Product
Everything you can hand overSee every feature →
Use cases
What Felesh looks like for youAll use cases →
PricingTrust
Resources
Log in

If ChatGPT ran on my work itself, not in a tab beside it, what would change?

The tab we open answers well, but keeps nothing of ours. The difference starts where Felesh writes down what happened itself.

6 min read · · The Felesh team

Where you are

It’s morning. Before anyone arrives, I open a tab and type a few lines. A message for the customer who called yesterday. A summary of last week’s meeting. The answer is good. Sometimes very good.

Then I close the tab. Tomorrow I open it again and have to explain from the start who this customer is, what we promised them and where the work stands. Every time, I am the one describing my work to it.

And the things that matter never reach that tab. The call my colleague took yesterday. The promise made across the desk. The message that came in on the site at night. They stay in people’s heads, or in a spreadsheet only one person updates.

Sometimes someone on the team asks, “Where are we with that customer?” and I don’t know. Not because nothing was done. It was done, but it wasn’t written anywhere I can see. So I call, or we set up another meeting.

My question isn’t whether the tool is clever. It’s this: if the same intelligence ran on our work itself, not beside it, what would change?

Fig. 1 A tab beside the work answers well but keeps nothing, while Felesh running on the work writes down what happened itself.

How it usually goes

Most of us have three ways: the meeting, the sheet, and that tab.

The meeting is for finding out where the work stands. An hour of several people’s time goes on everyone saying what they did. Someone takes notes, and the notes get lost too.

The sheet is so nothing is forgotten. But a sheet is only right when someone fills it in. A tired person after a call doesn’t. So the sheet is always a week behind.

The tab is for writing faster. But it knows only what we paste in. It keeps nothing, hands nothing to anyone, and has forgotten by tomorrow. Each time we copy, explain, and carry the answer somewhere else.

All three share one thing: the writing is on people. And because people are busy, what is written always trails what actually happened.

How it goes in Felesh

  1. After a call, I type one sentence into “Felesh”, the way I’d tell a colleague.
  2. Felesh drafts a note of the call, and a promise with its due date if I gave one.
  3. I open Review draft and see what will be written and why. Nothing is written without my yes.
  4. Now the record is current without anyone writing a summary: the note and the promise with its date are there.
  5. Tomorrow I don’t explain anything from scratch. If I ask “Where are we with Reza?”, Felesh answers from this record, not from whatever I pasted somewhere today.

What is still ours. The decision. Felesh drafts, keeps the record, and sends what it cannot answer to Needs you. Saying yes, picking up the phone and making promises are still our work.

That is the difference. A tab beside the work needs the work described to it every time. A tool on the work already has it. Felesh knows it, drafts it, and we come in only where we are needed.

To start, we don’t have to change everything. One real piece of work is enough: yesterday’s call, the promise due Saturday. If it works on that, the rest follows.

Next question Can AI run my business for me? No. It can keep the business known and do the work you approve, while the judgement calls wait for you. Here is where the line sits.