9 Best AI Customer Service Tools in 2026 (and the Resolution Tax)

Klarna reported that its AI assistant did the work of 700 full-time agents within a month of launch in February 2024. The buying question is what you pay for that work: a resolved ticket, a conversation, or an outcome whose definition can include a handoff.

Based on pricing nine tools on real cost per resolution rather than feature lists, Fin (formerly Intercom), HubSpot Breeze, and Zendesk AI agents are the best AI customer service tools for the buyers below. Prices and billing definitions were re-verified against primary pages on October 24, 2026.

That verdict will annoy two camps. The enterprise crowd wants Sierra or Decagon on the podium, and the bargain crowd wants the cheapest chatbot. Both miss the point of the billing model.

The point is the billing unit. In 2026, you may pay for AI usage alongside seats, workspace subscriptions, or ticket quotas. Get the unit wrong and a tool that looks cheap on the pricing page costs more per solved ticket than you expected.

A support agent is a metered service, and the question that matters is what you pay per unit of real work done. Read the usage definition alongside the base subscription before comparing the headline rates.

So that is how I priced all nine. Not by sticker. By cost per resolution after the resolution rate, which is the only number that survives contact with a real support volume.

What an AI customer service tool actually is in 2026

An AI customer service tool is software that answers customer questions, performs configured support actions, and passes cases to a human when needed. Some products charge per resolution; others meter conversations, sessions, or credits.

The distinction is whether the agent can complete the work your customers need. An answer to a help-center question and a completed refund workflow are different outcomes, even when both appear under the same AI label.

A 2026 AI customer service tool does three jobs. It resolves tier-one tickets end to end, the refund status, the password reset, the where-is-my-order. It assists human agents on the harder tickets by drafting replies and summarizing long threads. And it routes, reading intent and sending the urgent case to the right person.

When a vendor quotes a resolution rate, check how it defines the share of conversations completed by the AI. Drafting assistance, routing, and autonomous resolution are different jobs, and their billing units can differ too.

Now resolution is an invoice line. Zendesk announced outcome-based AI-agent pricing on August 28, 2024, according to its own pricing announcement. Fin charges per outcome. HubSpot uses credits for resolved text conversations. Help Scout and Gorgias also publish resolution-based rates.

The pitch is seductive. You pay when the AI delivers a defined result. But minimum commitments, included allowances, and the definition of success still determine the invoice.

In practice, the distinction that matters is whether a failed attempt consumes a paid unit. Per-resolution products meter defined outcomes; per-conversation products count handled conversations, whether or not they resolve them. Monthly credit bundles add another layer: unused capacity can raise your effective cost.

That distinction is the difference between paying for completed work and paying for attempts. A feature list does not tell you which one is on the invoice.

Some operators ask why not skip all nine and build the agent on a raw model API. The model call is only part of the system.

The rest is the retrieval layer, the escalation logic, the resolution ledger that explains what you were billed for, and the support knowledge that covers your edge cases. The vendors below sell that scaffolding as well as the agent. Build it yourself only if owning that system is worth the maintenance.

So before the tools, the math.

How I priced these, and the per-resolution tax

I priced every tool on one number: real cost per resolution. For conversation billing, that means dividing the usage cost by the number of tickets actually resolved. For outcome billing, it means checking that a billed outcome matches the resolution you wanted.

Here is the method. Take the billable unit and its price. If the tool bills per resolution, its unit price is the usage cost of a billable resolution, before base fees, minimums, or allowances. If the tool bills per conversation, divide the sticker by the resolution rate, because you are paying for the failures too. A hypothetical $0.40 per-conversation tool that resolves 57% of tickets has a real cost per resolution of $0.70. The 43% it could not solve still landed on your invoice.

That gap is the per-resolution tax, and it is the single most expensive thing buyers miss.

Run it on a mid-size support volume. A SaaS company handles 5,000 customer conversations a month. Assume a 60% resolution rate for comparison, not as a promise about any vendor. That is 3,000 resolutions and 2,000 unresolved conversations.

