AI work you can check and approve
Each agent task has a Show me why view and shows its confidence where the agent recorded one; you approve it, reject it, or note that it looks wrong.
Live in partWhat does Agent runs do? Hand repetitive work to agents while keeping accountability, with a record you can show an auditor of what each agent did and who approved it.
How it works
Agents post their work to one queue, split into 'Waiting on you' and 'Working now'. Each task shows its status in plain words, a 'Show me why' button, a readable list of changes, and what data and AI model it used. 'Looks wrong? Fix it' records a correction. Anything with an amount opens for review rather than a one-tap approval.
Confidence labels like 'sure' or 'not sure' appear only on AI suggestions, not fixed rules, and anything with an amount opens for review rather than a one-tap approval.
Step by step
Here is how an accounts executive or finance head works with agent output in a normal day.
- An agent does a piece of work. It might draft a collections reminder, match a bank line or flag a bill. Its work is posted to one queue rather than acted on.
- Open the agent panel. It has two lists: Waiting on you and Working now. When there is nothing for you it says "Nothing needs you right now". A line shows how many jobs finished today.
- Open a task. You see what the agent proposes, the record it concerns, its status in plain words, and a confidence figure when the agent recorded one. A missing figure is never shown as zero.
- Press Show me why. It expands what the agent looked at, with a readable list of changes, and shows which data and which AI model it used, or says that no AI model was used.
- Approve, or reject with a reason. Approving accepts the proposal. Rejecting sends it back and asks for a reason, which is kept with the run. When you decide, the record shows your name plus Kinnow.
- If it looks wrong, say so. "Looks wrong? Fix it" records your note with the run. It does not change the run, but anyone looking later can see what people said.
- Look back later. The Agent runs page has Active and History tabs, so you can open a run days or weeks afterwards. A failed task has a Try again button.
Ask us to show an agent run from start to approval on your own sample data.
What you see on screen
The agent panel opens beside any screen. The Agent runs page and each task have their own pages.
| Where | What you see |
|---|---|
| Agent panel | Waiting on you, Working now, and a list of your agents showing how many need you, how many are turned off, and when each last ran. |
| Agent runs, Active and History | Active holds runs queued, running or waiting for you. History holds finished, failed or cancelled runs. Click any row for detail. |
| A task | A status pill (Waiting for you, Done or Failed), what the agent proposes, confidence if recorded, Open the record, Approve and Reject. |
| Run details | Which AI model, how much text in and out, the cost, how long it took, who started it and who approved it. An agent that follows fixed steps says it used no AI model. |
| A run page | For agents that rank or classify: the headline answer, a confidence or track-record badge, Show me why and when it was recorded. It is read-only. |
Who uses it
Anyone who approves; anyone who answers for the result.
| Who | When | What they do |
|---|---|---|
| Accounts or operations executive | Daily | Works down Waiting on you and approves or rejects. |
| Finance head | Weekly | Looks at rejections, corrections and failed runs. |
| Administrator | When trust changes | Turns an agent off; a stopped agent's runs say so. |
| Auditor | At review | Opens a run to see what the agent decided and who approved it. |
Ask us to show how a rejected run is recorded.
Related
- Test business rules before enforcing them: rules that say what is allowed, tried out before they apply.
- One approval inbox: where people approve.
- Build and tune your own agents safely: your own agents.
- All agents: what each agent takes off a desk.
Questions
What is Agent runs in Kinnow?
Each agent task has a Show me why view and shows its confidence where the agent recorded one; you approve it, reject it, or note that it looks wrong.
What changes for the business?
Hand repetitive work to agents while keeping accountability, with a record you can show an auditor of what each agent did and who approved it.
Is it available today?
It is live in part: the core is built and parts are completing with early customers.
Can I see why an agent made a decision months later?
Yes. Each run is kept as a record. Open it from the Agent runs History tab and press Show me why to see what it looked at and, where recorded, its confidence.
What if the agent cannot finish?
The task shows Failed with the reason, and you can press Try again. If the AI part cannot run, the run says it did only the part that does not need AI.
More in the AI agents that do the work
Claim every rupee of GST input credit
Purchase bills are matched against your GSTR-2B file from the GST portal; your team only handles the exceptions.
AI agents that do the work · Collections AgentOverdue invoices chased for you, politely
Overdue customers get reminders that grow firmer as days pass, each one drafted for a person to approve before it goes out.
AI agents that do the work · GatherForms for applications, surveys and sign-offs
Collect applications, data, anonymous feedback or sign-offs, each with its own AI helper and privacy rules.
AI agents that do the work · Close Checklist agent + 12-gate period closeMonth-end close that runs itself
Kinnow runs the month-end checks, plans the fixes for what fails and hands you one close pack. A person closes the month.
AI agents that do the work · Buy somethingFrom supplier quote to purchase order
Drop in a supplier's quote as a PDF or photo; Kinnow drafts the purchase order and flags doubts.
Tell Kinnow what slows you down.
Answer three questions. Get the applications, the agents and a two-week plan for your business.