How Much Does an AI Chatbot Cost in 2026? Real Prices

Quick answer: most businesses pay $0 to $200 per month for an AI chatbot. A solo operator runs on a free or entry plan (€0 to €19/month with Heeya). An SMB with an active support desk lands between $20 and $120 per month. Mid-market teams pay $200 to $1,500 per month, and enterprise deployments $1,500 to $5,000+ per month. A custom-built chatbot is a different purchase: $5,000 to $50,000+ upfront plus $200 to $2,000 per month to run it. What really decides your invoice is the pricing model, not the headline price. A flat subscription stays fixed, while per-conversation billing grows every time the bot does its job.

AI chatbot cost by business size (typical 2026 market ranges):

The 4 Types of AI Chatbots and Their Price Ranges

Not all chatbots are priced the same way because they are not the same product. The four categories below reflect fundamentally different architectures, deployment models, and cost structures. Choosing the wrong category for your needs is the most common source of both overspending and underperforming.

1. SaaS AI Chatbot Platforms ($0–$200/month)

SaaS platforms are the most accessible entry point. You create an account, configure your agent, upload your content, and deploy it to your website — often within a single afternoon. The provider manages infrastructure, model updates, and uptime. Your cost is a predictable monthly subscription.

• Examples: Heeya (€0–€99/month, ~$0–$113), Tidio ($0–$39/month for live chat, $0.50–$1.00/conversation for AI), Chatbase ($19–$199/month)
• Best for: SMBs, startups, e-commerce sites, professional service firms
• Advantages: fast deployment, predictable cost, no ML engineering required, knowledge base updated by uploading documents
• Limitations: customization bounded by the platform's feature set; some platforms charge per AI conversation on top of the base plan

The most important variable within this category is whether AI is included in the flat rate or billed per conversation. A platform advertising "$29/month" that charges $0.99 per AI resolution is not a $29 product for a business handling 1,000 support conversations monthly — it is a $1,019 product. Read the pricing page carefully before committing.

2. No-Code Chatbot Builders ($50–$500/month)

No-code platforms offer a visual builder for designing conversation flows and decision trees via drag-and-drop. They are more flexible than simple SaaS platforms but demand significant configuration time upfront — and ongoing maintenance every time your product, policy, or pricing changes.

• Examples: ManyChat ($15–$169/month), Chatfuel ($15–$300/month), Landbot ($40–$400/month)
• Best for: marketing automation, lead qualification sequences, multi-channel campaigns (Messenger, WhatsApp, SMS)
• Advantages: fine-grained control over conversation paths, strong multi-channel support
• Limitations: conversation trees become brittle at scale; every product update requires manual flow editing; limited generative AI capability compared to RAG-based platforms

3. Enterprise AI Chatbot Platforms ($500–$5,000+/month)

Enterprise solutions target large support teams with requirements around SLAs, advanced reporting, CRM integration, dedicated account management, and compliance certifications. Pricing reflects the operational depth, not just the AI capability.

• Examples: Intercom (from $39/seat/month + $0.99 per Fin AI resolution), Zendesk AI Agent ($55–$115/agent/month), Salesforce Einstein Service Cloud (from $75/user/month)
• Best for: large support organizations with hundreds of agents, regulated industries requiring audit trails, companies running thousands of AI-handled conversations per month
• Advantages: deep integrations (Salesforce, HubSpot, Jira, ServiceNow), SLA guarantees, enterprise-grade security certifications
• Limitations: high cost of entry, long implementation timelines (8–16 weeks typical), per-resolution billing can make costs unpredictable at volume

4. Custom-Built Chatbots ($5,000–$50,000+)

A fully custom chatbot is built from scratch by an agency or internal engineering team. The AI layer is assembled using open-source components or API-accessed models. Every integration, every conversation flow, and every retrieval mechanism is purpose-built for your specific requirements.

