What SME Owners Actually Need From a Chatbot
Not a chatbot that pretends to be human. Not a chatbot that hallucinates product specs. What most small businesses need is simpler: a system that handles the 60–70% of incoming queries that are repetitive, and escalates the rest to a human with context.
The bar is: reduce the volume of queries that need human attention, and make the human responses better when they are needed.
3 Use Cases That Actually Work
1. Customer Support (FAQ + Escalation)
You train the AI on your knowledge base — FAQs, return policy, product specs, shipping info. Customer asks a question → AI answers from that knowledge base → if it can't answer confidently, it collects contact details and flags to your team with a summary.
Result: 60–80% of support queries handled automatically. Human team only deals with genuinely complex cases. Works best when your FAQs are well-documented.
2. Booking and Appointment
Customer messages asking about availability → AI checks your calendar → proposes slots → customer confirms → booking created and confirmation sent. No back-and-forth email chains. Works via your website, WhatsApp, or Telegram.
Result: Bookings happen 24/7 without staff involvement. We built this for a spa owner who was spending 3 hours/day managing WhatsApp bookings manually.
3. Lead Qualification
See how we built a lead qualification bot in a weekend using the same n8n + LLM approach.
Inbound lead arrives → AI asks qualifying questions (budget, timeline, use case) → scores the lead → if qualified, books a call with your team → if not, sends a polite response and logs the interaction. Your team only talks to people worth talking to.
Result: 5+ hours/week saved on discovery calls that go nowhere. Conversion rate on remaining calls typically improves because the AI filters noise.
Build vs Platform: Honest Comparison
Off-the-shelf platforms (Tidio, Intercom, Drift): Quick to set up, reasonable for basic FAQ handling. Get expensive fast ($100–$500+/month), limited customisation, often feel robotic, and lock you into their ecosystem.
Custom build (GPT-4o/Claude + n8n/webhook): More setup time (1–3 days), but you own it entirely, it integrates with your actual systems. For more on where AI fits in your business, see our practical AI integration guide., and API costs are typically under £5/month for most SME query volumes. No monthly platform fee.
Our recommendation: if you're handling under 500 queries/month, a platform is fine. Over that, build custom.
What to Avoid
- Generic AI with no business context — A chatbot that doesn't know your products, policies, or tone will damage trust
- No escalation path — Customers need to reach a human when the AI can't help
- Hiding that it's AI — Most customers accept AI chat when it's disclosed. Pretending otherwise creates backlash
- Deploying before testing — Run 50+ test scenarios before going live
- Platform lock-in without thinking — Read the contract before committing to a platform
Real Cost Breakdown
n8n self-hosted — ~$12/month for the VPS. No per-message fees.
Build cost — Typically $500–$1,500 depending on complexity and integrations.
Ongoing — Almost nothing. Update the knowledge base when your info changes.
Want a Custom AI Chatbot Built?
We build AI chatbots for SMEs — trained on your business, integrated with your tools. Fixed price.
[ View AI Automation Service → ]