Build a company list where every row has a source

The request comes in on a Monday: "Can you pull together every US drone maker we should be talking to?" It might be for next quarter's territory plan, an investment committee memo or a market sizing slide. A database export gives you hundreds of rows, some stale and some in the wrong category. A chatbot gives you a tidy list that falls apart the first time someone asks where a name came from. This page shows how to build a market map or prospect list where every row carries its source, how list size works by tier, and how to get the result into Excel or your CRM.

Why last year's list is probably wrong

Markets that policy touches move fastest. In June 2025, an executive order on drones told federal agencies to "prioritize the integration of UAS manufactured in the United States over those made abroad to the maximum extent permitted by law." Six months later, the FCC added drones and drone components made outside the US to its Covered List. A list of US drone makers built before those changes is missing the context that now decides which companies matter to a buyer.

The people side of prospect data is getting harder to keep, too. Since January 2026, Californians can send one deletion request to every registered data broker through the state's DROP platform, and more than 475,000 had done so by August 2026. Since August 1, 2026, brokers have had to process those requests, checking the platform at least once every 45 days. Contact records churn. Company facts with a public source behind them are the part of a list that lasts.

How CiteJury builds the list

Choose Find companies and describe the market in a sentence: what the companies do and where they are. Claude, ChatGPT and Grok each search the web for candidates. Every candidate gets its own search, and its own website is read. Each AI proposes companies, then every AI checks every proposed company against the same sources. A company is confirmed only when most of the AIs find a source that names it and shows it fits.

Each confirmed row gives you what the company does, where it is, why it fits and which AIs confirmed it, with links to the passages behind it. Below that sit a "likely" list and a "ruled out" list with reasons. In the app you can filter by name, product or place, and sort by best match, name, location or how many AIs confirmed each row.

The download is built for a CRM import. The Excel and CSV files have columns for Company, Website, Location, What they do, Why they fit, Founded, Size, Notable, AIs confirming, Confirmed and Sources, where Sources holds the URLs. Likely companies are included and marked as not confirmed, so you can import them to a separate stage or leave them out. Founded, Size and Notable are filled in only when a source states them.

These rows come from a real market map, numbered by their place in the list.

A real list: drone makers in the USA

We asked for "Drone manufacturing companies in the USA" on Standard. It took 12 minutes, drew on 53 sources and confirmed 12 companies, with 13 more on the likely list and 17 ruled out. Claude, ChatGPT and Grok each checked every company that any of them put forward.

Standard aims for up to 25 companies. This search confirmed 12 and didn't pad the list to reach the target, which is what you want when the list goes to a partner or a sales team. Several well-known names, among them BRINC, Inspired Flight and Shield AI, landed on the likely list, because no source in the pack showed both that they build drones and that they build them in the US. That's the rule working: a famous name isn't evidence. The 13 likely companies are the natural next round of research.

The rows show why a market map needs segments. Skydio makes autonomous drones and docks. Easy Aerial in Brooklyn designs and makes drone-in-a-box monitoring systems. Lucid Bots in Charlotte makes drones that clean the outside of buildings. Equinox Innovative Systems in Columbia, Maryland, builds tethered and heavy-lift multi-rotor systems, and Teal Drones in Salt Lake City makes the Golden Eagle. A seller of cleaning equipment and a seller of defense components would keep very different parts of this list.

Two details are worth copying into your own review. Skydio's location reads only "USA", because that's all the source stated. And read the ruled-out list: ModalAI was ruled out because it makes autopilots and flight controllers, not drones, which is exactly the line a buyer would draw. Claims such as "built with domestic engineering", as the Aero Systems West row puts it, usually come from a company's own site, so open the source before you lean on it. For a sales list, a homepage is often enough. For a list that feeds a procurement or compliance decision, "made in the USA" needs more.

How to build yours

1. Write the market in one sentence with a place. "Drone manufacturing companies in the USA" is a good start. Add a segment if you know it, such as "tethered drones for public safety".
2. Say what doesn't count. Resellers, service providers, companies that only design. The ruled-out list shows you how that was applied.
3. Pick the list size. Quick returns up to 10 companies, Standard up to 25 and Deep up to 50. A Standard search takes about 10 minutes and costs about 48 credits.
4. Review the confirmed rows first. Open "Why it fits" for anything you'll act on.
5. Decide what to do with the likely list. Research it, import it as a lower stage, or drop it.
6. Download the CSV and map the columns. Company and Website map to your account fields. Put Sources in a notes or URL field so the reason travels with the record.
7. Keep each market in a project. Group the searches, rename them, and run them again when the market moves.

Tips for sales, investors and researchers

• Dedupe on the website domain. Company names vary, as with "Teal Drones" and "Teal Drones (Red Cat Holdings)" in this very search. Domains don't. Match on the domain when you import to avoid duplicate accounts.
• Segment before outreach. One search per segment gives cleaner lists than one broad search you split by hand.
• For a market map, run the adjacent searches. Makers, component suppliers and software providers are separate searches, and the gaps between them are often the finding.
• Treat blanks as honest. An empty Founded or Size field means no source stated it. Fill it from a filing or the company, not a guess.
• Add people through your own process. CiteJury gives you companies. Contacts come from your own compliant sources.
• Date the list. A market map is a snapshot. Put the run date in the file name.

What it won't do

A list reflects what the search found. It isn't a complete census, so 12 confirmed companies doesn't mean the US has 12 drone makers. Many rows are confirmed only by the company's own website, which shows what a company says it does and little else.

It gives company facts only: no names, emails or phone numbers for people. It doesn't estimate revenue, headcount or funding beyond what a source states. It can't read paid databases or pages behind a login, some sites block automated access, and social media pages are skipped. A company that launched last month may not have enough coverage to be confirmed yet. Results can be wrong or incomplete, which is why every row links to its source.

The point of a sourced list isn't length. It's that every row survives the question "why is this company here?" For suppliers rather than prospects, see finding suppliers and manufacturers you can verify. If the list feeds an investment, pair it with checking a pitch deck before the partner meeting, and if you're sizing a market for your own product, start by testing the idea against real evidence.

Questions people ask