• Best for: businesses with highly specific integration needs (proprietary ERP, legacy CRM), organizations requiring on-premise deployment, companies with extreme compliance constraints that no SaaS vendor can satisfy
• Advantages: full control over code, data residency, and model selection; no vendor dependency
• Limitations: 2–6 month build timeline, significant upfront investment, and ongoing maintenance cost ($500–$3,000/month in engineering time) — the total cost of ownership over 24 months typically exceeds $80,000 for a well-maintained system

Understanding how RAG works helps clarify why custom builds are expensive: the retrieval pipeline, vector database, embedding model, and LLM orchestration each require specialist expertise to implement and tune correctly. A managed platform like Heeya abstracts all of that infrastructure — which is where the cost efficiency comes from. If you are evaluating an AI agent rather than a standard chatbot, the AI agent pricing section below covers how tool calls and multi-step reasoning change the cost equation.

AI Agent Pricing: What Changes Compared to a Chatbot

An AI agent does more than answer. It understands natural language, reasons over your documentation, and triggers actions: capturing a lead, logging an appointment request, filling in a form. Our guide to AI agent vs chatbot key differences covers the distinction. On the invoice, four factors move the price:

• The AI engine: every answer consumes LLM calls. Some platforms absorb that cost in a flat plan; others pass it on per conversation or per resolution.
• The knowledge base (RAG): how much content the agent can draw on, metered in characters, documents, or pages depending on the vendor.
• Tools and actions: an agent that chains several steps (search, check, act) makes more model calls per request. On a custom build, every tool adds tokens and integration hours.
• Volume: conversations or messages handled per month.

On a flat-rate SaaS platform, the gap between a chatbot and an agent stays small. Heeya's Standard plan (€19/month, ~$22) includes one AI tool of your choice (lead capture or appointment requests), and Premium (€99/month, ~$113) includes all of them (contact, leads, appointments). The gap widens with two other models:

• Per-action or per-conversation agent platforms: Salesforce Agentforce charges $2 per conversation, or about $0.10 per action through Flex Credits, so a conversation that needs many actions costs more. Our per-resolution pricing breakdown runs the same math for Fin, Zendesk, Gorgias, and Agentforce.
• Custom-built agents with RAG, tools, and CRM integrations: $10,000 to $50,000 to build, then $1,000 to $3,000 per month for hosting, LLM API usage, and maintenance, or $22,000 to $86,000 in the first year.

For multi-agent, autonomous deployments at enterprise scale, see our guide to agentic AI and autonomous agents.

Chatbot Pricing Models Explained

Two chatbots advertised at the same monthly price can produce invoices that differ by a factor of ten. The variable is the billing model. Six models dominate the market in 2026, and each one shifts the risk to a different place: to you, to the vendor, or to your support volume.

One question separates a safe contract from an expensive one: what does this cost at three times my current volume? Run that number before you sign. Under a flat subscription the answer is usually the same price, or one tier up. Under per-resolution billing it is three times the bill, and the better your chatbot performs, the faster that happens.

Chatbot Agency Pricing: What a Retainer Covers

Agencies typically split their pricing in two. A one-time setup fee, commonly in the $1,500 to $15,000 range depending on how many integrations and how much content preparation are involved, covers discovery, knowledge base preparation, conversation design, integration work, and go-live. A monthly retainer, usually $500 to $5,000, then covers hosting or the platform license, monitoring, content updates, and reporting. Project work outside the retainer is generally billed hourly, with market rates in the $75 to $200 per hour range.

Three questions are worth asking before signing an agency contract. Who owns the agent and the knowledge base if the engagement ends? Is the underlying platform license included in the retainer or rebilled with a margin? What does it cost to bring the chatbot in-house later? For a standard support or pre-sales use case, a self-serve platform plus a few internal hours of setup covers the same ground at a fraction of the retainer. An agency earns its fee when the project involves heavy content work, complex integrations, or multi-channel rollouts.

What Training Data and Knowledge Base Ingestion Cost

This line item surprises people, because the answer depends entirely on which side of the build-versus-buy line you sit.