For 3,000 billable resolutions, Fin at $0.99 costs $2,970, HubSpot Breeze at $0.50 costs $1,500, Zendesk at $1.50 committed costs $4,500, and Help Scout at $0.75 costs $2,250. These are gross usage calculations before included allowances or credits, and assume each billed outcome is a completed resolution. Fin Procedures can make those counts differ.

A conversation-priced tool counts the attempts too. eesel now sells a 5,000-credit monthly plan for $1,749, or about $0.58 per resolution at the assumed rate, if all credits go to support tickets. Freshdesk at $0.49 a session costs $2,450, or about $0.82 per resolution, assuming one session per conversation and the starter allowance already used.

Here is the same scenario across the eight usage-priced entries. Identical inputs: 5,000 conversations, a 60% resolution rate, and 3,000 resolutions. These figures exclude base helpdesk fees, included allowances, trial grants, and discounts; Tidio requires a volume-specific quote rather than an extrapolation from its smallest plan.

Two things jump out of that table. First, eesel's monthly bundle changes the comparison: its effective rate depends on how much of the quota you use. Second, Crisp is absent because its published conversation counts are estimates from bundled AI credits, not a fixed tariff for this workload. That does not establish a hard volume ceiling; price the workspace and the required AI capacity together.

The base seat fees I excluded are not a rounding error. Add a ten-seat Zendesk Suite Professional plan and you are paying roughly $1,150 a month before the first AI resolution. Help Scout on Standard for the same team adds $250. HubSpot Service Hub Professional adds $900. The AI line is only one of the meters running, and a fair comparison reads all of them.

Now the opposite trap, because per-resolution billing has its own catch. When you only pay on a resolution, the vendor decides what a resolution is. That definition is written into the contract, and it is rarely symmetrical in your favor.

Fin makes the distinction explicit. Its pricing page charges $0.99 per outcome, including a completed Procedure that hands off to a human. It also says a customer request to speak to a human is not charged, and a reopened conversation gets its outcome refunded. A configured Procedure handoff and a customer-requested escalation are different billing cases; neither should disappear into a blanket 'pay only when solved' claim.

Help Scout runs the definition the other way. A resolution counts only if the customer receives an AI answer and does not escalate, search the knowledge base, ask a follow-up, or click the "I still need help" button. If the customer asks for a human, you are not charged. That is a clear definition to compare with the outcome rules of the other products.

This is the same lesson the per-seat era taught and buyers keep forgetting. I went through it pricing AI meeting note takers against the per-seat cliff and again pricing the AI-tier upcharge on social media tools. The unit always moves to wherever the vendor captures the most margin. In 2026 that unit is the resolution, and you should read the contract clause that defines it before you read the price.

The decision rule I would actually run: compare the conversation price divided by your measured resolution rate with the outcome price, then add subscriptions, minimums, and any billable handoffs. There is no universal resolution-rate cutoff. And whatever you pick, set a monthly spend limit before launch.

The same logic I used to price AI SDRs against human SDRs on cost per booked meeting applies here without a single change: never buy the outcome tool on the sticker, buy it on the outcome.

What separates a tool worth paying for

Price is how I ranked these. It is not the only thing I checked. Five criteria decided which tools made the list of nine and which got sent to the avoid pile, and you should run the same five against any vendor a sales team puts in front of you.

The first is a resolution rate you can validate on your ticket mix. Fin reports 76%, Help Scout 73%, and HubSpot Service Hub says 65%+. These are vendor-reported figures from different customer populations, not a controlled accuracy comparison. They can inform a pilot, but your own resolved-ticket count is the denominator that matters.

The second is a billing unit you can forecast. Per-resolution billing avoids paying for many failed attempts, subject to the outcome definition and minimum commitment. Per-conversation billing can still win at a sufficiently low price. Neither is wrong, but you need to know which one you are signing.

The third is an escalation path that does not bill you. When a customer asks for a human, the best tools, Help Scout among them, treat that as a non-event and charge nothing. The weaker contracts count a handoff as a billable outcome. Read that clause before the price, because it is worth more than a ten-cent rate difference.

The fourth is a spend control. Help Scout disables AI at its monthly cap, Fin offers limits and alerts, and eesel pauses at the credit limit unless you enable capped overages. A metered tool without a spending limit is an open invoice.

