14 days, one bot: a day-by-day plan for a chatbot free trial
A good chatbot free trial test plan spreads across the full 14 days: connect real channels and data in week one, run real conversations and edge cases in week two, and always confirm you can export your flows and contacts before deciding. Define what 'good enough' means for your business before you start, not after.
- Decide your success criteria before day one — what does 'this works for us' actually mean for your business?
- Connect a real channel (your own WhatsApp number or website) early, not a demo sandbox — trial behaviour should match production.
- Test edge cases deliberately in week two: rude messages, off-topic questions, and requests to talk to a human.
- Ask support how exports work before you need to leave — flows, contacts and templates should be portable.
- Involve the person who will actually run the bot day to day, not just the person evaluating it.
Set your success criteria first
Before opening any trial account, write down what 'this is working' means for your specific business. Vague goals like 'better customer service' are hard to test against. Specific ones are not, and writing them down before you start stops the final decision being shaped by whichever demo went best or worst on the last day.
- Answers our top 10 FAQs correctly without escalating to a human.
- Books an appointment end-to-end in under two minutes on WhatsApp.
- A lead created by the bot appears correctly in our CRM pipeline within a minute.
- Handles a rude or off-topic message without inventing an answer.
- Works acceptably in the language mix our own customers actually use: Hindi, English or Hinglish.
Share this list with everyone who will test the trial, not only the person managing the evaluation, so the whole team is judging the same bar on day 14 rather than forming separate impressions.
Week one: connect real channels and real data
Week one is about making the trial resemble production as closely as possible. A demo sandbox number and made-up FAQs will not tell you anything useful about how the bot behaves with your own customers, so resist the temptation to 'just try it quickly' on throwaway data.
| Day | What to do |
|---|---|
| 1–2 | Sign up, connect your actual WhatsApp number or website widget, not a demo number. Upload your real knowledge base documents (price list, FAQs, policies). |
| 3–5 | Build one real flow end-to-end for your top use case (booking, lead capture, or order status) using the Flow Builder. Use Preview / Try-it to dry-run it before it goes live. |
| 6–7 | Invite a colleague to message the bot as a real customer would, without coaching them on what to ask. Note every place it stumbles. |
By the end of week one you should have one real flow live on a real channel, answering from your own documents, not a generic demo script that never touches your actual data.
Week two: stress-test and check the edges
Week two is where most trial decisions are actually made, because it is where a bot's real limits show up. A bot that handles the happy path well in week one can still fail badly on an unclear message, an unhappy customer, or a handoff to a human, so deliberately go looking for these cases rather than hoping they do not come up.
| Day | What to do |
|---|---|
| 8–9 | Test edge cases: an off-topic question, a rude or unclear message, someone asking for a discount you do not offer. Confirm it does not make things up. Check its knowledge-base grounding. |
| 10–11 | Test the human handoff. Can a live-chat agent take over cleanly when the bot cannot help? Does the bot pause correctly once a human replies? |
| 12 | Check reporting: open Analytics and see if the numbers match what you would want to report to your own manager or partner. |
| 13 | Test on a second channel if relevant (Instagram DM, website widget) to see if the same flow logic reuses cleanly. |
| 14 | Score against your day-one success criteria and use the decision table below. |
Questions to ask support during the trial
- What exactly happens to my data, flows and contacts if I do not continue after the trial?
- Is there a limit during the trial that will not exist on a paid plan, or is it genuinely full-featured?
- How do I export my flows, contacts and templates if I switch platforms later?
- What happens if my AI provider key runs out of credit or errors: will I be alerted, or will the bot just go silent?
- Who owns the WhatsApp number and Business Account I connect, and can I take it with me?
- If I add a second team member during the trial, can I give them limited access instead of the full owner login?
- What is actually backed up during the trial period, and how often?
How support answers these, quickly, clearly, without dodging, is itself useful trial data about what post-purchase support will be like once you are a paying customer and less of a priority.
