Efficiency onlyBuild toolEstablished, AI retrofitted

Clay

Clay Labs Inc ·www.clay.com· As of: 2026-07-30

A list of companies or people rarely carries the attributes you actually decide on. Clay is the workbench where you assemble those attributes, query several data sources in sequence and put your own scoring logic on top.

Our verdict

For a sales team that wants to target rather than blanket, enrichment and signal processing are no longer optional. Without them you work through lists whose selection criteria nobody can explain any more. It still does not amount to an edge: Clay sells the same workbench and the same data marketplace to everyone, and any competitor can query the same providers in the same order. What sets you apart is the judgement about which attributes matter at all, and that judgement lives in the operator's head and in your own customer data, not in the tool. Calling it a case-by-case efficiency gain would understate it, because what hangs on it is not one task but the entire stage that precedes outreach.

Who it fits

Solo

suitable with caveats

The free tier with 500 actions and 100 data credits a month is enough to try it out, not to work with. The first workable plan costs 167 USD a month, plus the time it takes to build the logic. A solo seller mostly buys themselves work here.

Mid-market

suitable

This is the group the tool is built for. Seats and tables are unlimited, one person can maintain the logic and the whole team reads from it. The precondition is that somebody keeps an eye on consumption.

Enterprise

suitable with caveats

Single sign-on is only an add-on on the Growth plan, the storage location of the data is not publicly documented, and there is no public commitment on access logging. For a corporate review that means an enterprise contract or nothing.

Buy, switch or build

Off-the-shelf

Buy a data vendor with ready-made filters

You subscribe to a database and filter it by industry, size and role. The start is faster and the bill is a fixed annual amount with no consumption surprises. You pay for that in selection: you can only sort by what the vendor carries, and your competitors buy the same database and filter it the same way. Coverage gaps only surface when reply rates drop.

AI-native

An agent researches and qualifies on its own

A newer class of tools takes the target definition and returns researched, scored contacts without anyone building columns. That saves exactly the work Clay demands. The price is that the criteria disappear inside the tool: once nobody can say why a company is on the list, there is no way to correct course when the list stops converting.

Build it yourself

Your own register with direct provider APIs

A script calls the data providers' APIs directly and writes into your own database, with the scoring living in code and under version control. This pays off when the criteria are stable, the volume is large and the data is needed elsewhere in the company anyway. It breaks where the criteria change weekly, because then sales waits on engineering.

Our advice

  • Solo: Try the free tier, but count the setup time honestly.
  • Mid-market: Buy it, put one person in charge and review consumption monthly.
  • Enterprise: Enterprise contract only, with storage location and audit logging settled in writing first.

Who owns what gets built

bound to the platform · portable with limits

Under its terms of service the vendor claims no ownership of the data you upload, while retaining all intellectual property rights in the platform and its content. What the workbench produces is an enriched and scored list. It is not an end product but moves on into the CRM, into an outreach or messaging tool, into a data warehouse or as a CSV file into the hands of the sales team. The work itself, meaning the table structure, the order of data providers and the wording of the AI steps, stays tied to the account and has no interchange format. On top of that, exporting counts as an action and consumes quota.

Who is behind it

Staying power: Funded

The vendor is well funded and growing fast, but the valuation sits far above revenue. This is not a forecast about the company's survival, it is a note that today's prices and terms rest on a growth expectation.

  • Legal entity and seat: Clay Labs Inc, 119 N 11th Street, 3C, Brooklyn, NY 11249, USA · Source · As of 2026-07-30
  • Funding: USD 277m across seven rounds, most recently a USD 100m Series C in August 2025 led by CapitalG · Source · As of 2026-07-30
  • Valuation: USD 5bn at an employee share sale in January 2026, previously USD 3.1bn after the Series C · Source · As of 2026-07-30
  • Revenue: estimated USD 150m annual recurring revenue in May 2026, up from USD 108m at the end of 2025, a third-party estimate not confirmed by the vendor · Source · As of 2026-07-30
  • Years in market: first funding round in 2017 per the vendor, founding year 2019 per Sacra, the sources contradict each other · Source · As of 2026-07-30
  • Security audit evidence: SOC 2 Type 2 per the vendor's own announcement of 9 September 2024, report available only through the trust center on request · Source · As of 2026-07-30

Cost of leaving: Moderate

The data comes out: every table can be downloaded as CSV, and there are live integrations into CRM and outreach tools. What does not come out is the work itself, meaning the table structures, the order of data providers and the wording of the AI steps. There is no interchange format for those, so they get rebuilt on a switch. On top of that, the vendor counts exporting as an action, so it consumes quota.