With a managed SaaS platform, ingestion is included and priced by volume of content rather than by the hour. Heeya meters the knowledge base in characters: 20,000 on the free plan, 1,000,000 on Standard, and 3,000,000 on Premium. You upload PDF, Word, PowerPoint, or text files, or point the agent at a website URL, and re-uploading an updated document costs nothing extra. There is no data labeling step and no per-document fee.

With a custom build, every step is billable. Collecting and cleaning source content, deciding a chunking strategy, running the embeddings, hosting the vector database, and re-indexing each time the content changes all consume engineering hours, plus embedding API usage. Budget it as part of the build and the monthly maintenance, not as a separate product. Worth noting: a RAG-based chatbot needs no labeled training dataset at all. It reads your documents at query time, which removes the data annotation cost that older intent-based bots required.

Pricing Comparison Table: SaaS, Custom-Built, Agency-Built, Open-Source

The table below covers the six tools most commonly evaluated by SMBs and mid-market companies in 2026. Prices reflect publicly available plans, verified 2026-07-16. Enterprise pricing requires direct quotes and is noted where applicable.

The critical insight from this comparison: per-conversation pricing is a risk multiplier. A business generating 3,000 AI-handled conversations per month pays $2,970/month in resolution fees with Intercom Fin — before a single seat license. At that same volume, Heeya's flat-rate model keeps costs fixed and predictable. Visit the Heeya pricing page for the full tier breakdown, or see the detailed Heeya vs Intercom Fin cost per resolution comparison at 500, 1,000, and 5,000 monthly conversations.

For e-commerce businesses evaluating chatbot costs specifically for their storefront, our guide on AI chatbot pricing for e-commerce in 2026 covers the additional dimensions of product catalog integration, order status automation, and abandoned cart recovery — with pricing benchmarks specific to Shopify, WooCommerce, and BigCommerce deployments.

Hidden Costs You Should Anticipate

The advertised price is rarely the total cost. Independent audits of chatbot total cost of ownership (TCO) over a 12-month period commonly find actual costs running 2–3x the advertised subscription price once usage fees, onboarding, integrations, and overages are added in. Here is where that gap comes from.

Per-Conversation or Per-Resolution AI Fees

Several platforms separate the base subscription from the AI usage charge. You pay a platform fee for the interface, integrations, and agent seats — then a separate charge every time the AI successfully handles a conversation. At low volumes this is manageable. At scale, it becomes the dominant cost line. Always model your expected monthly conversation volume against the per-resolution fee before signing.

• Intercom Fin: $0.99 per resolved conversation
• Tidio Lyro: $0.50–$1.00 per AI conversation
• Zendesk AI: approximately $1.50 per automated resolution (add-on required)

At 2,000 AI-handled conversations per month, those fees add $1,000–$3,000 to your monthly bill regardless of the base plan cost. A flat-rate model eliminates this uncertainty entirely.

Onboarding and Implementation Fees

Enterprise platforms often charge a one-time setup fee separate from subscription costs. Zendesk professional onboarding packages range from $1,500 to $15,000 depending on the complexity of your configuration. Intercom offers implementation support at additional hourly rates. These fees are disclosed in contracts but rarely in public pricing pages.

Integration Development Costs

Connecting your chatbot to your CRM (Salesforce, HubSpot), your helpdesk (Freshdesk, Zendesk), your e-commerce platform (Shopify, Magento), or your internal systems often requires custom development. Budget $500–$8,000 per non-native integration, depending on API complexity and whether your vendor provides pre-built connectors.

Scenario Maintenance for No-Code Builders

No-code chatbots built on conversation trees require manual updates every time your product, pricing, or policy changes. Industry estimates suggest 3–8 hours of maintenance per month for an active no-code chatbot — at a fully-loaded internal cost of $50–$150/hour, that is $150–$1,200/month in hidden labor. A RAG technology-based platform like Heeya eliminates this: update your documents, and the AI adapts automatically.

Team Training Costs

Enterprise platforms with complex dashboards, ticketing logic, and analytics layers require agent training. A rough industry rule of thumb is about a day of training per agent for enterprise chatbot platform onboarding. For a team of 10 support agents at an average fully-loaded cost of $35/hour, that is roughly $2,800 in training cost before the platform goes live.

