An AI inventory manager
Running out of a common part in the middle of a job costs a day, and over-ordering to avoid that ties up cash on a shelf. Both are the same failure: finding out at the wrong time. This job keeps the count current and puts a reorder in front of you while it is still early.
Concretely
- Tracks stock levels against your inventory records rather than a memory of them.
- Watches what moves and what has not moved in months.
- Flags the lines running low while there is still time to order.
- Recommends reorders with the quantity and the supplier already filled in.
- Produces the inventory report nobody has time to build by hand.
An annual count is a guess for eleven months of the year
Most businesses count stock once a year, and the number is wrong by February. Everything between counts runs on what somebody believes is on the shelf, which is why the part that stops a job is almost always a common one. A count that stays current is worth less for its accuracy than for the fact that it is there on the day the decision gets made.
Cash sitting on a shelf is the other half of the problem
Over-ordering is the natural response to running out, and it is expensive in a way that never shows up as a line item. Money in stock that has not moved in a year is money that was needed somewhere else, and nobody notices because nothing goes wrong. Watching what does not move is as much of this job as watching what does.
It recommends, a person orders
Buying is a commitment of money, so it is not something this job does on its own. It puts the reorder in front of a person with the working shown. A count that does not reconcile, or anything that looks like shrinkage, is flagged for a person rather than quietly corrected, because a corrected discrepancy is a hidden one.
The jobs this comes from
Sanaf holds thirty five written job descriptions. These are the ones behind this page.