The fifth is context fit. An agent that can read your order data, your CRM record, or your subscription state resolves a different class of ticket than one answering from a help-center article. Gorgias reads a Shopify order. HubSpot Breeze reads the CRM. That fit is why two tools at the same rate are not the same purchase. I applied the same pricing-first method when I ranked the AI SEO tools worth paying for: the sticker gets you in the room, the fit and the unit decide the buy.

The 9 best AI customer service tools at a glance

The prices below were checked on October 24, 2026. The billing model matters as much as the starting figure; the worked comparison above separates usage from base subscriptions.

Read the price and the allowance together. A conversation bundle can look inexpensive per unit while becoming costly per resolved ticket if you leave credits unused or the agent fails often. A free helpdesk tier does not necessarily include a recurring AI allowance.

The 9 best AI customer service tools, ranked

The ranking is for an operator buying against a real P&L, not an enterprise procurement team. Self-serve, auditable pricing, and a billing unit you can forecast all count. Hype does not.

1. Fin: best for transparent outcome pricing

• Best for: teams that want a published outcome rate on an existing helpdesk
• Standout: $0.99 per outcome with no seat fees and no platform charge when run on an existing helpdesk
• Pricing: $0.99 per outcome; $49/month minimum includes 50 outcomes on an existing helpdesk
• Free trial: 14 days

Fin is the cleanest published outcome rate in this comparison, provided you read what counts as an outcome.

Intercom renamed the company Fin on May 12, 2026; Intercom remains the helpdesk product. Since September 10, 2026, Fin has been part of Salesforce, which positions it for rapid deployment on existing systems alongside the more customizable Agentforce platform. For buyers today, Fin still lists $0.99 per outcome and a $19/month Intercom seat in its calculator. Essential and Advanced offer monthly or annual self-serve plans without a contract; enterprise needs, Premier Support/onboarding, and custom terms require one.

The usage rate is $0.99 when Fin produces a billable outcome, with no Fin seat fee when you use an existing helpdesk. Fin's pricing calculator reports a 76% average resolution rate. That is a vendor average, not a forecast for your own tickets, and your existing helpdesk subscription still costs what it did before.

Run the multiplication before you start the trial. At the published 76% average on an illustrative 4,000 monthly conversations, 3,040 resolutions would cost $3,009.60, about $3,010, at $0.99 each. That leaves 960 conversations for other handling. The estimate excludes helpdesk seats and any additional billable Procedure outcomes.

The honest catch is the definition of outcome. A completed Procedure handoff can bill $0.99, while a customer asking to speak to a human does not. Your effective cost per fully autonomous resolution depends on that mix. Fin documents the distinction, which is why it still ranks first.

If you operate in a bilingual market, the language coverage matters as much as the price. Use the Falcon H1 bilingual support comparison to frame the trade-off, then evaluate the language and regional context of your own tickets.

Pricing in full: $0.99 per outcome, with a posted $49/month minimum including 50 outcomes on an existing helpdesk. No Fin seat fee on that setup. If you want Intercom's helpdesk underneath Fin, the calculator shows $19 per seat per month. The verdict: if you cannot decide, start with Fin, then check the outcome definition against your workflow.

2. HubSpot Breeze Customer Agent: best value if you already run HubSpot

• Best for: teams whose CRM, marketing, and service already live in HubSpot
• Standout: $0.50 per resolved text conversation under the standard credit rate sheet
• Pricing: 50 standard HubSpot credits per resolved text conversation, plus an eligible Professional or Enterprise subscription
• Free trial: 14 days of Customer Agent access on an eligible subscription

Breeze has the lowest published text-resolution rate among the main per-resolution picks here, and if your business already runs on HubSpot it is close to a default.

The mechanics: Breeze Customer Agent charges 50 credits per resolved text conversation at $0.010 per additional credit, or $0.50 under the standard rate sheet. That is about half of Fin and a third of Zendesk's committed rate. Voice uses a different billing unit, and flexible seats-and-credits subscriptions have a separate rate sheet. The agent can use your approved content, enabled CRM data, and configured actions rather than starting with an isolated chat widget.

