> ## Documentation Index
> Fetch the complete documentation index at: https://docs.hired.tools/llms.txt
> Use this file to discover all available pages before exploring further.
# Research and interview prep
> Building a company record that is worth reading the night before, and turning it into a prep sheet the morning of.
Company research has one purpose: to be there the night before an interview, when you do
not have time to do it. Everything below is aimed at that moment.
## Researching a company
> Research Helios and put it on their record.
[`research_company`](/workflows#research-a-company-into-the-crm) reads what is already on file — including
every application and contact you have there — tells you what is missing, then writes the
result back.
The writing-back step is the one that needs care, and the workflow spells it out:
`update_company` **replaces** the notes field. The combined text has to be written back
whole — what was already there plus what is new. Sending only the new part deletes the
old research, which is the worst outcome available in this part of the product.
### What is worth recording
Not a company profile. The four things you will actually want later:
* **What they do and how they make money.** In your words, not theirs.
* **The interview loop**, if you know it. Rounds, who is in them, what each one is for.
* **Who you know there.** The single highest-value line in the record.
* **Why you do or do not want this** — the honest version, including the reservations.
### Two details that matter
**Set `website` to their own domain.** A Greenhouse or Ashby link is the job board, not
the employer. The domain is what puts their logo on your pipeline, and it costs nothing to
include.
**Mark unconfirmed things as unconfirmed.** An assistant is told not to invent facts about
a company, and to say when it is working from what you told it rather than from something
it knows. Research that quietly mixes the two is research you cannot trust in three weeks.
## Preparing for a round
> Get me ready for my system design round at Helios.
[`prep_for_interview`](/workflows#prepare-for-an-interview) gathers four things before writing
anything: the application with its full posting **and its whole timeline** — what has
already been said matters more than the posting does — the company research on file, the
contacts at that company, and your own evidence for the two or three themes the posting
leans on hardest.
Then it gives you:
The three things, read from the posting and the timeline together.
With the specific numbers from your brain. If a number is not on file, the story comes
without one and you are told to check — never a metric invented for the occasion.
Including the weak spots in your own history for this role, which is the half nobody
prepares.
Drawn from the company research rather than generic ones.
Plus anything in the timeline worth following up on or referring back to — the thing the
recruiter mentioned three weeks ago that you would otherwise have forgotten.
If the research on file is thin, it says so and offers to run `research_company` first.
## Afterwards
The interview is new material about you, and it belongs back in the brain. Two minutes
while it is fresh:
> That was the system design round at Helios. They asked about scaling the billing
> pipeline and I walked through the Go rewrite — went well. The follow-up on multi-region
> I fumbled. Interviewer was Dana, staff engineer, seemed to own the platform.
That lands as a timeline entry on the application, a contact if Dana is new, and — the
part people miss — material in the brain about how you answer that kind of question. The
next prep sheet is better because of it.
Why the loop back matters more than it sounds.