Regulation and data

Data processing agreementon request, including standard contractual clausesThe privacy policy states that Clay offers data processing agreements on request and that these include standard contractual clauses. A signed copy has to be requested via security@clay.com and is not publicly posted.Source · As of 2026-07-30
Storage locationunclearNeither the privacy policy nor the terms name a country or region for data storage. The trust center only releases its contents on request and could not be read for this review. Third parties claim a US region, which is not solid evidence.no source given · As of 2026-07-30
Subprocessorsa list exists, but it is not publicly retrievableThe vendor maintains a subprocessor list in its trust center. The page only loads its contents after a request, so it could not be read openly. The only subprocessor named in the privacy policy is Stripe for payment processing.Source · As of 2026-07-30
Third-country transferprocessing in the US, covered by standard contractual clauses in the data processing agreementThe vendor is based in New York and New York law applies. For European customers, the privacy policy points to the data processing agreement with standard contractual clauses as the basis.Source · As of 2026-07-30
Training on customer dataaccording to the vendor, customer data is not used for model trainingThe product documentation states explicitly that contracts with all AI providers prohibit training on customer data. How long those model providers retain the content passed to them is not stated.Source · As of 2026-07-30
Retention and deletiondata persists as long as the account does and is deleted when the account is deletedThe terms put this briefly and name neither retention periods for backups nor a grace period after the contract ends.Source · As of 2026-07-30
CertificationsSOC 2 Type 2 evidenced, ISO 27001 not evidenced first-handThe completion of the SOC 2 Type 2 audit is announced by the vendor itself. An ISO 27001 certificate is mentioned in third-party market overviews but could not be confirmed on any reachable vendor page. Ask for the evidence before procuring.Source · As of 2026-07-30
EU AI Act, Article 50unclearOn its publicly reachable pages the vendor says nothing about the transparency duties under Article 50. In practice your own duty matters more anyway: whoever sends machine-generated outreach built with this tool carries the disclosure obligation.no source given · As of 2026-07-30
Audit loggingunclearThere is no public commitment on auditable access and change logs. The only documented point is that single sign-on is an add-on on the Growth plan and included on Enterprise.Source · As of 2026-07-30

What it really costs

Entry price

Free with 500 actions and 100 data credits a month. Launch costs 167 USD a month with 15,000 actions and 3,000 data credits, Growth 446 USD a month with 40,000 actions and 6,000 data credits. Seats and tables are unlimited on every plan, annual billing saves 10 percent.

Source · As of2026-07-30

What gets expensive

The monthly price is not the price. Two currencies run side by side. Actions cost fractions of a cent and are charged for every step, including exporting data out of the system. Data credits cost several cents each and are charged as soon as third-party data is bought. The vendor puts a typical enrichment at 6 to 20 data credits. That makes the 6,000 credits on the Growth plan enough for roughly 300 to 1,000 enriched rows a month, not 6,000. Topping up carries a 30 percent premium, actions expire monthly, and credits only roll over up to twice the monthly allowance. Comparing plans by the base price means comparing the wrong number.

What it displaces

  • Manual list research done person by person
  • Several parallel subscriptions with individual data vendors
  • Scoring logic that lives only in a spreadsheet on one machine

Interfaces

  • CSV import and CSV export
  • HTTP calls to any API straight from the table
  • Integrations with CRM, outreach and messaging tools
  • Your own provider API keys instead of data credits
  • MCP connection

At Convios: Not used by us

Not used in our own operation. The statements in this profile come from the vendor's pages and product documentation, not from first-hand use.

Matching methods

Evidence

  1. 1.Plans, included actions and data credits, top-up premium, single sign-on as an add-on · As of 2026-07-30
  2. 2.Definition of actions and data credits, typical enrichment at 6 to 20 credits, expiry and rollover · As of 2026-07-30
  3. 3.Customer data is not used for model training, contracts with the AI providers prohibit it · As of 2026-07-30
  4. 4.Legal entity, address, data processing agreement on request including standard contractual clauses · As of 2026-07-30
  5. 5.Retention only while the account exists, automatic renewal, fees non-refundable · As of 2026-07-30
  6. 6.SOC 2 Type 2 completed, report available through the trust center on request · As of 2026-07-30
  7. 7.Valuation, funding rounds and estimated annual recurring revenue · As of 2026-07-30
  8. 8.CSV export of any table and live integrations as exit routes · As of 2026-07-30

Last reviewed: 2026-07-30byDr. Oliver Gausmann, Convios GmbH

Details out of date? Let us know.

Work through this method with an AI