The catch is the credit pool, and it is a real one. Breeze does not bill you cleanly per resolution like Fin. It draws from a shared HubSpot credit balance that also feeds other Breeze features, so a busy month of AI usage elsewhere in HubSpot quietly shortens your support runway.

Budget the pool as one expense. HubSpot also defines a resolution more broadly than a simple closed-ticket count: its current documentation uses a 72-hour window for agent replies, and manual or workflow assignment to a human does not itself cancel the charge. A qualifying visitor-initiated handoff does. I walked the broader credit-budget trade-off in the HubSpot Breeze outcome-pricing analysis.

The other cost is the platform underneath. Customer Agent is available with eligible Professional or Enterprise subscriptions across HubSpot, not only Service Hub. If you buy Service Hub Professional, it starts at $90 per seat per month paid annually, plus a required $1,500 onboarding fee. The free Service Hub tier is not a free Customer Agent plan.

Model it honestly. Five Service Hub Professional seats at $90 cost $450 a month. With 1,000 text resolutions, standard credits total 50,000. If all 3,000 included monthly credits remain available for support, the remaining 47,000 cost $470: a $920 monthly total, plus the one-time $1,500 onboarding fee. Other credit-consuming features raise that usage bill. Count the platform as a new expense if you do not already need it.

Pricing in full: Service Hub is free for up to 2 users, but Customer Agent requires an eligible Professional or Enterprise subscription. Service Hub Professional starts at $90 per seat per month paid annually plus $1,500 onboarding; Enterprise starts at $150 plus $3,500 onboarding. Standard Customer Agent text usage adds 50 credits, $0.50, per resolution before included credits. The verdict: if HubSpot already runs your business, Breeze is strong value. If it does not, count the platform cost before choosing it for the usage rate.

3. Zendesk AI agents: best for established support organizations

• Best for: support teams already on Zendesk that need depth, channels, and governance
• Standout: LLM-verified resolutions, with Forethought AI agents now named in Zendesk's offering
• Pricing: $1.50 per automated resolution committed, $2.00 pay-as-you-go, on top of a Suite plan
• Free trial: 14 days

Zendesk is the most expensive per-resolution rate on this list, and for a large, mature support organization it is still the right buy.

Zendesk's own pricing page still lists $1.50 per automated resolution on a committed plan and $2.00 pay-as-you-go. It now explicitly names Forethought AI agents by Zendesk: self-improving agents for complex workflows that can operate within Zendesk or on another provider’s stack. That is the current offering to discuss with sales, rather than assuming a migration is required.

What you buy for the premium is a defined verification step and a broad support platform. Zendesk's resolution tiers, introduced May 18, 2026, distinguish assisted escalations, contained resolutions, and verified resolutions. Only the verified tier, checked by a large language model, consumes the allowance. Existing accounts may still use the prior resolution platform. For a support org with strict QA, audit, and channel requirements, the definition and account setup matter more than a generic accuracy claim.

The cost stack is the problem for everyone else. The resolution fee sits on top of a plan, where Suite Team is $55 per agent per month and Suite Professional is $115, paid annually. Optional products add further charges. A ten-agent Professional team pays $1,150 in seats before additional AI usage. Zendesk's mix of included capabilities and optional products echoes the packaging question I covered when CallRail folded conversation intelligence into every tier: compare what your actual subscription includes.

Pricing in full: Zendesk Support Team is $19 per agent per month, Suite Team $55, Suite Professional $115, all billed annually. The pricing comparison lists $1.50 per committed automated resolution or $2.00 pay-as-you-go; account allowances and configuration determine the final bill. The verdict: the right call for an established support org that needs verification and governance, an overbuy for anyone who only needs a small standalone agent.

4. Gorgias: best for Shopify and e-commerce stores

• Best for: Shopify and e-commerce brands that need order-aware support
• Standout: included seats and AI actions against store data
• Pricing: $0.85 to $1.00 per automated interaction by plan, on top of a ticket-based helpdesk
• Free trial: free trial available

Gorgias is built for stores, and its native Shopify integration makes it a strong fit for order-aware support.

