How to Find Decision Makers in Clay Without People Finder

Clay's people search is the Find People source. You add it to a workbook, filter by company attributes, job title, experience, location and a handful of profile fields, and Clay returns matching people from its database, up to 50,000 records a search and up to 100 per company. When those filters can express the title, that is the right first move. When your niche is thin in that database, or the title you need takes judgment no filter can express, a Claygent finds the people instead, and the job becomes getting its list of names out of one cell and into rows.

How does Clay people search work?

Clay people search runs through the Find People source. In a workbook you click Add, search for Find People, and set filters: company size, industry and description keywords, job title, experience, location, profile, certifications, languages, education and specific companies. Each match becomes a row. Clay's docs cap a search at 50,000 records and let you limit results to 100 people per company.

The job title filter has three modes, and picking the wrong one is the usual reason a search looks thin. The default includes synonyms, similar titles and translated titles, so "Software Developer" also returns a "Frontend Engineer". "Job title must contain exact" requires one of your terms in the title, ignoring capitalisation, with no synonyms. "Job title must match exactly" requires an exact match, capitalisation included. You can also exclude up to 300,000 people, 100,000 per source, from earlier tables or lists, as long as each has a LinkedIn URL, which keeps a new search from re-surfacing people you already have.

Teams that still wonder whether this enrichment tool overlaps with their outreach platform will find a cross category platform comparison that clarifies how the two tools actually complement each other.

Sometimes the built-in database just doesn't cover your niche, or you need highly specific titles that take an AI agent's nuance. For those tricky cases there is an alternative method that works just as well, if not better.

The Challenge: Locked Data

Let's say you're targeting a specific niche (like schools, malls, or specialized firms) and need to find the "Head of School" or a specific Director. You import your list of companies into Clay and run a Claygent, Clay's AI research agent, to find these people.

The agent does a great job. It returns a list of decision-makers for each company.

But here's the problem: That data is often returned as a list (array) inside a single cell. You have one row for the company, and a clump of people stuck in one column. You can't enroll them in a campaign like that.

You need to "explode" that list so each decision-maker gets their own row, while keeping the company's information (Website, Name, etc.) attached to them.

Here are two ways to solve this.

Method 1: The "Clay Native" Way (Recommended)

This is the fastest method and keeps everything inside Clay. Clay's own docs call it flattening a list: turning a list of five people into five separate rows.

1. Open Send table data: In the table that holds the Claygent output, click Exports, then Send table data.
2. Pick the destination: Select the table the people should land in, or create one first (e.g., "Decision Makers").
3. Choose the method: Select Send row for each item in a list, then the column where the Claygent wrote its list of decision-makers.
4. Select the data to send:
   • Map the decision-maker's name, title, and email from the list.
   • Crucial Step: Also select the "Parent" data from your original table, like the Company Name, Website, and Phone Number. Clay sends it along with every flattened item, so each new row has the company context.
5. Save, then extract: In the destination table the data arrives in the leftmost column, labelled "rows from:" plus the source table's name. Turn on Auto-extract new columns when you configure the send, or hover over each field in that cell and click Add to column.

Result: You now have a clean table where every row is a unique person, complete with their company data. You can now enrich them with emails or push them straight to your CRM.

Once those contacts land in your CRM, you may wonder whether that platform pairs well with Clay, a question explored in this comparison of Clay and Copper.

Earlier versions of this guide pointed at a column action that wrote each item to a new row in another table. That was Clay's Write to Other Table action, and Clay's docs now mark it deprecated: existing actions keep working, but new ones are not supported. Send table data replaces it, and its limits are worth knowing before you chain tables together. At most 20 tables can be connected, and data only moves in one direction, A to B to C, never back to A. To pull data the other way, Clay points to its Lookup Single Row and Lookup Multiple Rows actions.

Before: one row, a list in one cell

Company	Decision makers
School A	[Head of School, Deputy Head, Director of Admissions]

Send row for each item in a list, with the parent columns selected

After: one row per person, company attached

Name	Title	Company
Person 1	Head of School	School A
Person 2	Deputy Head	School A
Person 3	Director of Admissions	School A

Method 2: The "AI Assistant" Way (For Complex Data)

If the data is messy or truncating, or if you want to do complex cleaning before creating new rows, you can use an AI code editor like Cursor.

1. Export to CSV: Export your Clay table (the one with the list in a cell) as a CSV.
2. Upload to AI: Drag that CSV into Cursor.
3. The Prompt: Ask the AI:

   "I have a column here called 'Decision Makers' which contains a list of objects. Please create a new CSV where each item in that list gets its own row. Keep the other columns (Company Name, Website) for each new row."

4. Review & Re-Import: The AI will write a script to process the file and give you a clean CSV.
5. Import to Clay: Upload this new CSV back into Clay.

Note: This method adds a few steps but is a lifesaver if you need to apply complex logic, like "Only keep decision makers if the confidence score is > 80%" or "Reformat the titles before splitting."

Pro Tip: Add Logic Before You Split

Whether you use Method 1 or 2, consider adding a filter before you create the new rows.

For example, if your AI Agent returns a "Confidence Score" for each person, you can tell Clay (or your AI script) to only keep people with a score of High or Medium. This saves you credits on enrichment later by filtering out low-quality matches early.

1

Claygent returns people

With a confidence score for each person

2

Keep High and Medium

Low-quality matches dropped in the source table

3

Send row for each item

Only the kept people become rows

4

Enrich the kept rows

Emails and phone numbers, credits spent on people you will contact

Does Clay have a phone number finder?

Yes. Clay's Sculptor pages describe a work phone finder: paste a list of target accounts, and Sculptor searches Clay's people database for contacts at each account, then runs them through a waterfall of data providers in sequence until one returns a verified work phone. The page says you only pay credits when a match is found.

The mechanics matter for this workflow. Clay's page says the waterfall reaches through 150+ providers and stops the moment one returns a valid work phone, so an empty lookup costs nothing. Each matched number then runs through Trestle verification, which returns a line type, an activity score and a name match grade, and the finished rows can be pushed to Outreach, Salesloft or HubSpot. The same logic of trying providers in order until one answers is what this site's guide to waterfall enrichment walks through outside Clay.

Treat the coverage figures on that page as the vendor's own claims, and note that they differ from each other. One section says a single provider typically covers 30-50% of a list and the waterfall routinely triples that, while the FAQ says a single provider covers around 30% of contacts and the waterfall pushes that to roughly 80% or higher. The FAQ also reports a Clay test of 9,806 numbers across 10 B2B phone data providers, in which 23% qualified as verified, high-propensity numbers.

For the decision-maker workflow above, the phone step belongs at the end. Split the list, filter it, and run the phone waterfall only on the rows you kept, so the credits land on people you actually plan to call.

Summary

Don't let a "list in a cell" stop your workflow. Whether you use Clay's native Send table data feature or wrangle the data with an external AI tool, unlocking this data gives you access to decision-makers that standard databases often miss.

Related Reading

• How to Extract Data from Any API in 2 Minutes Without Writing Code
• A Clay Lead Generation Workflow, End to End, With Real Costs

If you would rather have this run for you, RevenueFlow books qualified meetings on a pay-per-meeting basis and publishes client results.