Custom AI Chatbots for Small Business
A custom AI chatbot for small business answers your customers' questions and qualifies leads, using a knowledge base built from your own documents so its answers come from your information. We build each one on Botpress, a platform for building and hosting AI chatbots, and every project starts with a free 20-minute consult.
- Free 20-minute consult
- Based in Tulsa, OK ยท working nationwide
- Plain English, no jargon
Who is a custom AI chatbot for?
A custom chatbot is for small and mid-sized businesses that answer the same customer questions every week, or that want an AI chatbot for their website to sort visitors into real leads before anyone picks up the phone. It works best when your answers are already written down somewhere, even if they're scattered.
- Businesses that answer the same questions by phone and email every week
- Teams that want leads qualified, with the key details gathered, before a salesperson steps in
- Owners whose answers already live in FAQs, policies and service pages that customers can't find
- Small and mid-sized businesses, roughly 2 to 200 employees, that want to add capacity, not cut people
If most of your customer conversations need a person, like sensitive or one-of-a-kind requests, a bot may not be the answer, and we'll tell you so.
What's included in a custom AI chatbot for small business?
You get a bot built around your business: a knowledge base made from your own documents, conversations designed to answer questions and ask the right follow-ups, and lead qualification that gathers what your team needs to know. We build it, test it and launch it, then help you keep it current if you want.
- A knowledge base built from your own documents, organized so the right answer is easy to find
- A custom chatbot built on Botpress and designed around the questions your customers actually ask
- Conversation design: when the bot answers, when it asks a follow-up and when it qualifies a lead
- Lead qualification that asks the questions your team would ask and gathers the details
- Routing that guides each conversation toward the right next step
- Weekly or biweekly progress updates while we build, and optional monthly support after launch
How do we build your AI chatbot?
We build your AI chatbot in four steps: Discover, Design, Build and Deploy, with you in the loop the whole way. It starts with a free 20-minute call, and we ship a first version fast, then improve it.
Discover
A free 20-minute consult, then a scoping conversation about the questions your customers ask most, the leads you want and where your answers live today.
Design
We plan the knowledge base and the conversation flow, then send a written proposal with a fixed price or a range.
Build
We build the knowledge base and the chatbot, test it with the questions your customers really ask and send weekly or biweekly progress updates.
Deploy
We launch a first version, then improve it. Ongoing support is available as a monthly plan.
What do we build chatbots with?
We build custom chatbots on Botpress and train each one on a knowledge base made from your own content. We also design the conversation: when it answers, when it asks a follow-up question and when it qualifies a lead.
- Botpress, the platform your bot is built on and runs on
- Your own documents: FAQs, policies, service pages and product or service details
- A knowledge base that organizes those documents so the right answer is easy to find
- Conversation design: a plan for when the bot answers, asks a follow-up or qualifies a lead
How do you train an AI chatbot on your own documents?
You train a chatbot on your own documents by turning them into a knowledge base that it searches before it answers. In most cases, the AI model itself isn't retrained at all. It looks up the relevant part of your content and answers from it, so better documents make better answers.
That's why an AI chatbot trained on your data is only as good as the data. The custom chatbot in our portfolio runs on Botpress and answers from a knowledge base we built from the business's own information, then asks follow-up questions based on what the visitor has said so far.
The best source material is whatever your team already uses to answer customers:
- Your FAQ page, plus the questions your team answers by email and phone every week
- Policies, such as returns, warranties, cancellations and service areas
- Product or service details, including what's included and what isn't
- Anything customers ask about that only lives in someone's head (write it down first)
How do you build a knowledge base for an AI chatbot?
You build a chatbot knowledge base by gathering the answers your business already gives, clearing out anything outdated or contradictory, and organizing the rest by topic so the bot can find the right piece. Then you test it with real customer questions and fill the gaps it finds.
When there's a lot of source material, the gathering and checking can be automated. We built a multi-agent AI pipeline that does this: specialized AI subagents each handle one job, such as research, extraction or document analysis, and a quality-check step validates the results before they're saved as structured records in a database.
For most small businesses, the job is simpler. Step by step:
- Gather: collect your FAQs, policies, service pages and the answers your team types out every week.
- Clean: remove old prices, retired policies and answers that contradict each other.
- Organize: group the content by topic, with one clear answer to each question.
- Test: ask the bot the questions real customers ask, and add whatever it can't answer.
- Maintain: update the knowledge base whenever your prices, hours or policies change, or it will keep giving the old answer.
Can an AI chatbot qualify leads?
