---
title: "Company research with web data"
description: "Company research from the public web: read a company site, filings, news, job posts and reviews into one brief for due diligence or sales prep."
canonical: https://www.zenrows.com/use-cases/company-research
updated: 2026-11-02
---

# Company research with web data

Company research builds a picture of one company from its public web footprint: its site, filings, news, hiring and customer reviews. Zenrows crawls the company site, fetches filings, news and review pages as clean Markdown, and gives AI agents the same access through the MCP Server, so a brief cites the pages behind it.

## Sample output: a sourced company brief

```json
{
  "company": "Example Robotics Inc.",
  "domain": "example-robotics.com",
  "researched_at": "2026-10-17",
  "what_they_do": "Autonomous forklifts for mid-size warehouses, sold as a monthly subscription",
  "filings": [
    { "type": "Form D", "filed": "2026-05-02", "note": "equity offering", "url": "https://www.sec.gov/..." }
  ],
  "recent_news": [
    { "date": "2026-09-18", "headline": "Example Robotics opens Dallas service hub", "url": "https://news.example.com/..." }
  ],
  "hiring": { "open_roles": 22, "functions": ["field service", "sales", "robotics software"], "new_locations": ["Dallas, TX"] },
  "customer_sentiment": { "source": "g2.com", "rating": 4.3, "common_complaint": "slow parts delivery" },
  "litigation_found": false,
  "open_questions": ["Revenue not disclosed", "Two leadership pages disagree on founding year"],
  "pages_read": 37
}
```

Illustrative agent brief from pages fetched via Zenrows; each claim links to its source.

## What company research covers

Investors, compliance teams and sales reps research the same companies: shared sources, different questions.

| Job | Data you collect | Typical sources |
|---|---|---|
| Sales account research | What they sell, launches, priorities, hiring | Company site, blog and changelog, careers page, press releases |
| Investment and M&A screening | Business model, growth, funding, ownership, risks | SEC EDGAR, Companies House filings, Crunchbase, investor relations pages |
| Supplier and partner due diligence | Legal status, addresses, litigation, sanctions, adverse news | Business registries, court record portals, news search |
| Reputation check | Customer and employee sentiment, recurring complaints | G2, Trustpilot, BBB, app store reviews |
| Competitive briefing | Pricing, positioning, product changes, new markets | Pricing and product pages, Google Search, news |
| Agent-run research | Any of the above, gathered and summarized on request | Whatever the agent reads, via MCP |

## Why researching a company by hand stalls

A thorough brief takes 30 to 50 pages from a dozen sites: hours by hand. An agent without reliable access stops at the first protected page, or invents.

| What breaks | What it looks like | How Zenrows handles it |
|---|---|---|
| Too many pages per company | 40 tabs, careers page still missed | Crawl collects the site in one job as Markdown |
| Agents blocked on protected sources | "Unable to access" on review or news sites | The MCP Server scrapes with Adaptive Stealth Mode, escalating only when needed |
| Noisy HTML in the model context | Navigation, cookie banners and scripts eat tokens | Markdown output: clean text an LLM reads directly |
| Filings and registries in odd formats | Data in tables, forms or multi-step searches | Output filters return tables as JSON; [Browser](/products/browser) handles search forms |
| Unsourced claims | "Raised $20M" with nothing to check | Every page keeps its URL, so each claim carries its source |

