The most reliable way to sell AI chatbots to clients is to package them as a service, not software: a one-time setup fee for content preparation, build and testing, plus a monthly retainer for hosting, conversation review and updates. Lead with the client’s problem (repetitive questions, after-hours inquiries, buried information) and scope the project around the content the bot will answer from. Price from your own costs and time, then sanity-check against the value of the questions the bot will handle.
This playbook walks through packaging, scoping, proposals, pricing and the objections you’ll hear.
Start with the problem, not the technology
Clients don’t buy “an AI chatbot.” They buy fewer “what are your opening hours?” emails, faster answers for prospects, and a website that works while the office is closed. Good opening questions in a sales conversation:
- What questions does your team answer over and over?
- Where do those answers live today: website pages, PDFs, someone’s head?
- What happens when a visitor has a question at 9 p.m.?
- Which questions should the bot never answer (pricing exceptions, legal, medical)?
The answers tell you whether a chatbot is a fit and how much content work the job involves.
Good-fit clients
- Businesses with a lot of informational content: service providers, education, property, hospitality, software, professional services
- Teams that answer the same questions by email or phone every day
- Organizations with policy documents, guides or manuals that visitors struggle to find
Poor-fit clients
- A five-page website with almost no content and no documents (fix the content first)
- Clients who expect the bot to take payments, make bookings or replace staff judgment
- Anyone who won’t give you access to review and update content
How to package a chatbot service
A two-part structure is easy for clients to understand and keeps your revenue predictable.
The setup fee (one-time)
Covers everything needed to launch:
- Discovery session and question inventory
- Content audit of the website and documents
- Writing gap-filling FAQ documents
- Bot configuration, branding and welcome message
- Structured pre-launch testing
- Embedding the widget or publishing a hosted page
- Handover and a short client walkthrough
The monthly retainer
Covers keeping the bot accurate:
- Platform and hosting costs
- Monthly review of conversation history
- Adding or updating content when the website or offers change
- A short monthly summary: common questions, gaps found, changes made
Optional tiers
Many agencies offer two or three tiers that differ by number of bots, volume of content, review frequency and response time for changes. Keep the tiers about your service, not about features the platform may or may not support.
How to scope the content
Scoping is where chatbot projects go right or wrong. Before quoting, estimate:
| Scoping item | What to count | Why it affects price |
|---|---|---|
| Website pages | Number of pages with useful answers | More pages = more to audit and test |
| Documents | Number and length of PDFs, DOCX, Markdown files | Long or scanned docs need cleanup |
| Content gaps | Top questions with no written answer | You’ll be writing new content |
| Topics to exclude | Legal, medical, pricing exceptions | Needs explicit handling and testing |
| Update frequency | How often the site or offers change | Drives retainer effort |
| Number of bots | One site, or several brands/locations | Each bot needs its own setup and testing |
A common trap is quoting a flat setup fee and discovering 40 unwritten answers during testing. Put a content-volume assumption in the proposal. For the practical side of content prep, see how to train a chatbot on your website and PDFs.
Sample proposal outline
Keep the proposal short and concrete. A structure that works:
- The problem. Two or three sentences in the client’s words, for example “your team answers the same 20 questions every week.”
- What we’ll build. A branded AI assistant on your website that answers from your pages and documents and shows the source for its answers.
- Scope. Which pages and documents are included, content-volume assumption, topics excluded.
- Process and timeline. Discovery, content prep, build, testing, launch, with rough durations.
- What you’ll need to provide. Document access, website access or a developer to paste the embed code, a reviewer for test answers.
- Pricing. Setup fee, monthly retainer, what each includes, and what counts as extra work.
- After launch. Monthly review, content updates, summary report.
- Next step. Sign-off and kickoff date.
How to price an AI chatbot service
Your price should cover three things: platform cost, your time, and margin. Work it out from the bottom up, then check it against value.
Build the setup fee from hours
Illustrative numbers (replace with your own rates and estimates):
| Task | Hours |
|---|---|
| Discovery and question inventory | 2 |
| Content audit | 3 |
| Writing FAQ gap documents | 4 |
| Configuration and branding | 2 |
| Testing and fixes | 4 |
| Launch and handover | 1 |
| Total | 16 |
At an illustrative internal rate of $75/hour, 16 hours is $1,200. That’s your setup fee floor before any margin.
Build the retainer from recurring effort
Illustrative numbers:
- Platform cost per client: assume $50/month (get the real figure from your platform)
- Conversation review and updates: 3 hours/month × $75 = $225
- Cost total: $50 + $225 = $275/month
If you price the retainer at $400/month, your gross margin is $400 − $275 = $125/month, or about 31%. At $500/month it’s $225, or 45%. Decide what margin your agency needs and set the price accordingly.
Sanity-check against value
Illustrative example: if a client’s staff spend 5 hours a week answering routine questions at a loaded cost of $30/hour, that’s $150/week, or roughly $650/month (5 × $30 × 52 ÷ 12 = $650). A $400 retainer that meaningfully reduces that workload is an easy conversation. If the bot only deflects a few questions a month, the value case is weaker, and you should say so.
To run these numbers for different scenarios, use the AI chatbot pricing calculator.
Handling common objections
“Won’t the bot make things up?” Explain that the bot answers from the client’s own content and shows the source of its answers, so they can check any response. Show them your test plan. Our guide to chatbots with source citations covers how to explain this in plain English.
“We already have an FAQ page.” An FAQ page only helps visitors who find it and whose question matches a heading. A chatbot answers the question the visitor actually asked, drawing from the FAQ and every other page and document.
“Why do we need a monthly fee after it’s built?” Websites change, prices change, new questions appear. Without review, answers go stale. The retainer is maintenance, not rent.
“Can it replace our receptionist?” Be honest: it handles routine informational questions. It doesn’t replace judgment calls, and you should only promise what the platform does.
“What if visitors ask things it can’t answer?” You’ll configure and test how the bot responds when the answer isn’t in the content, for example by pointing visitors to the contact page.
Choosing a platform you can resell
You need a platform that lets you brand each client’s bot, keep clients separate and see what visitors ask. Techvia AI Bot is a white-label platform for agencies: client bots answer from website content and PDF, DOCX and Markdown documents with source citations, sit in separate workspaces, and can be delivered by script embed, hosted page or custom domain. Conversation history and usage for every client are visible from one agency dashboard. For a full checklist, read what to look for in a white-label AI chatbot.
Frequently asked questions
How much should an agency charge for an AI chatbot?
There’s no standard price. Build your setup fee from estimated hours and your retainer from platform cost plus monthly maintenance time, then add your target margin. Scope and content volume vary so much between clients that a fixed market rate is misleading.
Should I charge per conversation or a flat monthly fee?
A flat fee is easier for clients to budget. If your platform costs scale with usage, include a usage band in each tier and state what happens above it.
How long does it take to set up a client chatbot?
Mostly it depends on content. A client with a well-organized website and a few clean documents is much faster than one where you need to write answers from scratch. Estimate content prep and testing separately from configuration.
What should I report to clients each month?
The most common questions, questions the bot couldn’t answer well, content you added or changed, and anything the client needs to decide on.
Next step
If you’re building a chatbot service for your agency, book a demo of Techvia AI Bot to walk through the agency dashboard and confirm pricing and onboarding for your client mix.