We research the company — live, from the open web.
"Poor company research" is on every recruiter's list of what sinks candidates, and it is the most fixable thing on that list. When you add a position, the company is looked up for real: what they sell, who to, what has happened lately — and your questions are built from that world instead of a generic one.
What gets looked up
The product and the market it sits in, roughly how big the company is, recent news that an interviewer would expect you to have noticed, and the themes their interviewers are likely to push on. It runs when the position is created, so the brief reflects the day you started preparing, not a training cut-off.
It changes the questions, not just a sidebar
The brief is fed straight into question generation, so at least one question comes from their actual situation — "you're moving into the enterprise segment this year; how would you handle a customer who…". That is the difference between practising "why do you want to work here" in the abstract and practising it about them.
You can read the brief yourself on the position page, and it feeds the cover letter and the recruiter intro message too.
When there is nothing to research
If you have no posting yet, a realistic job and company can be invented around your target role — and in that case the web research is deliberately skipped. Researching a company that does not exist would produce confident nonsense, which is worse than nothing.
Same principle if the search comes back thin: the brief is dropped rather than padded. Nothing invented is presented to you as fact about a real employer.
The interviewer remembers it, like a person would
In the room it is never quoted at you as a document. The interviewer simply behaves like someone who works there and knows what is going on — which is exactly the level of company knowledge you are practising to match.