## How to research a company with Zenrows

1. **Collect the company's own site**: one capped crawl to Markdown brings about, product, pricing, careers and press.
2. **Add outside sources**: filings, registry records, news and reviews as Markdown.
3. **Let the model read, with sources**, from your pipeline or an agent on the MCP Server; require a URL per claim.
4. **Save the brief and re-run it** before the next meeting.

```text Claude Code (MCP)
# Connect once
claude mcp add zenrows -e ZENROWS_API_KEY=YOUR_ZENROWS_API_KEY -- npx -y @zenrows/mcp

# Then ask:
Research example-robotics.com for a first sales call. Use Zenrows to read their homepage, pricing, careers and press pages, plus the first page of news results for the company name. Summarize what they sell, what changed in the last six months and who they are hiring, and cite the URL for every point.
```

```bash cURL
curl -X POST "https://api.zenrows.com/v1/crawls" \
  -H "X-API-Key: YOUR_ZENROWS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example-robotics.com/",
    "max_items": 60,
    "exclude_patterns": ["/blog/page/", "/tag/"],
    "output_format": "markdown"
  }'
# Then: GET https://api.zenrows.com/v1/crawls/{crawl_id}
```

```python Python
# pip install requests
import requests

SOURCES = {
    "filing": "FILING_PAGE_URL",       # e.g. an EDGAR or Companies House filing page
    "reviews": "REVIEW_PROFILE_URL",   # e.g. the company's G2 or Trustpilot page
}

pages = {}
for name, url in SOURCES.items():
    response = requests.get("https://api.zenrows.com/v1/", params={
        "apikey": "YOUR_ZENROWS_API_KEY",
        "url": url,
        "mode": "auto",  # Adaptive Stealth Mode
        "response_type": "markdown",
    })
    pages[name] = {"url": url, "markdown": response.text}
# Pass `pages` to your LLM and ask for a brief that cites each URL.
```

```bash CLI
npx -y @zenrows/cli init
zenrows crawl create https://example-robotics.com/ --max-pages 60 --exclude-patterns /tag/ --output-format markdown --follow
zenrows crawl download CRAWL_ID --out example-robotics.jsonl
zenrows fetch REVIEW_PROFILE_URL --output markdown --out reviews.md
zenrows extract FILING_PAGE_URL --outputs tables --out filing-tables.json
```

## Which Zenrows product fits

| If you need to | Use | Why |
|---|---|---|
| Read a company's whole site | [Crawl](/products/crawl) | Follows internal links on the domain, Markdown per page |
| Fetch a filing, news or review page | [Fetch](/products/fetch) | Clean Markdown from protected and JavaScript-heavy pages |
| Let analysts research in Claude or Cursor | [MCP Server](/mcp) | The assistant calls Zenrows to read live pages |
| Build your own research agent | [Agent Toolkit](/agents) | MCP setup, skills and templates that teach a coding agent which Zenrows tool fits each step |

## What success looks like

| Measure | What good looks like | How to check it |
|---|---|---|
| Time per brief | Minutes, not an afternoon | Request to finished brief |
| Citation coverage | Every factual claim links to a page | Claims without a URL in sampled briefs |
| Source breadth | Site plus filings, news and reviews | Distinct source domains per brief |
| Accuracy on review | No claim contradicted by its source | Five claims per brief against the cited page |
| Fetch failure rate | Agents rarely miss a page | Failed fetches per run |

## Company research FAQ

### How do I research a company online?

The company's site first: about, product, pricing, careers, press. Then filings (SEC EDGAR in the US, Companies House in the UK, or the local registry), news, G2 or Trustpilot reviews, and court or sanctions records for due diligence. Zenrows fetches all of it as clean text for you or a model.

### Can an AI agent do company research for me?

Yes. Connect Claude, Cursor or another MCP client to the Zenrows MCP Server and ask: "research this company before my call and cite every source". The agent picks pages and fetches them through Zenrows. For a custom pipeline, the Agent Toolkit sets up MCP, skills and templates so a coding agent starts on the right Zenrows primitive. See deep research agents .

### What sources are used for company due diligence?

Registries (legal status, officers, filings), court portals (litigation), news (adverse media), sanctions lists, annual reports and investor relations pages (financials), and the company's own claims. At volume: adverse media screening .

### How do I research a private company?

Use other sources. Registries like Companies House in the UK publish officers, ownership and accounts; Crunchbase, rounds and investors; careers pages and job boards, growing teams; reviews and news, customers and problems. The full site says what it sells, to whom.

### How is company research different from lead enrichment?

Lead enrichment fills fixed fields for thousands of accounts; company research writes a deep, judged brief on one. Enrich all accounts, research those that reach a meeting or investment committee.

### How do I keep a company brief up to date?

Keep each brief's URLs, not just the summary. Before the next review, re-fetch them with fresh news and ask the model what changed. For accounts you watch continuously, news monitoring catches coverage in between.