Testing this on Convo360
Convo360's trial is 14 days with every feature, no credit card required (fair-use limits: up to 100 messages and 10 AI calls of 3 minutes in any 24 hours, one WhatsApp number), so there is no artificial feature gate to work around during the test plan above. You can connect your real WhatsApp number via embedded signup, build in the Flow Builder with Preview / Try-it, set up a knowledge base for RAG answers, and check Analytics and CRM pipeline exactly as they would run on a paid plan. Because you own your WhatsApp Business Account and API keys under BYOK, nothing is locked into the platform even during the trial. See avoiding vendor lock-in for the fuller picture. You can also invite a colleague under a limited staff role during the trial itself, so week-two testing does not require sharing your one login.
A decision table for day 14
| What you found | What it usually means |
|---|---|
| All success criteria met, no major stumbles in week two | Move to a paid plan and plan a soft launch to a slice of real traffic first. |
| Most criteria met, one specific gap (a language, an edge case, a report field) | Ask support whether the gap is a setting you missed or a genuine limitation, then decide. |
| Core answers are wrong or the knowledge base is not grounded properly | Revisit your source documents and system prompt before blaming the platform, then retest before deciding. |
| Handoff to a human did not work reliably | Treat this as a blocker: a bot that traps a frustrated customer is worse than no bot. |
| Support was slow or vague answering your ownership/export questions | Weigh this heavily: it previews what post-sale support will feel like. |
Deciding at the end of day 14
Go back to your success-criteria list from day one and mark each item met, partially met, or not met, with a short note on why. Involve whoever will run the bot daily in this final call. Their read on 'would I trust this in front of a customer' matters as much as the feature checklist, and they are the one who will live with the decision after the evaluation is over.
Step by step
- Write your success criteria (before day 1)List 3–5 specific, testable outcomes for your business, not vague goals, and share them with everyone who will test.
- Days 1–2: connect real channelsSign up and connect your actual WhatsApp number or website widget, and upload your real knowledge base documents.
- Days 3–5: build one real flowBuild your top use case end-to-end in the Flow Builder and dry-run it with Preview / Try-it before going live.
- Days 6–7: get an outside testerHave a colleague message the bot as a real customer would, uncoached, and log every place it stumbles.
- Days 8–9: test the edgesSend off-topic, rude or unclear messages and confirm the bot stays grounded in your knowledge base instead of inventing answers.
- Days 10–11: test human handoffConfirm a live-chat agent can take over cleanly and that the bot pauses correctly once a human replies.
- Day 12: check reportingOpen Analytics and confirm the numbers are ones you would actually want to report upward.
- Day 14: score and decideCompare results against your day-one criteria using the decision table, then commit to a plan or extend testing on the specific gap.
Example conversation
Start your free 14-day trial
Start free in minutes — no code, no credit card. Bring your own AI keys and pay no per-message markup.
Frequently asked questions
Do I need a credit card to start a Convo360 trial?
No. The 14-day trial is free, unlocks every feature on the platform, and does not ask for a credit card at sign-up, so there's no accidental charge to remember to cancel if you decide not to continue afterwards.
Can I test with my real WhatsApp number during a trial?
Yes, and you should — testing with a demo sandbox number can hide issues that only show up with your real Business Account, templates and customer flows.
What should I do if the trial period is not enough time?
Use the structured day-by-day plan to prioritise the highest-risk tests early (real channel, real data, edge cases) so the most important information is gathered even if you cannot test everything.
Can I export my work if I decide not to continue?
Ask this directly during the trial — a trustworthy vendor should let you export flows, contacts and templates regardless of your decision. See also chatbot vendor lock-in and what to check before committing.
Should I test with my own team or real customers?
Start with a colleague acting as a customer to catch obvious issues safely, then, if comfortable, soft-launch to a small slice of real traffic before switching over fully.
Related reading
Available across India — see WhatsApp & AI calling automation by state and city.