CEA Circular
An advisor that never clocks out
A dedicated WhatsApp AI assistant for CEA Circular, a driving school in Bogotá, built on the WebMinds WhatsApp AI platform. It answers prices, requirements and schedules in about eight seconds, sends the right flyer or reel, turns every chat into a structured lead tagged by source, and hands the conversation to a human advisor the moment a student is ready to enroll — with a full admin panel behind it.
Duration
Ongoing · live since October 2026
Started
October 6, 2026
Measured on real production traffic, October 2026
Built from 24 exported sales chats and 32 transcribed voice notes
Agent, RAG, leads, human handoff, alerts, attribution, automations and admin panel
Secrets readable only by the server, column by column
CEA Circular sells driver’s license courses in Bogotá, and almost all of its demand arrives on WhatsApp, answered by hand by one sales advisor.
The questions repeat — price per license category, documents, course length, schedules, payment options, the new CALE theory exam — and the official information was scattered and contradictory: three versions of the opening hours, different phone numbers on the flyers, prices only in images.
- — WhatsApp is the main sales channel, handled one chat at a time by a single advisor
- — Most inquiries repeat the same questions: price per category, documents, duration, schedules and payment options
- — Leads arriving at night, on Sundays and holidays went cold, or went to a competitor
- — Business knowledge lived in flyers, voice notes and the advisor’s head, with conflicting versions
A dedicated deployment of the WebMinds WhatsApp AI platform with CEA Circular’s brand, code owned by the client, and eight systems working together.
- + AI advisor · a Gemini agent on the official WhatsApp Cloud API that answers prices, requirements, payment options and the CALE exam, and sends the right flyer or reel from the catalog
- + Knowledge from the real operation · 24 sales chats and 32 voice notes distilled into a pgvector knowledge base, with 39 questions resolved by the owner; personal data never entered the index
- + Lead engine and source attribution · every chat becomes a structured lead with license category and stage, tagged by channel: website, social, QR, Google or Meta ads
- + Human in the loop · one-click takeover pauses the AI; escalation by WhatsApp template and email, with a safety net that catches an “an advisor will contact you” promise even when the model forgets to flag it
- + Admin panel and automations · role-based dashboard, analytics, wa.me links and QR per channel, a website widget, hourly close-out of inactive chats with an approved template, branded emails and a 49-page user manual
Inside the build.
Brand, interface and system — shown as they shipped.
The chat window is the tip.
Underneath is a conversational sales backend: a webhook with more than fifteen decision branches, row-level security on every table, vector search, scheduled jobs on two schedulers, five external integrations and up to five AI calls per message. By effort, roughly two thirds of the work is backend.
We read the operation, not a brochure
24 real sales chats and 32 voice notes surfaced conflicting hours, prices and phone numbers. We turned them into 39 questions for the owner, seven knowledge documents and a 40-question test the agent passes in full.
If the AI promises a human, a human gets notified
When the model writes “an advisor will contact you” but forgets the escalation flag, a future-tense promise detector fires the escalation anyway, by WhatsApp and email, with per-conversation cooldowns and an audit trail.
Attribution without hidden parameters
wa.me links cannot carry hidden data, so each channel injects a short source tag into the prefilled text. The webhook reads it on the first message and always strips it before the AI or the history sees it.
Honest by rule
Domain guardrails stop the agent from inventing prices, schedules or license promises, and an explicit honesty rule keeps it from ever advising a student to misstate information in any official step of the process.
Secrets the API can never return
Credentials live in a single-row settings table; column-level privileges in Postgres mean neither admins nor viewers can read tokens through the public API — only the server role can.
Hybrid AI Coding
A senior engineer owns the architecture, the decisions with the client and every trade-off, while AI agents handle implementation under direction, and an independent end-to-end QA pass gates every change before it ships.
Technologies & Tools
The tech stack powering this solution
Beyond the numbers, we delivered a system that transformed how this client operates.
Every inquiry answered, at any hour
Prices, requirements, payment options and the new CALE exam are answered in about eight seconds, including nights, Sundays and holidays, when the advisor is off.
WhatsApp becomes a pipeline, not a pile
Each chat lands as a structured lead with license category and source channel, so the team knows what students want and where they come from.
AI does the first mile, people close
One-click takeover and guaranteed escalation mean the advisor steps in exactly when a student is ready to enroll, never left waiting on a promise the bot made.
- Apr 2026Proposal and dedicated deployment created on the WebMinds platform
- Aug 20, 2026Human takeover, inactivity close-out, source attribution, branded email, analytics and legal pages; domain live
- Sep 24, 2026Official WhatsApp number connected; training from 24 real chats (40/40); functional QA 16/16
- Sep 25–28, 2026Website widget with live preview; owner-confirmed prices, hours and payments; five reels
- Oct 5, 2026Delivery QA, security package, approved Meta templates and the 49-page manual
- Oct 2026Live with real students; tuning with production traffic, ongoing
Ready to achieve similar results?
Let's discuss how we can transform your business with a tailored digital solution.