Build Personalised Chatbot For a Family-Run Restaurant

Timur Restaurant contacted us to automate responses to common inquiries. The restaurant owner repeatedly answered the same questions, and missed chats resulted in lost covers. This created inefficiencies and negatively impacted customer service during busy times.

About Timur

Timur is a family-run Indian & Nepalese restaurant on Crystal Palace’s Westow Street. Open since 2020, it serves street-food-style curries, has lots of vegan options and takes bookings, takeaway and delivery via its own site plus the usual apps.
About _Build Personalised Chatbot For a Family-Run Restaurant
Challenges_Build Personalised Chatbot For a Family-Run Restaurant

Challenge

Owners often spend time answering the same questions repeatedly, causing delays in responding during service hours. This leads to missed chats, resulting in lost customer engagement and revenue opportunities.

Steps We Followed to Turn Their Challenges into Solutions

We have built a personalised chatbot and integrated it with their existing system.

Discovery & Mapping
  • Conducted 4 short calls with the owner and head, mapped processes, and counted question types such as dietary, opening hours, booking, and delivery radius FAQs.
  • Identified pain points, including repetitive questions, delayed replies during service, and missed chats, causing lost covers.
Use Case ID
  • Listed and scored 3 ideas based on value versus effort.
  • The chosen solution was an FAQ autoreply on WhatsApp to answer common questions in seconds.
Solution Design

Sketched the flow using Twilio WhatsApp integrated with a small function and a guard-railed large language model response.

  1. The pilot handled 20% of questions fully, reducing owner phone time by an hour daily.
  2. The top FAQs were compiled into a question-and-short-answer list, with a flow where inbound texts are checked against FAQs, answered if matched, or routed to GPT-4o with a controlled prompt for related questions.
Strategy & Roadmap (3 months)
  • Weeks 1-4 involved finishing the FAQ list, setting up, testing the prompt, and staff training.
  • Weeks 5-8 covered the live pilot, collecting unknown questions, expanding FAQs, and tuning the prompt.
  • Weeks 9-12 focused on launching the full version and linking it to the online booking page.
Turn Their Challenges into Solutions

Results

Build Personalised Chatbot
Fast Guest Response

Guests get answers in less than 5 seconds, even on busy nights.

Staff Time Savings

Staff can save 1 hour per day on average on calls.

Cost Savings

The restaurant can now save roughly £5,000 on staff costs

Remarks

“Working with ARCQ was a smooth process for us. They sat with us, studied every call and chat, and showed how a WhatsApp autoreply could save us heaps of time without changing how guests reach us. I’m looking forward to the pilot.”
Bishwas
SEO

Got an idea for your product?

We will build it for you from scratch and share the structured documentation to help you keep your system in top shape.
Scroll to Top