“I read your website” is the weakest thing you can say in an interview. Good company research finds the specific 3 facts that actually matter — a product decision, a market move, a recent change — and turns them into talking points and questions only an insider would think to ask.
These 12 templates do that. Each one names what you want, so the model returns specifics instead of a corporate-brochure summary.
TL;DR
- Use a model with web search or Deep Research for anything time-sensitive (recent funding, launches, leadership changes). A plain chat without browsing will confidently invent dates and numbers.
- As of June 2026, Gemini Deep Research (Google AI Pro, $19.99/mo) is the strongest for multi-source competitive briefs; ChatGPT Deep Research (Plus, $20/mo, 10 runs/mo) is close and reads more conversationally; Claude with web search is great for tight, well-cited summaries.
- Always end a research prompt with a verify-list instruction. Then cross-check dates, headcount, and exec names against the company’s own pages, LinkedIn, and a 10-K if it’s public.
- Budget 60–90 minutes for an important interview. Past that, returns diminish fast.
Who this is for
Candidates prepping for interviews, networkers pitching a target company, and salary negotiators who want leverage before the first call.
Pick the right AI first (June 2026)
Company research is the one career task where the model and mode matter as much as the prompt. The big differentiator is whether the model can actually browse the live web and synthesize sources, or is just recalling stale training data.
| Tool / mode | Best for | Plan needed (June 2026) | Notes |
|---|---|---|---|
| Gemini Deep Research (Gemini 3.1 Pro) | Multi-competitor briefs, deep dives | Free (5 runs/mo) or Google AI Pro $19.99/mo (much higher cap) | Strongest research depth; can pull from Gmail/Drive/Docs if you connect Workspace |
| ChatGPT Deep Research (GPT-5.5) | Structured 90-day news + risk reports | Plus $20/mo (10 runs/mo); Pro $100 (50) / $200 (250) | Not on Free or Go ($8); reads more like a narrative briefing |
| ChatGPT with web search (GPT-5.5) | Quick fact-checks, single-question lookups | Free or any paid tier | Faster than Deep Research; good for the day-of re-scan |
| Claude (Opus 4.7 / Sonnet 4.6) | Tight, well-cited summaries; 10-K parsing | Pro $20/mo | 1M-token context handles a full annual report in one paste |
Practical rule: do the heavy research run in Deep Research mode (Gemini or ChatGPT), then paste the findings into Claude to compress them into a one-page brief and a verify-list. Two-model workflows beat any single chat here.
One hard rule no matter the tool: if a prompt asks for a date, number, or name and the model is not browsing, treat the output as a draft. Models hallucinate exec titles and funding rounds. Every template below ends with a verify instruction for this reason.
Prompt anatomy
Every company research prompt should carry six elements:
- Role — candidate, hiring manager, recruiter; name the persona the AI plays.
- Context — target role, company, level, your background.
- Goal — one deliverable: analysis, script, answer, or plan.
- Constraints — word count, banned phrases, must-include facts, “cite sources.”
- Tone — confident / curious / measured; pick 2–3 anchors.
- Examples — paste 1–2 of your past answers or a sample of your voice.
12 copy-ready prompt templates
In each template, replace the [bracketed] placeholders with your details. Run the research-heavy ones (1, 2, 3, 6, 11) in a browsing or Deep Research mode.
1. 60-minute briefing
Company: [company]. Use web search. Write a 60-minute research briefing:
(1) What they make + for whom, (2) Stage / size / funding, (3) 3 specific
recent moves (product / hire / partnership) with source links and dates,
(4) Strategic risks they face, (5) 3 questions an inside engineer might be
debating right now. Don't invent — write "[verify]" anywhere you're unsure.Swap: company
2. 3 things you can name
Find 3 specific things about [company] I can name in an interview: a shipped
feature, a public talk or blog post, a recent hire or departure. For each:
source link + how I'd connect it to my background as a [your role/background].Swap: company, your role/background
3. Product / market position
Map [company] against 3 named competitors. For each axis (target audience,
pricing, scope, geography), say who wins. Output the position they'd describe
themselves as vs. the position competitors might describe them as.Swap: company
4. Culture signals
From these public sources [paste links] (engineering blog, hiring page,
Glassdoor highlights), extract 5 culture signals (decision-making, hours,
transparency, growth path, conflict). Mark each: well-supported / weak.Swap: source links
5. Team I’m joining
Within [company], the team is [team], led by [leader]. From public signals
(blog, talks, LinkedIn posts), what does this leader value? What's their
recent focus? Anything they've published I should reference?Swap: company, team, leader
6. Recent news brief
Last 90 days, using web search: what newsworthy events at [company]? Funding,
hires, departures, launches, controversy. For each: link + date + relevance
to my role as [your role]. Skip routine PR.Swap: company, your role
7. Strategic risks
What are 3 strategic risks [company] faces in the next 12-24 months?
