Chat Bot for Business: The Complete Guide to Getting One Built in 2026

You want a chat bot for your business. Maybe for customer support, maybe for sales, maybe to answer the same 50 questions your team gets every day. You've seen the demos. They look impressive. But when you actually try to get one built, the questions pile up fast.

What kind do I need? How much does it cost? Do I build it myself or hire someone? What about GDPR? Will it actually work or will it annoy my customers?

Here's the honest guide.

What kind of chat bot do you need?

There are three tiers, and the price and complexity differ enormously:

Tier 1: Rule-based chat bot (£500-£2,000)

A flowchart. The customer clicks buttons or types keywords, and the bot follows a scripted path. "Press 1 for billing, press 2 for support."

Good for: Simple FAQ responses, appointment booking, basic lead capture. Bad for: Anything requiring understanding of natural language. If the customer phrases their question slightly differently from expected, the bot fails. Build with: Intercom, Tidio, ManyChat, Chatfuel. No custom development needed.

Tier 2: AI-powered chat bot (£2,000-£8,000)

Uses a large language model (GPT-4, Claude, etc.) to understand natural language and generate responses. Trained on your business knowledge — products, pricing, policies, FAQs. Handles variations in how people ask questions.

Good for: Customer support, product recommendations, lead qualification, FAQ handling at scale. This is what most businesses actually need. It handles 70-80% of customer queries automatically and hands off the rest to your team. Build with: Custom development using AI APIs, or platforms like Voiceflow, Botpress.

Tier 3: Full AI agent (£5,000-£12,000+)

Does everything Tier 2 does, plus takes actions — looks up orders in your system, processes returns, schedules appointments, takes payments. Integrated with your CRM, booking system, payment processor, and internal tools.

Good for: Businesses with high support volume that want near-complete automation. Build with: Custom development. No off-the-shelf solution handles deep integrations well.

Most businesses reading this need Tier 2. If you're not sure, start there.

What it actually costs

Real pricing, not marketing ranges:

What you get	Typical cost	Timeline
Rule-based bot (platform)	£500-£2,000	1-2 weeks
AI chat bot (custom)	£3,000-£8,000	2-4 weeks
AI agent with integrations	£5,000-£12,000	3-6 weeks
Running costs (AI API + hosting)	£100-£400/month	Ongoing

The build cost is a one-time investment. The running costs are forever — they scale with conversation volume. A business handling 500 conversations/month pays less than one handling 10,000.

The hidden cost most people miss: compliance. If your chat bot handles customer data (it will), you need compliance documentation — a DPIA, data processing agreements with your AI provider, and updated privacy notices. Budget an extra 30-50% if this isn't included in the build quote.

Full pricing breakdown: How Much Does an AI Chatbot Cost?

Build it yourself vs hire someone

DIY with a no-code platform

Platforms: Voiceflow, Botpress, Chatfuel, Tidio, Intercom Cost: £50-£500/month for the platform Good for: Simple bots, businesses with technical staff who can maintain it

The tradeoff: You save on build cost but invest your own time. If the bot breaks, you fix it. If you need a complex integration, you're stuck. And no platform handles compliance documentation for you.

Hire a freelance developer

Cost: £2,000-£5,000 for a custom build Good for: Businesses that want a tailored solution without agency prices

The tradeoff: Quality varies enormously. A freelancer at £30/hour builds a very different bot than one at £100/hour. Most freelancers don't handle compliance — you'll need a separate engagement for that.

Hire a build-and-comply firm

Cost: £3,000-£8,000 for the build including compliance docs Good for: Businesses that want it done right without managing multiple vendors

The tradeoff: Higher upfront cost, but you get one package — working bot plus DPIA, DPA documentation, and privacy notices. No separate compliance engagement.

Detailed comparison: How to Automate Customer Support With AI

What to look for in a chat bot builder

Whether you're hiring a freelancer, agency, or build-and-comply firm:

1. Ask what AI model they use. "We use AI" isn't an answer. GPT-4, Claude, Gemini — which one, and why? Different models have different strengths, costs, and data handling terms.

2. Ask about data handling. Where does conversation data go? Is it stored? For how long? Does the AI provider retain it? This matters for GDPR compliance.

3. Ask about handoff to humans. The bot won't handle everything. How does it transfer a conversation to your team? Does it pass context so the customer doesn't repeat themselves?

4. Ask about training and updates. Your products change. Your pricing changes. How does the bot's knowledge get updated? How often? Is there a maintenance cost?

5. Ask about compliance. Specifically: will they deliver a DPIA, review the DPA with the AI provider, and update your privacy notices? If the answer is "that's not our area," budget for it separately.

6. Ask for a demo. Not a video — a live demo where you ask real questions a customer would ask. Watch how it handles edge cases, off-topic questions, and requests it can't fulfil.

The compliance side (don't skip this)

Your chat bot processes personal data from the first message. At minimum: a user identifier, their IP address, and whatever they type. Often: names, emails, phone numbers, order details, payment information.

Under GDPR (UK and EU) and the EU AI Act, you need:

1. A Data Protection Impact Assessment. AI processing customer data triggers GDPR Article 35. The ICO fined MediaLab.AI £247,590 partly for not doing one. A DPIA for a chatbot takes 1-2 weeks and costs £1,500-£3,000 if done separately.

2. A Data Processing Agreement with your AI provider. Every major provider offers one. Most developers never sign it. If you're sending customer data to OpenAI without a signed DPA, you're technically in breach of GDPR.

3. Updated privacy notices. Your customers need to know their conversations are processed by AI. One paragraph added to your privacy policy.

4. AI Act transparency. From August 2026, you must tell users they're interacting with an AI system. A disclosure at the start of the conversation handles this.

The cheapest approach: get a builder who includes compliance documentation in the build price. Retrofitting costs more.

Common mistakes

Building for the website when your customers are on WhatsApp. If you're in a market where WhatsApp dominates (Nigeria, much of Europe, parts of the UK), a website chat widget is the wrong channel.

Over-engineering the first version. Start with FAQ + handoff. Add order tracking later. Add payment processing later. Don't try to build the full agent from day one.

No handoff to humans. A bot that can't escalate to a person frustrates customers more than no bot at all. Always build the escape hatch.

Stale knowledge base. If your prices changed last month and the bot quotes old prices, you have a bigger problem than before. Build an update process.

Ignoring compliance. GDPR, the AI Act, and data protection laws aren't going away. The fines are real. Build compliant from the start — it's cheaper than retrofitting.

Getting started

Pick your tier. Most businesses need Tier 2 (AI-powered, knowledge-trained).
Define the scope. What questions should the bot answer? Export your last 3 months of customer support queries to find out.
Get quotes. At least two. Ask every question above.
Check compliance is included. If not, budget 30-50% extra.
Start before August 2026. The EU AI Act deadline makes this more urgent, not less.

Need a chat bot built for your business? We build AI chat bots with compliance documentation included — DPIA, DPA review, privacy notices, AI Act transparency. Fixed-price, no hourly billing. Get a quote.