For AI agents: the complete documentation index is available at https://docs.clickmax.io/en/llms.txt, the full documentation bundle is available at https://docs.clickmax.io/en/llms-full.txt, and this page is available as Markdown at https://docs.clickmax.io/en/features/contacts/playbook-test.md.
  • English

      • Test the playbook before using it in a meeting

      The Test tab in the playbook editor simulates a conversation with a client and runs the same copilot that would join the real meeting. You write what the client would say, and the screen shows two things: the cues from your keywords, which appear instantly, and the copilot suggestion, which arrives a few seconds later — exactly as the seller would see them.

      The test uses what is on screen right now, saved or not. You can tweak a stage, go back to the Test tab and see the effect before clicking Save.

      Role: workspace owner, admin or editor.

      Before you start

      • Write at least one stage. An empty playbook makes the copilot answer with no method at all.
      • To test with your own AI agent, it must be active under Settings › Connections › AI Agents. With no agent selected, the test uses the Clickmax default model.

      Set up the scenario

      1. Open the playbook and click the Test tab.
      2. Under Agent that answers, pick the AI agent. Leave Default model to test without your own agent.
      3. Under Start from a scenario, click the selector to open the Pick a scenario window: nine ready-made tough clients, each stalling at a different stage of the method — no pain in sight, price up front, already uses a competitor, stalls the close, the person on the call can't decide, technical and sceptical, burned before, wants it yesterday, and wants it badly with no budget. Pick one and click Use scenario. The conversation opens with ten lines already exchanged, and the goal and persona come filled in. To write it yourself, pick Blank conversation.
      4. Adjust the Meeting goal and Who is on the other side if you want. Both go to the copilot alongside the playbook — that is what makes the suggestion talk about the right client.

      Have the conversation

      1. In the selector below the conversation, choose whether you are speaking as Client or as Seller.
      2. Type the line and press Enter (Shift+Enter for a new line) or click the send button.
      3. Only the Client line wakes the copilot — the same rule as in the meeting. Write as Seller to build context without asking for a suggestion.
      4. To ask outside that rule, click Ask for a suggestion. It is the equivalent of the manual request inside the meeting.
      5. Clear wipes the conversation and starts over.

      What shows up in the column:

      • Trigger: "<word>" — a playbook keyword was said. The card is the same one the seller would see in the meeting, in the same colour.
      • Copilot — the AI suggestion, with the title and tone it chose.
      • No suggestion — the copilot looked and found nothing worth saying. When the model explains why, the reason shows underneath. This is a normal answer: most turns in a meeting don't call for a card.

      Speak instead of typing

      1. Click the microphone button next to the field and allow microphone access when the browser asks.
      2. Say the client's line. The footer shows the seconds recorded.
      3. Click the square to stop. The audio becomes text and lands in the field, not straight into the conversation — read it, fix it if needed, then send.

      If the browser has no microphone available or you deny access, the button disappears or the footer says so; type the line instead.

      FAQ

      Is the test conversation saved? No. It only exists while the tab is open; leaving the playbook or clicking Clear wipes it. None of it becomes a contact, an opportunity or an activity.

      Does the test count as a meeting? No. It creates no room, records nothing and does not show up in the contact's history.

      Will the suggestion be identical in the real meeting? The keyword cue will: it is always the text you wrote. The AI suggestion varies — it is a model answering, and the same request may come out in different words. What the test guarantees is that the copilot is receiving the right method.

      Why did the copilot stay silent? Either it looked and found nothing to say (the reason shows in the No suggestion card), or the chosen agent is paused, in draft, or has no key configured. Check under Settings › Connections › AI Agents.

      Can I test without saving? Yes — that's the point. The test reads the draft on screen. Remember to save afterwards, otherwise the real meeting keeps the old version.