Why does the AI tool we adopted end up unused?
Because it became a second job on top of the work. A tool that takes one typed sentence and drafts the rest has a better chance of surviving a busy week.
6 min read · · The Felesh team
The short answer
You chose a tool, set it up and showed the people you work with how it went. For a fortnight it was used. Then a busy week came, and it was quietly left alone.
The usual cause is not carelessness. The tool became a second job. The work still had to be done, and then the tool wanted its own attention on top.
A second job is the first thing to go when time is short. The fix is a tool that takes what you already do, asks for one sentence, and does the writing itself.
Why does a tool that looked useful get dropped?
Think about how most tools ask to be used. You finish a call with a customer. Then you open the tool, find the right record, and fill in the fields. You write the note. You set the reminder and pick the date. You add the next step.
Each of those is small. Together they are another piece of work, and nobody asked for it on the day the customer called.
On a calm day you do it. On a day with three customers waiting, you tell yourself you will do it tonight. Tonight you are tired. The record falls behind, and a tool with out-of-date records is worse than none, because you stop trusting it.
Then the tool is unused. Nobody decided that. It happened one skipped entry at a time.
Is it a problem with the people?
It is tempting to think so. Perhaps the team resisted change, or someone did not try hard enough.
I would look at the work instead. If a tool only helps when someone does extra typing after every conversation, it asks for steady discipline from people who have a business to run. That is a design problem, not a character flaw.
A quiet tool tells you about the tool and the way it was set up. It tells you nothing about the people who were meant to use it. Blaming them tends to produce a stricter rule, and a stricter rule produces more of the same extra work.
What would make it stay in use?
Three things would help, as far as I can tell.
The tool should ask for less than the job itself. One sentence after the work, in your own words, is about the least it can ask.
It should do the writing. If you say what happened, the note, the date and the next step should come back as drafts rather than blank fields.
And it should not act without you. Drafts that wait for your yes are easier to trust than actions taken in your name. You can read a draft in a few seconds and fix it if it is wrong.
None of this is clever. It simply takes the second job away.
What does that look like with Felesh?
Here is one call, start to finish.
- The call ends. You have promised the customer a quote by Thursday.
- You type one sentence about what happened, as you would say it to a colleague.
- Felesh drafts three things: the note, the promise with its date because you said one, and the next step. Each draft shows its reason beside it.
- You read the drafts. You say yes to those that are right, and you change or leave any that are not. A draft you leave stays waiting.
That is all. You did not open a form or pick a field from a list.
Tasks are simpler still. A task is added directly, with no approval step, because there is nothing to review.
Later you can ask Felesh a question in your own words about your work, such as what you promised a customer. It answers from your own records. That is why the one sentence is worth typing. What you type is what you can ask about later.
What about the messages that keep arriving?
Customer messages are a second kind of extra work. They arrive from the website chat, Instagram, WhatsApp and Telegram, and they are often the same few questions.
In Felesh they arrive in one inbox. Felesh answers from what you taught it, using the documents you uploaded and the instructions you wrote. A tool that makes you answer every message yourself is another job. One that handles the routine ones from your own wording is not.
When Felesh cannot answer, it does not guess. The message goes to “Needs you” with the reason it could not answer. Nobody is notified, so looking at that list is part of your day. When a person joins a conversation, Felesh stays quiet.
Will it start empty?
An empty tool is another reason for abandonment. If you must type in every customer before it is any use, many people never get that far.
Contacts and companies can be brought in from one Excel or CSV file. You see the column mapping and the rows before anything is saved. Nothing goes in that you have not reviewed.
Felesh knows what is done on Felesh, the files you import and what you upload or teach it. It does not read your mail or your drives, and it does not connect to other tools. That is a limit. It also means the records are what you and the people you work with chose to put there.
What if it still goes quiet?
Then look at the work, not at people. A quiet customer or a stalled job is worth a question. A quiet person is not, and no tool should be used to watch anyone.
If you find the drafts have piled up unread, the likely cause is that they are asking for too much. Ask for less. One sentence after the calls that matter is a better habit than a full record that lasts a fortnight.
Where to start
Do not start with the whole business. Pick one kind of conversation, perhaps the calls where you promise something with a date. After the next one, type what happened and see whether the drafts save you the step you usually skip. You can read how after the call works in practice, from the sentence to the drafts waiting for your yes.