Overage Charges

SaaS platforms with message or conversation caps charge overage fees when you exceed your plan limits — typically $0.01–$0.05 per additional message. If your chatbot goes viral, experiences a seasonal spike, or your team deploys it more broadly than expected, a single month's overages can exceed your annual base subscription cost. Always check whether your vendor offers hard caps (the bot stops responding when the limit is hit) or soft caps with automatic overage billing.

How to Calculate the True ROI of an AI Chatbot

The ROI calculation for an AI chatbot is straightforward once you have the right inputs. Use this formula: (Monthly labor cost offset + Revenue impact) - Monthly chatbot cost = Net monthly ROI.

Here is a concrete example for a mid-sized SMB with an active customer support function:

In this scenario, the chatbot pays for itself roughly 120x over in its first month of operation. Even using conservative estimates — 30% automation, lower lead value — the payback period for a flat-rate SaaS chatbot is typically measured in days, not months. Salesforce's State of Service report found that service leaders expect AI agents to cut service costs by roughly 20% on average — realized savings for any specific deployment depend heavily on deflection rate and knowledge base quality.

The ROI picture changes substantially for custom-built solutions. A $30,000 development project with $2,000/month in ongoing maintenance costs requires 18–24 months to break even against a comparable SaaS solution — assuming the SaaS solution could have covered your requirements. The build-vs-buy decision should factor in this payback timeline explicitly, not just the upfront sticker price.

For the AI customer service chatbot use case specifically, deflection rates tend to be higher because support questions are repetitive and well-suited to retrieval-based answering. Real deflection rates of 60–80% are achievable with a well-structured knowledge base — which is the upper end of what the calculation above would yield.

Heeya Pricing: What's Included at Each Tier

Heeya operates on a flat-rate subscription model. There are no per-conversation fees, no AI usage credits to monitor, and no resolution-based billing. Every plan includes the full RAG pipeline: document ingestion, vectorization, semantic retrieval, and LLM-generated answers — all managed infrastructure.

Every plan — including the free tier — includes the full RAG engine. Upload PDFs, Word documents, PowerPoint files, plain text, or website URLs. The AI answers from your content, not from generic internet data. There are no hidden usage credits, no per-resolution charges, and no surprise billing at the end of the month.

The Enterprise plan adds dedicated support, GDPR data processing agreements, custom integrations, and SLA guarantees. It is designed for organizations running high conversation volumes across multiple departments — support, sales, HR, and internal knowledge management simultaneously.

Full plan details, including which features are available at each tier, are on the see Heeya's plans page.

Which Budget Fits Your Company Size?

Freelancers and Solo Operators ($0–$19/month)

The free Heeya plan handles the needs of a solo operator or early-stage startup: one AI agent, 30 messages per month, a knowledge base covering your core service offering. It is enough to qualify inbound leads, answer FAQ after hours, and test whether AI automation fits your workflow before committing a budget. The Standard plan at €19/month (~$22) extends that to 500 messages — appropriate for a professional services site or single-product business with steady inbound traffic.

Small and Mid-Sized Businesses ($19–$200/month)

Most SMBs with active customer support functions will find everything they need in the €19–€99/month range (~$22–$113) on a flat-rate platform. Heeya Premium at €99/month (~$113) covers three concurrent AI agents and 2,500 messages per month — sufficient for a business with separate support, pre-sales, and HR agents running simultaneously. Chatbase at a similar price point is a reasonable alternative if your team is developer-led and wants API-first control. Tidio works well if you need live chat alongside AI, with the caveat that Lyro AI conversation costs add up fast beyond 500 conversations/month.

Mid-Market Companies ($200–$1,500/month)

At this scale, the decision pivots between a well-configured SaaS platform (Heeya Enterprise or a comparable solution) and an entry-level enterprise tool. The key questions are: Do you need native CRM sync with Salesforce or HubSpot? Do you require audit logs for compliance? Do you have dedicated support staff who need an agent interface alongside the AI? If yes to two or more, an enterprise tier starts to make sense. If no, a well-configured SaaS chatbot at $200–$400/month handles the volume comfortably and avoids per-seat bloat.

