Your first job, step by step¶
From a connected AI to your first approved record, in five steps, with what you see on screen at each one.
Your first job is the moment the loop becomes real: you ask your AI for a piece of compliance work, it drafts it, Kanonik checks it, and you approve it onto your record. Your risks make a good first job, because the controls and policies you need follow from them.
1. Connect your AI and wake it up¶
On the Connect your AI page, follow the steps for your client. Then copy the prompt under Activate and paste it into your AI. It asks your AI to read your workspace profile through the Kanonik tools before asking you anything. That first call flips the page from "Waiting for your AI's first call" to Connected. Choose Go to my workspace.
If the page never flips, see Connecting your AI.
2. Pick the job¶
You can type your own request, or open Use cases in the left-hand menu. It lists jobs such as Find the risks that are actually mine, Answer a security questionnaire and Hand a pack to an auditor. Pick one and choose Copy prompt for your AI. Kanonik writes the prompt from what your workspace already knows. Paste it into your AI to run it: nothing on that page runs on its own.
A good first job is your risks. "Find the risks that are actually mine" asks your AI to look at your environment, draft a risk register and send it to you for approval. How risk discovery works explains what happens along the way.
3. Answer your AI's questions¶
Your AI does the work in its own chat. It may ask you about your company, your systems or your vendors. It reads your code or your cloud only with your own access, on your side. Kanonik never opens any of it.
Kanonik's safety check reviews each draft before it reaches you. Your AI can revise a draft the check is unhappy with and submit it again, and some items may need your input first. When you and the safety check disagree covers what the results mean.
4. Decide on the request¶
The finished work arrives as one or more approval requests, depending on its size and what it contains. You find it under Needs you in the left-hand menu, which shows how many requests are waiting on you. The link to the request is also emailed to you once your sign-in email address is verified.
Open each item, read it, and approve it, send it back with your reason, or reject it. What you approved lands on your last decision; an item that needs a second person waits until a colleague seals it. Deciding on an approval request, item by item covers every button.
Nothing goes on your record until you decide. Saying "yes" in your AI's chat does not count.
5. See it on your record¶
What you approved shows in the Workspace: that is the place to check whenever you are unsure whether something landed. Approved risks, controls and policies also appear in the Audit log, with the name of the person who approved them and the time.
What to do next¶
- Ask your AI for the controls that treat the risks you approved, then the policies those controls need.
- Invite the people who will approve or write with you. See Roles and inviting your team.
- If a buyer has sent you a security questionnaire, pick Answer a security questionnaire under Use cases.
More help¶
Browse every article in the Help center, where you can also ask the Kanonik assistant directly. For anything else, email [email protected] and a person who works on the product answers.