(Competitor, regulatory, hiring, churn.) For each: how my role as [your role]
could be part of addressing it.Swap: company, your role
8. Questions to ask the interviewer
Based on my research notes below, generate 8 questions to ask interviewers at
[company]. Vary them: strategic, role-specific, team-specific, personal-growth,
culture-test. Skip anything answered on their website. Notes: [paste notes].Swap: company, notes
9. Pre-recruiter brief
Generate a 1-page brief for my recruiter call: (1) Company TL;DR, (2) What I
find compelling, (3) 1 question I want answered, (4) 1 concern to validate,
(5) Salary range from my own research with the sources I used.Swap: none (paste your raw notes first)
10. Talking points for the final round
Final-round panel at [company]. From my research, prep 4 talking points:
(a) something I'd build or fix first, (b) a problem I notice from outside,
(c) a question for the most senior interviewer, (d) a tactful follow-up to
send after.Swap: company
11. 10-K / public filings angle
If [company] is public, summarize their last 10-K / annual report in 6 bullets:
revenue mix, growth rate, biggest risks, segment shifts, R&D priorities, exec
changes. Then name the 1 metric most relevant to my role as [your role].Swap: company, your role
12. Fact-check pass
Here are my current research notes: [paste notes]. Flag every specific claim
(date, number, name, title) I should verify before walking in. Output a
verify-list ordered by interview impact, with where to confirm each one.Swap: notes
Common mistakes
- No specific context (company / role / level) — output stays generic.
- Asking AI to “be honest” without your real record — it confabulates a story.
- Same answer for every company — interviewers on a panel compare notes.
- No tone anchor — answers land flat.
- No browsing for time-sensitive facts — the model invents dates, headcount, and funding.
- Treating the first draft as final — first drafts read AI-flavored; rework them in your voice.
A 60-minute workflow that actually works
- Minutes 0–25 — Run template 1 in Deep Research mode. Skim the briefing; flag anything marked
[verify]. - Minutes 25–40 — Run templates 3 and 6 for competitive position and recent news. Open the source links yourself; don’t trust the summary blind.
- Minutes 40–55 — Run template 12 against your combined notes, then confirm the top 5 verify-items against LinkedIn, the company blog, and a 10-K if public.
- Minutes 55–60 — Run template 8 to turn your verified facts into 8 questions. Pick the 3 sharpest.
For the AI side of this, our AI company + role research guide and organizing market research with AI go deeper on tooling. For the U.S. salary baseline in step 9, levels.fyi and Glassdoor are the standard starting points — validate against 2–3 sources.
FAQ
- Can recruiters tell when an answer is AI-written? Yes, when it’s generic. Specifics that only you could know — a feature you’d ship first, a named competitor risk — are the antidote.
- Which AI is best for company research? As of June 2026, Gemini Deep Research (Google AI Pro, $19.99/mo) leads on multi-source depth; ChatGPT Deep Research (Plus, $20/mo) is close and more narrative; Claude is best for tight, well-cited summaries and parsing a full 10-K.
- Do I need a paid plan? Free tiers help: Gemini gives 5 Deep Research runs/month, and ChatGPT free has web search but no Deep Research. For more than a couple of serious interviews a month, a $20 plan pays for itself.
- How much research is enough? 60–90 minutes for an important interview. Beyond that, returns diminish — spend the time practicing answers instead.
- Should I trust levels.fyi / Glassdoor numbers? As a baseline, yes — with caveats. Validate against 2–3 sources and weight recent, role-and-location-matched data points.
- How often should I refresh before the final round? Do a quick re-scan the day of the interview for news or launches in the past week, ideally in a fast web-search mode rather than a full Deep Research run.
Related
- Behavioral question prompts
- Tell me about yourself prompts
- Networking outreach prompts
- Recruiter reply prompts
- AI company + role research
- Organize market research with AI
- Career & Interview Prompts hub
Tags: #Prompt #Job search #Research