Enterprise Organizations ($1,500+/month)

Large enterprises with structured support organizations, compliance requirements, and thousands of monthly conversations typically require either a purpose-built enterprise platform (Zendesk Suite, Salesforce Einstein, or comparable) or a custom-built system. The relevant cost comparison at this scale is not chatbot vs chatbot — it is chatbot cost vs the headcount cost of manual handling at the deflectable volume. Enterprises that successfully deploy AI chatbots at scale commonly report cost-per-contact reductions in the range of 20–30% within 12 months, driven mainly by ticket deflection rather than headcount cuts. At enterprise contact volumes, that reduction covers platform costs many times over.

How AI Chatbot Pricing Compares to Live Chat and Human Support

The most useful pricing benchmark for an AI chatbot is not another chatbot — it is the alternative: human agents handling the same volume manually, or a traditional live chat tool that still requires a human on the other end.

Cost Per Interaction: AI vs Human Agent

Industry benchmarking puts the average fully-loaded cost-per-ticket for human-handled support at roughly $15–$20 in North America once salary, benefits, management overhead, and workspace costs are factored in — compared to a global blended average closer to $8–$12. For a business handling 2,000 support contacts per month, that is $30,000–$40,000/month in support labor.

An AI chatbot on a flat-rate platform at €99/month (~$113), automating 50% of that volume, saves $15,000–$25,000/month in support cost while handling 1,000 contacts — at a cost-per-automated-contact of roughly $0.10. That is a cost-per-contact reduction of 99% on the automated interactions. The human team handles the remaining 1,000 complex or escalated contacts — where their judgment actually adds value.

Traditional Live Chat Tools

Traditional live chat platforms (Intercom, Drift, LiveChat) require a human agent to be present and responsive. They reduce friction for visitors but do not reduce labor costs — they redistribute them. At $39–$75 per seat per month with a team of five agents, you are paying $195–$375/month in seat fees before a single conversation happens, plus full agent salaries. Adding AI to those platforms introduces per-resolution fees on top. The hybrid model — AI handles what it can, humans handle the rest — is operationally optimal, but only if the AI layer is priced predictably.

The Staffing Comparison

A single full-time customer support agent in the US costs $45,000–$65,000/year in salary alone, or $55,000–$80,000 fully loaded. An AI chatbot handling the equivalent of a full-time agent's repetitive inquiry volume costs €228–€1,188/year (~$260–$1,355) depending on platform and plan. The math is not a justification for eliminating support staff — it is a justification for redirecting your best agents toward complex, high-value interactions while AI handles tier-1 repetitive volume around the clock. For the smaller-scale version of this same comparison, hiring a part-time virtual assistant instead of a full-time agent, see our AI chatbot vs virtual assistant cost breakdown.

Support teams that offload repetitive tier-1 questions to AI commonly report higher agent job satisfaction — because agents spend more time on interesting, solvable problems and less time copy-pasting policy responses.

Further Reading

• Heeya AI Chatbot Platform — deploy a RAG-powered agent in minutes, no ML engineering required
• AI Customer Service Chatbot — automation strategy for support teams at any scale
• RAG Architecture Deep-Dive — how retrieval-augmented generation works under the hood
• E-Commerce Chatbot Pricing — cost breakdown specific to online retail use cases
• What Is RAG? Business Guide 2026 — the architecture powering modern AI chatbots explained
• Gorgias Pricing Explained (2026) — how per-ticket and per-resolution billing compares to a flat-fee AI chatbot
• Best AI Chatbot for Small Business (2026) — cost-conscious picks for lean support teams
• Heeya Pricing Page — complete tier breakdown, from free to enterprise
• AI Chatbot Pricing for E-commerce (2026) — vertical-specific cost breakdown covering Shopify, WooCommerce, and BigCommerce deployments, with ROI benchmarks for cart recovery, support deflection, and product Q&A use cases

FAQ