Real Results · AI Automation

Case Studies

How Gideon Wafula has used AI voice agents, n8n workflows, and intelligent chatbots to solve real business problems — with numbers to back it up.

Case Study 01 · Hospitality · Nairobi, Kenya

Nairobi Restaurant Group: AI Voice Agent + Chatbot Cuts Missed Reservations by 65%

Retell AI ManyChat Voice Agent Instagram Chatbot WhatsApp

A Nairobi-based restaurant group operating four busy locations was hemorrhaging bookings. Their front-of-house team couldn't answer every inbound call during peak service hours, and hundreds of reservation enquiries arriving via Instagram DMs and WhatsApp each week were going unanswered for hours. Management estimated they were losing 40–50 potential bookings weekly — bookings that were going straight to competitors.

They needed a system that could answer questions, take reservations, and confirm bookings automatically — without adding headcount.

Gideon deployed a two-channel AI automation stack:

  • Retell AI voice agent on the main restaurant phone number — trained on the menu, hours for all four locations, reservation policies, and special dietary information. The agent handles inbound calls, confirms bookings in real time, and escalates to staff when a caller asks for something outside its scope.
  • ManyChat chatbot connected to Instagram DMs and WhatsApp — flows for table enquiries, menu questions, special event bookings, and gift vouchers. Connects to the restaurant's booking system to confirm availability without human intervention.

Both systems were integrated with the group's existing reservation platform via API. Setup and go-live took 11 days from brief to deployment.

65%
Drop in missed reservations
+42
New bookings per week
91%
Enquiries handled automatically
"We had no idea how many customers we were losing just because the phone rang during service. The voice agent now handles everything and we actually had to hire an extra waiter to cover the new tables." — Restaurant Group Manager, Nairobi
Case Study 02 · E-commerce · Nairobi, Kenya

Kenyan Clothing Brand: n8n Automation Saves 22 Hours per Week on Order Processing

n8n WhatsApp API Google Sheets Mpesa Order Automation

A growing Nairobi-based clothing brand selling across Instagram, TikTok Shop, and their own website was drowning in manual work. Every new order triggered the same multi-step manual chain: copy customer details from DMs into a spreadsheet, send a WhatsApp confirmation, generate an Mpesa payment link, log it again once payment came through, update inventory, and message dispatch details. Two team members were spending over 20 hours a week on these repetitive steps — time they could have used for product and marketing.

Errors were frequent: double-booked items, missed confirmations, and payment links that expired before customers responded.

Gideon designed a single n8n automation workflow that handles the full order lifecycle:

  • Order ingestion — new orders from the website and TikTok Shop trigger the workflow via webhook, extracting customer details, item, size, and address automatically.
  • Instant WhatsApp confirmation — a personalised message with the order summary is sent via WhatsApp Business API within seconds of the order being placed.
  • Mpesa payment link — the workflow generates and sends a payment link via Daraja API, logs it to a shared Google Sheets tracker, and sets a reminder if payment isn't received in 2 hours.
  • Inventory update — on payment confirmation, Google Sheets inventory is decremented automatically and a low-stock alert fires when items fall below threshold.
  • Dispatch notification — when the team marks an order as shipped in Sheets, the customer receives an automatic WhatsApp dispatch message with tracking info.
22h
Staff hours saved per week
0
Missed payment follow-ups
Order processing speed
"Before this, Fridays were chaos — 30 new orders and both of us glued to the phone. Now the whole thing runs itself and we're actually thinking about new designs instead of copy-pasting delivery addresses." — Brand Co-Founder, Nairobi
Case Study 03 · Creative Services · Seoul, South Korea

Seoul Creative Agency: Full AI Stack Triples Lead Qualification Rate

Voiceflow Zapier HeyGen Notion CRM Lead Automation

A small creative agency in Seoul — four team members, offering branding and content — was generating leads through their website and social presence, but converting poorly. Enquiries would come in, sit unanswered for a day or two while the team was in client work, and the prospects would have moved on by the time anyone followed up. When leads did respond, they often weren't well-qualified — wrong budget, wrong scope — and the team was burning time on discovery calls that went nowhere.

They needed a system that qualified leads instantly, educated prospects before the call, and filtered out poor-fit enquiries automatically.

Gideon built a three-component AI stack:

  • Voiceflow website chatbot — a custom AI assistant trained on the agency's services, case studies, pricing tiers, and FAQ. It answers questions and runs a short qualification flow (budget, timeline, project type), then either books a discovery call or routes them to an email nurture sequence.
  • Zapier lead pipeline — qualified leads from the chatbot flow into a structured Notion CRM with all collected data (budget, service interest, urgency, contact details). Zapier triggers a personal-looking email from the team within 5 minutes of qualification, and adds the lead to a 4-email nurture sequence.
  • HeyGen onboarding videos — a three-part video series (what the agency does, how projects work, past results) delivered automatically after the first email. The videos dramatically reduce time spent explaining the same things on discovery calls.
Lead qualification rate
70%
Less time on sales follow-up
<5 min
Lead response time
"Leads now come to discovery calls already knowing our process, our pricing range, and having watched our work. The quality of the conversations completely changed. We're closing more with fewer calls." — Agency Creative Director, Seoul

Want results like these?

I build AI automation systems for small businesses and agencies worldwide — voice agents, chatbots, n8n workflows, and content pipelines. Tell me what you're trying to solve.

Get in touch →
© 2026 Gideon Wafula — AI Automation Engineer, Seoul, South Korea · Home · Who is Gideon Wafula? · Services · FAQ