The same three pieces — a trainable agent, a live inbox, honest metering — solve very different problems. Here's how teams put Tendabot to work.
Train the bot on your shipping, returns and sizing policies. It answers instantly on a cheap, fast model, captures the customer's email, and hands tricky cases to a human — while the whole thread stays in one inbox.
Each client is its own workspace: separate bots, knowledge, wallet and billing. Put budget models on low-traffic clients and flagships where it matters, and switch workspaces from the top bar. White-glove service, platform economics.
Feed the bot your documentation and changelog. It resolves how-to questions on the spot, escalates bugs to a human, and every conversation is searchable — with the model bill itemised per reply.
Clinics, salons, restaurants, repair shops: the widget takes bookings and FAQs all night, and the WhatsApp-ready architecture means your number can plug into the same bot and inbox when you connect it.
The pre-chat form captures name and email with a consent receipt. When visitors leave, follow up by email in one click — replies land back in the same conversation, answered by AI or by you.
Start on 300 free credits, prototype against 42 models, bring your own API key when you scale, and keep a to-the-token record of what every experiment cost. No vendor lock-in to regret later.
Not sure where you fit? Ask the demo bot — it runs on this exact platform.