AI assistant says a colleague is not in my team
17 per cent of live employees had no manager recorded, so an AI assistant said a real colleague was not in the team. The check, and why names stay out of the index.
Topic
AI
What AI genuinely accelerates in delivery, and where it does not.
14 articles
About this topic
These pieces are about using models in production rather than about models. The distinction matters because almost all of the difficulty sits outside the model: what the system does when retrieval returns nothing relevant, how you tell whether last week’s change made the output better or merely different, and what a confident wrong answer costs when a person acts on it.
The other half is AI inside our own delivery — where it genuinely earns its place, which is comprehension of code nobody documented and the mechanical bulk around a change, and where it does not, which is anywhere the answer has to be right rather than plausible. Both are written from having done it, including the times it produced something that read well and was wrong.
AI assistant said nothing was waiting; four things were
An AI assistant said nothing new was waiting while four requests sat unanswered. Tool descriptions and a date filter hid them. How to test your own.
Best AI image generator? The dearer one got it wrong
We asked two AI image models for a pizza with three of eight slices left. The dearer one drew a perfect 3/8 over a whole pizza. How to choose and check.
AI hallucination: our assistant invented four requests
An AI assistant listed four requests a child never made, though told not to invent. Why it happens and the rule that stops it.
AI for sales data: what it costs and whether to trust it
What it costs to let a sales team question their own data in plain English, why the model never produces the number, and the two defects still open.
Private AI for business: using AI on pricing safely
Can you use AI on your pricing without sending it to an outside provider? Only if it runs on your own machine. The design rules we set, and what is still unproven.
AI reply cut off mid-answer, showing garbage characters
An AI chat reply hit its length limit mid-draft and its control text reached the screen. Why replies get cut off, and what to check in your own AI feature.
AI app works for us, fails for new users: model not found
A model-not-found 404 hit every fresh API key while our own key worked. Availability is per account. What to check before you ship an AI feature.
Excel formula audit stalled at 40 of 362 pricing rules
Hundreds of pricing formulas to check and the review stalled. Sample prices, a flagged shortlist and a separate human sign-off are what finish it.
AI question generator gave twelve questions, not fifteen
A children's learning app returned twelve questions when fifteen were asked and said nothing. Three ways to handle a short result, and the one we chose.
AI feature keeps failing with a network error?
An AI feature failed half the time and blamed the network, but one supplier's blank success stopped the fallback chain. What to check in your own.
Vehicle damage detection: sorting photos without big AI
Reusing 67,000 staff-tagged handover photos sorted new ones with no training. Damage close-ups were only 36% right, so those stay with a person.
AI assistant for business: ask your system in plain English
Can staff ask a business system questions in plain English without seeing what they should not? Yes, if limits sit in what it can reach. What to check.
Will AI make my software project cheaper? Partly
Where AI genuinely speeds up building software, where it quietly costs you more, and the one review rule a supplier should never loosen.