The AI agent handles order tracking and returns against your store data. The pricing comparison lists $1.00 per automated interaction on Starter, $0.90 on Basic and Pro, and $0.85 on Advanced, with payment tied to AI-resolved conversations. The Advanced bundle example elsewhere on the same page implies $0.90, so confirm that plan’s quoted rate. The Pro example below uses the consistent $0.90 figure.

The structural advantage is included seats. Gorgias does not price per agent, but seats are not unlimited: Starter includes 3, while Basic, Pro, and Advanced each list 500. The base helpdesk is priced on ticket volume: $10 a month for 50 tickets at Starter, then annual-billing equivalents of $50 for 300 at Basic, $300 for 2,000 at Pro, and $750 for 5,000 at Advanced.

The limit is focus. Gorgias is e-commerce software. If you run a SaaS product or a B2B service without an order workflow, its best features may not apply. The base plan is also a metered ticket bucket: overages are $0.40 per ticket on Starter, $40 per 100 on Basic, and $36 per 100 on Pro and Advanced. Two meters run at once, and you should model both.

Price an illustrative store. A Shopify brand on Pro pays $300 a month billed annually for a 2,000-ticket bucket. If the AI resolves 1,200 tickets at $0.90, that adds $1,080, for a $1,380 monthly total with up to 500 seats included. The same resolution count at Zendesk's $1.50 committed rate is $1,800 before its included allowance and seat fees. The saving comes from the whole package, not just the AI rate.

Pricing in full: Starter $10 a month for 50 tickets; Basic $50 for 300, Pro $300 for 2,000, Advanced $750 for 5,000 when paid annually; custom plans beyond that. AI is $1.00 on Starter, $0.90 on Basic/Pro, and listed at $0.85 on Advanced in the comparison table. Seats are included up to the plan limit. The verdict: the default shortlist for Shopify support, a poor fit when you have no commerce workflow.

5. Help Scout: best for small teams that want it simple

• Best for: small support teams that want a buyer-friendly resolution definition
• Standout: a resolution does not count, and does not bill, if the customer asks for a human
• Pricing: $0.75 per AI resolution, plus $25 to $75 per user per month billed annually
• Free trial: 15-day plan trial plus 3 months of unlimited AI resolutions

Help Scout writes the resolution definition in the buyer's favor, and that is rare enough to carry the whole recommendation.

Envoy charges $0.75 per resolution, between HubSpot and Fin on rate. The difference is what counts as a resolution. Help Scout bills only when the customer receives an AI answer and then does not escalate, search the knowledge base, ask a follow-up, or click "I still need help." If the customer pushes back at all, the resolution does not count and you are not charged. After the per-resolution camp's usual generosity toward the vendor, that asymmetry is genuinely unusual.

Two more buyer-side features back it up. You can set a hard monthly cap on AI spend, and when you hit it Envoy switches off until the next cycle. New accounts get three months of unlimited AI resolutions free, starting at account creation. Help Scout reports a 73% average resolution rate; use the trial to measure your own rather than treating vendor averages as a league table.

The trade-off is the seat fee. Help Scout’s annual prices are $25 per user per month at Standard, $45 at Plus, $75 at Pro, so the AI rate is not the whole bill. For a small team that is fine, since a five-seat Standard plan is $125 plus usage. For a large org the per-seat layer grows with the team, so compare the complete bill against Gorgias or Fin on an existing helpdesk.

Pricing in full: Free plan for 5 users with one inbox, Standard $25 per user per month, Plus $45, Pro $75 on annual billing. Envoy at $0.75 per resolution, with a 3-month unlimited free trial on signup and a configurable spend cap. The verdict: the safest pick for a small team, because the contract is written so the bill cannot ambush you.

6. Tidio (Lyro): best free starting point

• Best for: small businesses testing an AI agent before committing budget
• Standout: 50 one-time free Lyro conversations, with standalone AI available
• Pricing: Lyro from $32.50 a month for 50 conversations, billed annually
• Free trial: 7-day full trial plus the one-time AI allowance

Tidio is the easiest way to find out whether an AI agent works on your tickets without spending anything, and that is exactly what it is for.

Every Tidio account starts with 50 free Lyro conversations. That is a one-time allowance, not a monthly refill. You can connect Lyro to your site, point it at your help content, and watch those conversations resolve or fail before you decide. Monthly renewal of the quota requires a paid Lyro plan.

