5, Fin may not save money)

  • What percentage of tickets could AI genuinely resolve, and do they need answers or actions?

  • How will costs scale if support volume doubles in 18 months?

  • Which add-ons will we realistically need within 12 months?

  • What price protection exists in our contract for renewal?

  • If we leave, what happens to our data, and how painful is migration?

Run this test before signing anything. Sample 50 random tickets from the past month. For each, answer two questions: could AI resolve this without human judgment? Does resolution require an action (process refund, update account) or just an answer? Tally the results. If 70%+ need actions and you're considering Fin at $0.99, you'll pay for every one of them, and your team still has to wire those actions up through Procedures and data connectors. An action-taking AI like Hay.chat at €135/month for 2,000 resolutions might deliver more value for your specific ticket mix.

What Happens If You Leave#

A huge moving box that won't fit through an open doorway, next to a neat stack of small boxes, showing how parts of an Intercom setup don't move with you when you leave
Contacts and help articles pack up neatly. Your workflows and Messenger setup are the big box.

Every platform decision includes an exit consideration. This is what leaving Intercom actually involves.

Data Export#

Intercom provides data export capabilities for conversations, contacts, and help centre content. Conversation data exports as CSV, but that covers metadata, not full message content. Transcripts come out as PDF or text in batches, or through the API. Help centre content exports in HTML format.

Export your data before you cancel. Not after. Once your subscription lapses, access becomes complicated. Some companies have reported weeks-long delays getting export access post-cancellation. Build exports into your migration timeline as step one: download everything while you still have full access. Set a calendar reminder for 30 days before your renewal date to begin exports, regardless of whether you're leaving.

Migration Complexity#

Moving off Intercom is more than a data transfer. Expect 6-10 weeks, depending on complexity:

Shorter timeline (6 weeks): Small team, simple workflows, few integrations, minimal historical data requirements

Longer timeline (10+ weeks): Large team, complex routing rules, many integrations to rebuild, extensive historical data to migrate, multiple team inboxes

Specific work involved:

  • Messenger widgets embedded across your product need replacement (1-2 weeks of developer work)

  • Workflows built in Intercom need recreation in the new platform (1-2 weeks, depending on complexity)

  • Agent retraining on new interface (2-4 weeks, expect 20-30% productivity dip)

  • Integration rebuilding with your tech stack (1-4 weeks depending on how many you have)

  • Historical data import and verification (1-2 weeks)

Making the Call#

Intercom has earned its place. Fifteen years of iteration show in how polished it is, though the company behind it is changing (renamed Fin in May 2026, now owned by Salesforce). Fin is good at conversation because it's trained on an enormous pile of real customer interactions. Messenger-first suits how SaaS products talk to their users. And your existing tools probably already connect.

But the gap between advertised and actual pricing is real. "$29/month" is a marketing anchor, not a budget. And $0.99 per outcome works well for companies where human tickets are expensive and AI quality protects revenue.

It's a problem for others. Lean teams whose human tickets are already cheap. High-volume transactional support where most tickets are routine. Growing companies, where doubling volume doubles AI costs with no published volume discount to soften it.

If you're in the second or third group, look at the alternatives properly. Action-taking AI like Hay.chat bundles 2,000 resolutions for €135/month, against Fin's $1,980 for the same volume. That's a big saving if your tickets are mostly transactional. You give up some platform maturity and conversational polish, and you get cost control and things actually getting done. Both approaches work. Your ticket mix decides.

Before signing anything: pull your ticket data. Calculate your human cost-per-ticket. Sample 50 tickets and categorise them. Do they need answers or actions? Project what 12 months of growth does to your bill. Set usage limits in whatever platform you choose.

Buyers who do this homework end up satisfied. Buyers who sign based on list prices end up surprised.

Sources

1. Intercom Official Pricing Page (intercom.com/pricing), checked October 2026

2. Fin AI Agent Pricing & Outcome Documentation (fin.ai/pricing), checked October 2026

3. Intercom Help Center: Plans Explained (intercom.com/help), December 2025

4. Capterra: Intercom Pricing Guide (capterra.com), 2025

5. Vendr Marketplace Negotiation Data (vendr.com/marketplace/intercom), 2025

6. Featurebase: Intercom + Fin Pricing Analysis (featurebase.app), October 2025

7. Customerly: Intercom Pricing Breakdown (customerly.io), 2025

8. BoldDesk: Complete Intercom Pricing Guide (bolddesk.com), October 2025

9. SaaSGenie: Intercom Plans and Pricing Analysis (saasgenie.ai), 2025

10. eesel AI: Intercom Fin AI Pricing Analysis (eesel.ai), 2025

11. Intercom: Today Intercom Becomes Fin (intercom.com/blog), May 2026

12. Salesforce: Salesforce Completes Acquisition of Fin (salesforce.com), September 2026

13. Intercom: Early Stage Programme (intercom.com/early-stage), checked October 2026

14. Salesforce, Inc. annual reports (Form 10-K) on SEC EDGAR, fiscal year ending 31 January (sec.gov), checked October 2026

About the Author

Damien Mulhall

Co-Founder, Strategy and Content

Damien spent 10+ years managing support operations and project delivery for global brands including Dell, Microsoft, Intel, and Google. He's PMP-certified, and that operations background shapes how Hay.chat is built.

Compare Hay.chat vs Fin