RoleGPSThe free map to your AI job
All capabilities · AI automation & no-code

Deploy a support/sales chatbot on WhatsApp or web

Voiceflow/Botpress/custom: intents, knowledge base, handoff to human, analytics.

~15 focused hoursintermediate
Explore 3 tools for this project
Market relevance

Which roles ask for this — and how often

Share of job postings in India, per role, that name this capability.

AI Automation Specialist6%
What employers mean

You should be able to…

  1. Design intents and a knowledge base that covers real user questions, not just happy paths
  2. Deploy a bot on WhatsApp or web with a working handoff to a human agent
  3. Wire a knowledge base (docs, FAQs) into the bot so answers are grounded, not hallucinated
  4. Track analytics (deflection rate, handoff rate, common failed queries) post-launch
  5. Handle multi-turn context (the bot remembers what the user already said)
  6. Set up fallback responses for out-of-scope questions instead of a dead end

Needs first: Add LLM steps to business automations

Learn — free, link-checked

The few resources that matter

Read · beginner · 40 min · docs.voiceflow.com

Voiceflow documentation

Covers Voiceflow's visual flow builder and knowledge base — the other chatbot tool JDs name alongside Botpress. — Voiceflow
Read · beginner · 45 min · botpress.com

Botpress documentation

Free, current docs for building an intent/knowledge-base bot with human handoff on a widely-used platform. — Botpress
Read · intermediate · 40 min · developers.facebook.com

WhatsApp Cloud API — Get Started

The official path to actually deploying a bot on WhatsApp, the channel most Indian support chatbots ship on. — Meta
Read · intermediate · 60 min · docs.n8n.io

Advanced AI in n8n

Official docs for wiring LLM nodes, agents and memory into a workflow — the exact skill 'AI + automation' JDs ask for. — n8n
Tools for practice

Choose a tool for the job

Start with one tool for each part of your project. You don’t need to learn them all.

Go to the practice brief

3 tools to explore

Botpress

Automate · Build

Prototype a chatbot with a knowledge source and test its conversation paths.

Visit toolGetting started

Voiceflow

Automate · Build

Design a conversational workflow and test the handoffs between intents and actions.

Visit toolGetting started

Selected for practice by RoleGPS. No paid placements or affiliate links.
Access can change; check the details before starting.

Think your tool is a good fit? Let us know

Practice

WhatsApp support bot for a D2C brand's order-status and returns queries

Build a bot (Botpress or Voiceflow) on the WhatsApp Cloud API test number — a free Meta developer account gives you one that messages up to five verified test numbers with no business verification. Ground it in a small knowledge base you write yourself: a returns-policy document and a sheet of sample orders. Add a 'talk to a human' handoff when confidence drops, and log query volume and handoff rate so you can state how often it actually deflected rather than guessing. If Meta's test-number terms have changed by the time you start, ship the same bot as a web chat widget instead — the knowledge base, handoff and metrics are the skill being assessed, not the channel.

Start from

WhatsApp Cloud API test number — free Meta developer account, no business verification; a web chat widget is an equivalent fallback if that changes

Milestones
  1. Get the test number talking to your bot and echoing a message back · ~2.5h
  2. Write the returns-policy knowledge base and sample-order sheet, and ground both intents in them · ~3.5h
  3. Build the multi-turn order-status flow that remembers what the user already said · ~3.5h
  4. Add the confidence-based human handoff and log query volume and handoff rate · ~3.5h
Done when
  • Bot is connected to a live WhatsApp sandbox (Meta test number or provider sandbox) and responds
  • At least 2 distinct intents (order status, returns policy) are handled with grounded answers
  • A human-handoff path exists and is triggered by a defined confidence/keyword rule
  • Basic analytics (query count, handoff count) are logged and shown in a simple dashboard or sheet
Prove it

Evidence a recruiter can check

  • A recording of a real conversation on the WhatsApp test number, including the moment it hands off to a human
  • Both intents answering from your knowledge base with a source line showing which document each answer came from
  • A deflection log — query count, handoff count and the rate — with the failed queries listed rather than summarised
  • The exported bot flow with the out-of-scope fallback path visible instead of a dead end
Signal it

Deployed a WhatsApp support bot grounded in a returns-policy knowledge base with a confidence-based human handoff, instrumented to report deflection and handoff rates.

Interview

Questions you'll get asked

  1. Walk me through how you'd design intents and handoff logic for a support chatbot.
  2. How do you ground a chatbot's answers in a company's actual knowledge base instead of the model's general knowledge?
  3. What analytics would you track after launching a WhatsApp support bot?
  4. How do you handle a multi-turn conversation where the user's intent changes halfway through?
  5. Tell me about a fallback strategy you built for questions the bot can't answer.
See where you stand for AI Automation Specialist