The billing unit still matters once you scale past free. The current Lyro starting price is $32.50 a month for 50 conversations on annual billing. A conversation counts once Lyro replies, whether or not it solves the issue. If you use the full quota, that is $0.65 per conversation and about $1.08 per resolution at the illustrative 60% rate. Leave part of the allowance unused and the effective cost rises.

Lyro can be bought on its own; a paid Tidio helpdesk plan is not mandatory. The helpdesk’s displayed annual rates start at $24.17 for Starter and $49.17 for Growth, while Plus starts at $300 a month plus usage. Premium offers quoted pay-per-resolution billing. For serious volume, compare that quote with an outcome-priced alternative instead of multiplying the smallest Lyro bundle into a fictional large-volume price.

Pricing in full: 50 one-time free Lyro conversations, then Lyro from $32.50 a month for 50 on annual billing. Helpdesk plans are optional; Plus starts at $300 a month plus usage, and Premium is quoted. The verdict: a good place to measure your own resolution rate before choosing a paid quota, not a promise of free recurring AI support.

7. Freshdesk (Freddy AI): best budget omnichannel

• Best for: budget-conscious teams that need email, chat, and social in one tool
• Standout: $29 per agent omnichannel support with session-priced AI
• Pricing: $49 per 100 Freddy AI sessions after a one-time 500-session grant, plus seats from $29/month annually
• Free trial: 14 days on the Enterprise plan

Freshdesk is the budget omnichannel option, with a low seat price and a separate session meter for Freddy AI.

The base value is real. Freshdesk Omni starts at $29 per agent per month billed annually, bringing ticketing, email, chat, and messaging into one tool. The first 500 Freddy AI sessions are included once per account on Growth, Pro, and Enterprise. Additional sessions cost $49 per 100, or $0.49 each.

The asterisk is the word session. Freddy bills per session, not per resolution: chat interactions are grouped into a 24-hour session and email into a 72-hour window. Assuming one session per conversation and a 60% resolution rate, $0.49 becomes about $0.82 per resolution. A ticket that generates more than one session changes that calculation.

The trade-off is the billing unit and workflow fit. A routine question and a longer case can both consume sessions. Check how your actual tickets map to those windows before comparing Freddy’s rate with a competitor’s outcome price.

Run the budget. Six agents on Freshdesk Omni Growth cost $174 a month billed annually. In the period when the one-time free allowance is available, 2,000 sessions leave 1,500 billable at $0.49, or $735, for a $909 total. After that allowance is exhausted, the same seats and 2,000 sessions cost $1,154. The starter grant is not a recurring monthly discount.

Pricing in full: Freshdesk Omni starts at $29 per agent per month billed annually. The first 500 Freddy AI sessions are included once per account on Growth, Pro, and Enterprise, then $49 per 100 sessions. The verdict: the budget shortlist for multi-channel support, provided you model sessions and seats together.

8. Crisp: best for a predictable workspace bill

• Best for: small teams that want a workspace subscription with bundled AI credits
• Standout: workspace pricing with included seats and AI credits
• Pricing: Free workspace; AI credits included from $45/month Mini
• Free trial: 14 days

Crisp makes the workspace subscription the starting point, and for a certain kind of operator that is the value.

Crisp prices a flat rate per workspace with included seats. Free covers 2 agents, Mini is $45 a month for 4, Essentials is $95 for 10, and Plus is $295 for 20, with additional Plus seats at $10 each. The workspace price is predictable, but seat additions and AI capacity still need budgeting.

The AI runs on bundled credits: $5 for around 90 automated conversations on Mini and $25 for around 450 on Essentials. That is roughly $0.06 in AI credits per conversation using Crisp's estimates, not a guaranteed unit rate and not the full cost once the workspace fee is included.

For a small team or an agency managing a fixed client retainer, included capacity can be useful. Just keep the distinction clear: a flat workspace subscription does not mean unlimited AI conversations.

The trade-off is the size of the included credit pool. A busy month can need more capacity than the plan includes. Compare your workload with the estimated allowance and confirm the additional capacity terms before promising a client a fixed bill.

