The 20-Minute Company Dossier
Most candidates Google a company for five minutes, skim the homepage, and walk into interviews knowing less than the receptionist. You can do better in 20 minutes with a structured pass and an AI assistant doing the synthesis.
The goal is not to know everything. It is to know enough to ask sharper questions than the other 12 candidates, and to tailor your application so it reads like you actually want this specific job — not any job.
Split your 20 minutes like this: 5 minutes on what the company actually does and how it makes money, 5 minutes on recent news and strategy, 5 minutes on the team you would join, and 5 minutes on the people who will interview you. AI handles the heavy reading. You handle the judgment.
What the Company Actually Does
Start with the boring stuff: revenue model, customer segments, competitors. Most company websites are written for customers, not job applicants, so you need to translate.
Open the company's homepage, "About" page, and most recent product launch page. If they are public, grab the latest investor letter or 10-K summary. Paste the raw text into Claude or ChatGPT and run this:
You are helping me prep for a job interview. Based on the
text below, answer in 5 short bullets:
1. What does this company actually sell, in plain English?
2. Who are their primary customers and what do those
customers care about?
3. How do they make money (subscription, transaction,
ads, services)?
4. Who are their top 2-3 competitors?
5. What is the single biggest risk to their business
right now?
Be skeptical. Strip marketing language. If something is
unclear from the text, say "unclear" rather than guessing.
[paste text]
The "strip marketing language" line matters. Without it, the model will parrot phrases like "revolutionizing the future of work" back at you. You want the version a junior analyst would write.
If the company is private and information is sparse, ask the model to flag what is missing. Then go fill those gaps using LinkedIn, Crunchbase, or a quick web search. Do not pretend you know things you do not — interviewers smell that immediately.
Recent News and Strategic Direction
A company's last 90 days tell you more than its mission statement. Layoffs, funding rounds, new product launches, leadership changes, expansion into new markets — these are the topics your interviewer is thinking about right now.
Pull the company name into a news search and grab headlines from the past three to six months. Drop them in and prompt:
Here are recent news items about [company]. Cluster them
into 3-4 strategic themes (e.g., "expanding into
enterprise", "cost cutting", "AI product push"). For each
theme, tell me:
- What it suggests about leadership priorities
- What it could mean for the role I'm applying to:
[paste job description]
- One thoughtful question I could ask in an interview
about it
[paste headlines + first paragraphs]
That last bullet is gold. You walk in with three to four real questions tied to real news, and you instantly sound like someone who did the work. Most candidates ask "what does success look like in this role" — yours will sound like a peer.
Mapping the Team You Would Join
Before you talk to anyone, sketch the team. LinkedIn is your primary source. Search the company, filter by department or location, and skim the first 15-20 profiles of people in your target function.
What you are looking for: average tenure, where people came from before, where people go after, ratio of senior to junior, any obvious gaps. A team with five-year average tenure and senior-heavy ranks is a different bet than a team where everyone joined six months ago from three different acquisitions.
Paste a list of titles, tenures, and previous companies into the model:
Here is a snapshot of the [team name] at [company]. Help
me read it:
- What is the seniority distribution and what does it
suggest about how decisions get made?
- Are there patterns in where people came from?
- Are there roles missing that you would expect on a team
this size?
- What questions should I ask a hiring manager based on
this structure?
[paste data]
You can also reuse the workflow from /courses/ai-job-search-career, which has a longer module on team-mapping if you want to go deeper than this chapter covers.
Researching the People Who Interview You
Once you have the names, spend the final five minutes per interviewer — not five minutes total. This is the highest-leverage part of the entire process.
Pull their LinkedIn profile and any public talks, podcasts, blog posts, or conference appearances. For senior leaders, check if they have written publicly about their philosophy or priorities. For engineers, scan GitHub. For PMs, check Substack and X.
Then run a focused prompt:
Here is the LinkedIn profile and other public content for
[name], who will interview me for [role] at [company].
Give me:
- A 3-sentence summary of their career arc
- 2-3 topics they seem to care about based on what
they have written or shared publicly
- 1 question I could ask them that connects their
background to the role
- 1 thing to AVOID saying based on signals in their
profile (e.g., they left a previous company that I
should not enthusiastically reference)
Be specific. No generic advice.
[paste content]
The "avoid saying" line saves you from landmines. If your interviewer recently left a company in a public dispute, you do not want to praise it in your first answer.
Build a Reusable Dossier Template
Do this once, save the structure, and reuse it for every company. A simple markdown file works:
# [Company] — [Role]
## Business model (3 lines)
## Strategic themes (last 90 days)
## Team snapshot
## Interviewers
- [Name 1]: background, topics, question to ask
- [Name 2]: ...
## My 3 questions
## Connection to my background
After five companies, you will have refined this into something tighter than any generic interview-prep blog. After ten, you will start spotting industry patterns that make you sharper in every conversation.
Twenty minutes of focused research beats two hours of unfocused scrolling. The candidates who get offers are not the ones who memorized the "About" page — they are the ones who walked in sounding like they already worked there.