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 hours·intermediate
Tools: Botpress or Voiceflow, WhatsApp Cloud API, a knowledge-base/RAG backend, analytics dashboard (built-in or custom)
Market relevance — share of job ads asking for this
What employers mean
You should be able to…
- Design intents and a knowledge base that covers real user questions, not just happy paths
- Deploy a bot on WhatsApp or web with a working handoff to a human agent
- Wire a knowledge base (docs, FAQs) into the bot so answers are grounded, not hallucinated
- Track analytics (deflection rate, handoff rate, common failed queries) post-launch
- Handle multi-turn context (the bot remembers what the user already said)
- 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
Practice
WhatsApp support bot for a D2C brand's order-status and returns queries
Build a bot (Botpress or Voiceflow) deployed to a WhatsApp sandbox that answers order-status and returns-policy questions grounded in a small knowledge base (a returns policy doc + sample order data), with a clear handoff to 'talk to a human' when confidence is low, and basic analytics logging of query volume and handoff rate.
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
- Screen-recording of the bot running on WhatsApp
- Public repo with bot flow export and README
- Analytics sheet/dashboard screenshot showing logged interactions
Interview
Questions you'll get asked
- Walk me through how you'd design intents and handoff logic for a support chatbot.
- How do you ground a chatbot's answers in a company's actual knowledge base instead of the model's general knowledge?
- What analytics would you track after launching a WhatsApp support bot?
- How do you handle a multi-turn conversation where the user's intent changes halfway through?
- Tell me about a fallback strategy you built for questions the bot can't answer.