Put a number on the included allowance. Essentials at $95 a month includes credits estimated to cover around 450 automated conversations. A team handling 600 chats needs to budget beyond that estimate. This is a capacity question, not evidence that Crisp cannot serve a larger team.

Pricing in full: Free for 2 agents, Mini $45 a month for 4, Essentials $95 for 10, Plus $295 for 20 with extra seats at $10. Bundled AI credits are $5 for around 90 conversations on Mini, $25 for around 450 on Essentials, and $75 for around 1,350 on Plus. The verdict: a useful workspace-priced option when the included allowance fits your actual volume.

9. eesel AI: best for wrapping an existing helpdesk

• Best for: teams that already own a helpdesk and want an AI layer on top
• Standout: draft mode before auto-response, with no separate seat or platform fee
• Pricing: Teammate from $299/month for 500 credits; one worked ticket or chat is one credit
• Free trial: 100 credits, no credit card

eesel is the wrapper, the tool you add to a helpdesk you already run rather than replacing it, and that remains its strongest fit.

eesel connects to your existing stack and learns from help content and past tickets. Its current pricing is a monthly credit bundle: $299 for 500 credits, $499 for 1,000, or $1,749 for 5,000. A ticket or chat the agent works on uses one credit regardless of length. There is no additional platform or per-seat fee, but the monthly subscription is a commitment to capacity.

The useful control before launch is draft mode: review its proposed replies before switching to automatic responses. The 100-credit trial lets you examine real drafts without a card. Use that to check whether the agent handles the ticket types you intend to give it.

The catch is still the billing unit. A worked ticket consumes a credit even if it is not resolved. On the $499 plan, using all 1,000 credits for tickets at an illustrative 57% resolution rate produces 570 resolutions, or about $0.88 each. On the 5,000-credit plan at the comparison’s 60% rate, it is about $0.58. Utilization and plan size now matter as much as the resolution rate.

Agents pause when credits run out unless you enable overages. Extra credits cost $0.80 each, with a monthly cap you set, and unused credits do not roll over. Slack questions, reports, and other tasks can consume the same pool. eesel also offers chat and email channels, so it is broader than a helpdesk-only plug-in, but keeping an existing support stack is still the clearest reason to choose it.

Pricing in full: Teammate starts at $299/month for 500 credits; 1,000 costs $499 and 5,000 costs $1,749. Enterprise is custom-priced. Annual upfront payment gets two months free. The verdict: a strong way to add AI to a helpdesk you already like, but buy the right credit batch and remember that a ticket touched is not necessarily a ticket solved.

The ones to avoid: the enterprise four

Sierra, Ada, Decagon, and Forethought offer enterprise AI agents. For an operator who needs an auditable self-serve budget, they require a different buying process: a quote and a careful review of the billable outcome.

The primary pages reviewed here do not establish a public dollar quote for your workload. Forethought does have a pricing page, while Sierra and Decagon explain their billing models. Treat a sales quote, including the base fee and usage commitment, as the comparison input.

Sierra documents outcome-based pricing, with agreed criteria for successful work and usually no charge for escalations. It also describes blended conversation pricing for routing or greeting tasks. Its essay does not publish the unit rate or an annual minimum, so ask for both rather than importing a reported deal size into your budget.

Decagon’s own pricing explanation offers both per-conversation and per-resolution contracts. Its resolution-pricing definition explains the model without publishing a unit price. Compare the quoted options using your own volume and resolution rate; no public dollar figure on those pages establishes the cheaper choice.

Ada offers a sales discussion and a custom business case. It does not establish a public per-interaction rate, platform minimum, or a universal billing unit. Get those terms in the quote before applying the resolution-tax calculation.

Forethought is presented as “Forethought AI agents by Zendesk” on Zendesk's pricing page, including deployment on other providers’ platforms. Forethought’s own pricing page describes platform access fees plus outcome-based charges and optional add-ons. Evaluate the overlap with Zendesk, but do not assume buying Forethought requires replacing your helpdesk.

Here is the real avoid rule, and it applies to all four and to any contract you are handed. Do not sign a per-resolution deal where you do not control the definition of resolution. When the rate is custom, the definition is custom too, and a broad definition written by the vendor's legal team turns every ambiguous interaction into a billable event.