Yes. A lead qualification chatbot on your website asks visitors the questions your team would ask, like what they need and when, and guides each conversation toward the right next step. Qualifying leads is one of the two main jobs we build chatbots for, along with answering customer questions.
Lead qualification means finding out which visitors are real prospects, and what they need, before your team spends time on them. The Botpress bot in our portfolio qualifies leads as part of the conversation, asks follow-up questions based on what each visitor has said and routes every conversation toward the right next step.
Getting those leads into another app, like a Google Sheet or your CRM (the software where you track customers), is a workflow automation job: connecting apps that don't talk to each other.
Questions a lead-qualifying bot might ask:
- What do you need help with?
- When do you need it done?
- Where are you located?
- What's the best way to reach you?
AI agent vs. chatbot: what's the difference?
A chatbot talks with people: it answers questions and gathers details in a conversation, and it waits for someone to ask. An AI agent does work: it follows several steps, uses tools and can run on a schedule without anyone typing a message. Some businesses need one, and some need both.
| Compare | AI chatbot | AI agent |
|---|---|---|
| Main job | Answers questions and qualifies leads in a conversation | Finishes a multi-step task, like research or turning documents into data |
| How it starts | Someone types a question | A schedule, or a task you hand it |
| What it works from | A knowledge base of your business's information | Tools and connections to your business systems |
| Who it talks to | Your customers and website visitors | Usually your team, when the work is done |
| Example from our work | A Botpress bot trained on a business's own knowledge base | A Claude agent that writes a short AI news brief every morning |
What does an AI chatbot cost to build and run?
A chatbot has two kinds of cost: the one-time build and the ongoing cost to run it. The build is quoted in a written proposal with a fixed price or a range. Running costs can include the platform that hosts it, and ongoing support is an optional monthly plan.
We don't publish prices, because every bot needs to know and do different things. Payments follow milestones, starting with a deposit, and any change to the scope goes through a written change order, so there's never a surprise invoice.
What drives the cost:
- How much it needs to know, and how ready your documents are
- What it needs to do: answer questions only, or qualify leads too
- Whether leads need to reach another app, like a spreadsheet or your CRM
- How many conversations it handles, which can affect the monthly platform cost
More AI chatbot questions
How much does an AI chatbot cost per month?
A chatbot's monthly cost can include the platform that hosts it, plus an optional monthly support plan if you want help keeping it current. The one-time build is separate and quoted in a written proposal with a fixed price or a range. What the bot needs to do, and how much it gets used, usually drive the monthly side most.
What's the best AI chatbot platform for a small business?
The best AI chatbot platform for a small business is one that answers from your own information, lets you shape how conversations go and has running costs you're comfortable with. No single platform wins for everyone. We build custom chatbots on Botpress, and if a simpler tool you already pay for would do the job, we'll tell you.
Get help choosing AI tools with AI consultingIs an AI chatbot worth it for a small business?
A chatbot is worth it when it saves your team real time or helps you respond to leads faster, and you should be able to see that in your own numbers, like the hours spent on repeat questions before and after launch. Start small, with the handful of questions your team answers most, and grow it from there.
How long does it take to build an AI chatbot?
It depends mostly on how ready your information is. If your answers are already written down, the knowledge base comes together faster. If they live in people's heads, gathering them is the first job. Either way, we ship a first version fast and then improve it, with weekly or biweekly progress updates so you know where things stand.
Will an AI chatbot make up answers?
Any chatbot can get an answer wrong, so the goal is to keep mistakes rare and easy to spot. A bot that answers from a clean, well-organized knowledge base, and has been tested with the questions your customers really ask, has far less reason to guess. We're upfront about that risk, and about which questions a person should answer instead.
Will a chatbot replace my staff?
No, and that isn't the goal. We build chatbots to add capacity, not to cut people. The bot handles the questions your team answers over and over, plus the first few questions for a new lead, so your people can spend their time on the conversations that really need a person.
Do I need a chatbot, or a better FAQ page?
Sometimes a clear, well-organized FAQ page is enough, and if that's true for you, we'll say so. A chatbot earns its place when customers ask the same questions in many different ways, when the answer depends on their situation, or when you want leads qualified before your team steps in. The free 20-minute consult is a quick way to find out.
See small business website and app developmentHave another question? Our AI consulting FAQ covers cost, timelines, data privacy and ongoing support, or you can ask us directly.
Tired of answering the same questions every day?
Book a free 20-minute consult. Tell us what your customers ask most and where your answers live today, and we'll tell you honestly whether a chatbot is the right fix.
Prefer email? Write to admin@legacyaisolutions.ai.