Clay
The spreadsheet where every cell can make an AI call and every row can enrich itself.
You have a list of 200 prospects. You need to find each one's LinkedIn, their current role, their company's latest funding round, and write a personalized outreach email. This is six hours of browser-tab drudgery — unless every cell in your spreadsheet can think.
Why this tool matters
Clay is a spreadsheet where every row is a subject of research and every cell can make an AI call, a web scrape, or an API request. It is the tool that has quietly become the backbone of modern sales and business-development operations — but its capabilities extend well beyond sales into research, journalism, recruiting, competitive intelligence, and any workflow that involves enriching a list of entities.
The experience is unusual. You start with a list of names, URLs, or companies. You add columns like: Find LinkedIn URL, Current job title, Company size, Recent news, Personalized opener based on their last post, Classify into these 4 segments. Each column is powered by a combination of: dozens of integrated data providers (ZoomInfo, Clearbit, LinkedIn, Crunchbase, etc.); web scrapers for any site with a URL pattern; and AI (GPT, Claude, Gemini) calls that can read enriched data and produce new insights per row.
For a researcher: imagine a spreadsheet of 300 academic papers that auto-populates columns for methodology, sample size, and a plain-English summary. For a journalist: a list of 50 public companies with auto-enriched quarterly earnings, recent litigation, and a short risk profile. For a non-profit: 200 potential donors with their giving history and personalized outreach drafts. Clay is the tool that lets one person do what used to require a research team.
Setup
Account: clay.com free tier includes 100 credits/month and 14 days of Pro access — enough to complete a real workflow end-to-end. Pricing starts at $149/mo for Starter (2,000 credits), scaling to $349/mo and custom enterprise tiers. This is the most expensive tool in this course; use the trial period deliberately.
Prereq: comfort with spreadsheets and the idea that “each cell is a formula.” If you've used Google Sheets formulas, you'll be fluent in Clay in about 30 minutes. If spreadsheets intimidate you, do Day 12 (Otter) and Day 21 (Gamma) first; Clay rewards spreadsheet thinking.
Walkthrough
Step 1: Start with a seed list
In Clay, create a new table. Paste or import a list of 20–50 entities you want to research — names + companies, or URLs, or LinkedIn profiles, or research-paper DOIs. Clay needs at least one identifying column to anchor the enrichment.
Step 2: Add one enrichment column
Click Add Column → Enrich. Pick a simple enrichment: Find LinkedIn URL or Find company website. Map the input column. Run it. Watch cells populate one by one. This is your “aha” moment.
Step 3: Chain enrichments
Add a second column that depends on the first. Given the LinkedIn URL, find their current company. Then a third: Given the company, find recent news. Each cell computes from the previous; errors propagate, so check your first column's accuracy before layering on more.
Step 4: Use AI to reason over enriched data
Add an AI column. Prompt it to read values from other columns and produce output: “Given their recent job title and company news, write a 2-sentence personalized note that references something specific.” Pick Claude or GPT. Run on the first 3 rows only to calibrate; then run on the full set.
Step 5: Filter and segment with conditions
Once enriched, use filters to segment: company size > 50, recent news contains “funding”, job title contains “research”. Export segments to CSV, or push directly to HubSpot / Salesforce / Airtable via Clay's native integrations.
Step 6: Build a reusable recipe
When a workflow is proven, save it as a Template. The next time you have a list of the same entity type, point the template at the new list and it runs end-to-end. This is where Clay pays back its price: the second use of any workflow is essentially free.
Your turn
Basic: Enrich a list of 20 entities
Build a seed list of 20 things you'd genuinely want to know more about — 20 companies you're tracking, 20 researchers in your field, 20 local non-profits, 20 podcasts you might pitch, 20 competitors. Enrich each row with 3–5 columns (including one AI-generated field).
Review the output. Spot-check 3 rows against the source — is the data right? Is the AI reasoning reasonable? Export to CSV and use the data for something real this week.
Advanced: Build a reusable research pipeline
Identify a recurring research task in your work: monthly competitor tracking, weekly investor outreach, quarterly partner scouting, term-by-term student outreach, annual donor research. Design a Clay workflow that automates it:
- Input column(s) that represent the seed list.
- 5–10 enrichment columns that cover everything you care about.
- 2–3 AI columns that produce reasoning or copy.
- A filter or segment that surfaces the high-priority rows.
- A push destination (CSV export, HubSpot, Airtable, email).
Save it as a template. Run it on a real batch this week. Compare the result to what you would have produced by hand in the same time. Run the same template again next month on fresh data.
Write a 200-word memo: what just became possible that wasn't before? What is the annual time savings? Is Clay's subscription priced correctly for your usage?
Pitfalls and pro tips
Credit costs add up fast. Every enrichment cell costs credits, and AI cells cost more per call than data-provider cells. A naive workflow on 500 rows with 10 columns can burn through 5,000 credits. Always test on 5–10 rows before running on the full list — this is the single habit that saves Clay subscribers the most money.
Enrichment accuracy varies by data source. Clay aggregates dozens of providers; the first one tried is not always the best. For mission-critical fields, use the waterfall feature to try multiple providers in sequence and use the first successful result.
AI columns hallucinate on thin context. If you ask for “a personalized opener” and the only data Clay has about the person is their job title, it will invent details. Always include enough enrichment columns to ground the AI, and review AI output before sending anywhere it represents you.
How it compares
Clay's direct competitors are Apollo.io (cheaper, narrower, sales- focused), Airtable + custom automations (more flexible, much more work to set up), ZoomInfo (data-provider-first, less workflow flex), and Instantly (outbound-specific, lighter research). For general- purpose “research every row in a list with multiple data sources + AI,” Clay is currently unmatched. For sales-only workflows at smaller companies, Apollo is the budget choice. Zapier (Course 1) can orchestrate pieces of what Clay does but cannot replicate the spreadsheet-as-canvas metaphor.
When to use — and when not to
Use Clay when you have a recurring need to enrich and reason over lists of entities: sales prospects, research subjects, journalism targets, donor lists, competitive scans, recruiting pipelines. The ROI is proportional to how often you'd otherwise be opening browser tabs.
Do not use Clay when your list is one-off and small (<20 entities, a manual afternoon beats the Clay learning curve), when your data is highly proprietary and can't go through third-party enrichment APIs (compliance teams say no), or when you just need a better CRM (use HubSpot; Clay feeds into CRMs, it doesn't replace them).