The self-serve tools above publish rates and definitions that let you start the calculation. For a custom quote, get the same detail in writing: which events bill, which escalations do not, what is included, and what happens when usage exceeds the commitment.

If you have complex support needs and a procurement team that can evaluate the agreement, these four belong on your shortlist. If you need to compare a budget before a sales process, start with the published Fin, Breeze, or Help Scout terms. A custom vendor’s total could differ; the public pages do not prove a universal six-figure floor.

Frequently asked questions

What is the best AI chatbot for customer service?

For transparent outcome pricing, Fin at $0.99 per outcome remains the first pick, with a vendor-reported 76% average resolution rate. A completed Procedure handoff can also be billable, so this is not a promise that every charge is a fully solved ticket. A Shopify store should shortlist Gorgias, a HubSpot shop Breeze, and a small team that wants a buyer-friendly resolution definition Help Scout.

How much does an AI customer service chatbot cost?

Published usage rates here include HubSpot's $0.50 per resolved text conversation under standard credits, Help Scout’s $0.75 per resolution, Fin's $0.99 per outcome, and Zendesk's $1.50 committed or $2.00 pay-as-you-go resolution. Seat fees, monthly bundles, included credits, and minimums change the total. Enterprise custom quotes cannot be reduced to a verified public starting price from the pages reviewed.

How are AI chatbots used in customer service?

They answer routine questions, perform configured actions, draft replies for human agents, summarize conversations, and route cases to the right team. The billing unit depends on the product and feature: a completed resolution, a handled conversation, a session, or a credit-consuming action.

How do you set up an AI chatbot for customer service?

Connect your help center articles and approved support knowledge, review the agent’s answers, and roll out gradually to a small share of live tickets. eesel offers draft mode before auto-response. Measure the resolved-ticket rate and billable usage on your own workload before committing to a larger plan.

What is good customer service AI?

Good customer service AI has three things: a resolution rate you have checked on your own tickets, a billing unit you can forecast, and a clear escalation path with explicit charging rules. A tool that scores well on the first and fails the other two can still surprise you on the invoice.

What is the best AI tool for customer service?

It depends on the stack you already run. For a Shopify store, shortlist Gorgias; for a HubSpot business, Breeze; for a small team that wants a clear resolution definition, Help Scout; and for published outcome pricing on an existing helpdesk, Fin. There is no single winner, only a best fit for your billing tolerance and your data.

Which AI customer service tool should you choose?

The choice routes cleanly off two questions: what stack do you already run, and how hard are your tickets.

If you sell on Shopify, start with Gorgias. It connects support to order data and includes seats instead of charging per person. Model the ticket bucket and your plan’s AI rate together.

If your business already lives in HubSpot, Breeze is the same kind of default. Its $0.50 standard text-resolution rate can make sense when you already pay for an eligible subscription. Budget the shared credits, the resolution definition, and any new platform expense together.

If you run a small team and want a clear cap, Help Scout. The resolution definition and spending control are explicit, and three free months of AI let you measure before paying usage charges. For an established support org with QA and governance needs, Zendesk's verified-resolution model deserves a close look.

If you want a published outcome rate on an existing helpdesk, Fin. Its current $0.99 rate is straightforward to model once you separate completed resolutions from billable Procedure handoffs and account for the minimum. It is the pick when you cannot decide, which is why it is first.

And if you are still testing whether AI resolves your tickets at all, start free. Tidio's 50 one-time Lyro conversations or eesel's 100-credit trial and draft mode give you a starting sample. Your own results should drive the purchase. The same instinct pays off across the automation stack, the way a tracked ManyChat DM flow exposes the real cost per booked call: measure the outcome first, buy second.

Whatever you choose, set a monthly spend cap before you route a single live ticket, and read the contract clause that defines a resolution before you read the price. The vendor controls that definition. Pricing it is your job, and it is the only job that decides whether the tool pays for itself.

If you want the break-even model I used to price all nine, including the per-resolution tax calculation and the volume thresholds where each tool flips from cheap to expensive, I built it into a free AI business workflow audit checklist. Take it, run your own numbers, and join the newsletter while you are there for the next teardown.