## Related use cases

- [Deep research agent](https://www.zenrows.com/use-cases/deep-research-agent.md): An agent that reads dozens of live sources as Markdown, including protected pages, and cites every claim.
- [Firmographic data](https://www.zenrows.com/use-cases/firmographic-data.md): Industry, headcount, HQ, locations and funding per company, built from registries, company sites and directories.
- [Adverse media screening](https://www.zenrows.com/use-cases/adverse-media-screening.md): Negative news on customers, directors and suppliers from global and local sources, with text and screenshots as evidence.
- [Lead enrichment](https://www.zenrows.com/use-cases/lead-enrichment.md): A domain in, an account record out: pricing, hiring, product and review signals from the company's own web.
- [News monitoring](https://www.zenrows.com/use-cases/news-monitoring.md): Alerts on new stories about your companies, people and keywords, polled from feeds, wires and publisher pages.
- [Court records API](https://www.zenrows.com/use-cases/court-records-api.md): Dockets, case status, hearing dates and attorney records from state court and bar portals, kept current.

Start free: https://app.zenrows.com/register. Docs: https://docs.zenrows.com

## Structured data



<script type="application/ld+json">
{"@context":"https://schema.org","@graph":[{"@type":"WebPage","@id":"https://www.zenrows.com/use-cases/company-research#webpage","url":"https://www.zenrows.com/use-cases/company-research","name":"Company research with web data","description":"Company research from the public web: read a company site, filings, news, job posts and reviews into one brief for due diligence or sales prep.","isPartOf":{"@id":"https://www.zenrows.com/#website"},"inLanguage":"en","datePublished":"2026-11-02","dateModified":"2026-11-02","about":{"@type":"Thing","name":"company research"},"keywords":"company research, company research tool, how to research a company, company due diligence research, account research, company background research, research a company online, ai company research agent","publisher":{"@type":"Organization","name":"Zenrows","url":"https://www.zenrows.com/"}},{"@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://www.zenrows.com/"},{"@type":"ListItem","position":2,"name":"Use cases","item":"https://www.zenrows.com/use-cases"},{"@type":"ListItem","position":3,"name":"Company research with web data","item":"https://www.zenrows.com/use-cases/company-research"}]},{"@type":"FAQPage","@id":"https://www.zenrows.com/use-cases/company-research#faq","mainEntity":[{"@type":"Question","name":"How do I research a company online?","acceptedAnswer":{"@type":"Answer","text":"The company's site first: about, product, pricing, careers, press. Then filings (SEC EDGAR in the US, Companies House in the UK, or the local registry), news, G2 or Trustpilot reviews, and court or sanctions records for due diligence. Zenrows fetches all of it as clean text for you or a model."}},{"@type":"Question","name":"Can an AI agent do company research for me?","acceptedAnswer":{"@type":"Answer","text":"Yes. Connect Claude, Cursor or another MCP client to the Zenrows MCP Server and ask: \"research this company before my call and cite every source\". The agent picks pages and fetches them through Zenrows. For a custom pipeline, the Agent Toolkit sets up MCP, skills and templates so a coding agent starts on the right Zenrows primitive. See deep research agents ."}},{"@type":"Question","name":"What sources are used for company due diligence?","acceptedAnswer":{"@type":"Answer","text":"Registries (legal status, officers, filings), court portals (litigation), news (adverse media), sanctions lists, annual reports and investor relations pages (financials), and the company's own claims. At volume: adverse media screening ."}},{"@type":"Question","name":"How do I research a private company?","acceptedAnswer":{"@type":"Answer","text":"Use other sources. Registries like Companies House in the UK publish officers, ownership and accounts; Crunchbase, rounds and investors; careers pages and job boards, growing teams; reviews and news, customers and problems. The full site says what it sells, to whom."}},{"@type":"Question","name":"How is company research different from lead enrichment?","acceptedAnswer":{"@type":"Answer","text":"Lead enrichment fills fixed fields for thousands of accounts; company research writes a deep, judged brief on one. Enrich all accounts, research those that reach a meeting or investment committee."}},{"@type":"Question","name":"How do I keep a company brief up to date?","acceptedAnswer":{"@type":"Answer","text":"Keep each brief's URLs, not just the summary. Before the next review, re-fetch them with fresh news and ask the model what changed. For accounts you watch continuously, news monitoring catches coverage in between."}}]}]}
</script>
