# What goes on a data readiness checklist?

> A data readiness checklist for licensing company data to AI: eight lines, what micro1 and Handshake AI publish, and a fictional filled-in sample.

By Matt Burns, Vaultfish. Canonical: https://vaultfish.com/answers/data-readiness-checklist  
Published: 2026-08-05  
Updated: 2026-10-14

## The short answer

Eight lines: headcount, usable history, systems and owners, work linked to results, rough volume, language, rights notes and personal data. Rights come first. micro1's pages ask for 30+ employees. Handshake AI's referral page asks for 20+ full-time staff. Neither floor sets a price.

Run this list before you talk to any buyer. It takes an afternoon and nothing leaves the building. It can also end in a no, and a no on paper costs less than a no after three calls.

Spoken remarks are marked undated and count as what was said then, not current written criteria.

Figure: steps Confirm rights, list systems -> Size records, note language -> Plan personal data removal. The readiness checklist runs from rights to a plan for personal data.

## What are the eight lines of a data readiness checklist?

Eight lines: headcount, usable history, systems and owners, work linked to its result, rough volume per tool, language, rights notes and personal data. micro1 and Handshake AI publish criteria for some of them.

1. **Headcount.** Full-time staff, with contractors counted apart. [micro1](/buyers/micro1)'s [data partnerships page](https://www.micro1.ai/data-partnerships) asks for 30 or more employees. [Handshake AI](/buyers/handshake-ai)'s [referral page](https://joinhandshake.com/ai/data-partnerships/referral) asks for 20 or more full-time employees.
2. **Usable history.** The oldest record in each system that still links to its outcome. Handshake AI's referral page asks for three or more years in operation, which is company age. Usable history is a different number, covered below.
3. **Systems and owners.** Every tool where work gets recorded, and who can export from it. In an undated recorded forum talk, David Revland of micro1's human data team named Google Workspace, Slack, project management tools and GitHub.
4. **Work linked to its result.** Whether a request ties to how it ended, like a ticket to its fix or a quote to a won deal. micro1's page lists project histories and QA processes among the data it wants.
5. **Rough volume per tool.** A size or a count per system, with the date you checked it. micro1's pages state no data-size threshold. In the same talk, Revland described a data manifest form that asks which tools you use and a rough size for each.
6. **Language.** The main language in each system. micro1's pages ask for companies that operate in English and give no percentage.
7. **Rights notes.** Contracts, NDAs and vendor terms that touch the data, and any policy on who owns staff messages. Mark what stays out, such as privileged material, HR files, payroll, health data and union communications ([Frankfurt Kurnit](https://technologylaw.fkks.com/post/102o455/selling-company-data-for-ai-training-what-to-know-before-you-say-yes)).
8. **Personal data.** What personal data the records hold, and who will remove it. Health, financial and education records carry their own rules and need specialist handling.

Figure: scale, Published staff floors at two companies: Handshake AI referral page 20 or more, micro1 pages 30 or more (emphasized). micro1's published floor is 30 employees, and Handshake AI's referral page starts at 20. Source: joinhandshake.com/ai/data-partnerships/referral and micro1.ai/data-partnerships, self-reported.

## Which line on a readiness checklist matters most?

Rights. The thresholds move from company to company and nobody publishes a pass mark. The rights work goes in order: bring in legal, privacy and HR, review your agreements, then set what stays out.

Daniel Goldberg, a partner at [Frankfurt Kurnit Klein & Selz](https://technologylaw.fkks.com/post/102o455/selling-company-data-for-ai-training-what-to-know-before-you-say-yes), writing in October 2026, advises bringing in legal, privacy, security, HR and communications people first, so business development doesn't run the deal alone.

He then puts a review of customer, partner, vendor and data license agreements before any deal.

_Rights come first, so do not start with the spreadsheet._

The [rights check](/rights-check) asks seven questions, including whose records these are and who can sign a license. It lists what to settle with counsel.

## Does company age count as history on a readiness checklist?

No. Usable history counts records whose links still work, and a founding date says nothing about that. System migrations break the context.

A CRM exported to flat files during a move can lose which email belonged to which deal, and a chat tool you left behind may have taken its threads with it.

Count usable history per system, from the oldest record that still ties to its outcome. A company founded 18 years ago that moved its CRM in 2022 has about four years of linked CRM history.

## Does a readiness checklist value your data?

No. The headcount and storage on a readiness checklist are screening signals. AI data companies use them to decide whether to talk, and on their own they don't set what a deal would pay.

micro1's page lists deal tiers from $100K+ and ties the higher tiers to how large and how unique the data is. Nobody publishes a price per employee or per gigabyte.

Older figures still circulate, and they are not micro1's current written criteria. In an undated recorded forum talk, micro1 staff named about 1 TB of data and 20 full-time employees, and David Revland put the English share at 80% or more. That talk shows what was said then.

[Troveo's business-data page](https://www.troveo.ai/business-data) states no minimum headcount. It says value scales with your systems, team size and years of history.

## What does micro1 publish as its criteria?

micro1's current public criteria list 30+ employees, mature operations and primarily English documentation, with US companies prioritized. Acceptance depends on its review and current demand. Confirm exceptions directly with micro1.

An undated micro1 [Enterprise Data Partnership post](https://req.micro1.ai/post/74520e8b-4673-4da5-b855-53609aaf6710) lists 30 to 200 employees. micro1's [data partnerships page](https://www.micro1.ai/data-partnerships) adds that demand is strongest in the US, UK and Canada.

Handshake AI's [data partnership estimator](https://joinhandshake.com/ai/data-partnerships/) asks for headcount, years of operation and country.

In an undated recorded forum talk, micro1's David Revland said micro1 wants the chain from a request to a finished deliverable across tools. Records tied to an outcome, such as a won deal or a resolved ticket, show that chain.

## What does a filled-in data readiness checklist look like?

A filled-in checklist gives one row per line: your answer, what it tells a buyer and the next step. Otterbrook Supply, a made-up 42-person plumbing wholesaler founded in 2008, shows how one reads, with descriptions and counts only.

**Fictional readiness checklist for Otterbrook Supply, a made-up 42-person plumbing wholesaler. Descriptions and counts only, no records.**

| Line | Otterbrook's answer | What it tells a buyer | Next step |
|---|---|---|---|
| Headcount | 42 full-time, 6 seasonal contractors | Above both published staff floors. Sets no price | List contractors apart |
| Usable history | Email since 2013, chat since 2020, CRM since a 2022 move | Founded 18 years ago, about 4 years of linked CRM history | Ask whether the old CRM export kept deal links |
| Systems and owners | Google Workspace and Slack (office manager), HubSpot (sales lead), Zendesk (support lead) | Four systems, three people who can export | Confirm admin access on each |
| Work linked to its result | Tickets carry order numbers. Deals link to quotes since 2022 | A request-to-result chain from 2022 on | Note the gap before 2022 |
| Rough volume | Drive about 600 GB, Slack about 40 GB, 1.4 million tickets | Describes the records. Sets no price | Screenshot each total with its date |
| Language | English, with Spanish in about 1 in 10 support tickets | Primarily English | Note the share per tool |
| Rights notes | Three distributor contracts with confidentiality clauses. Handbook silent on AI use | Open questions, so no buyer talk yet | Send both to counsel |
| Personal data | Customer names and addresses in Zendesk. Payroll and HR files in one Drive folder | HR and payroll are on Frankfurt Kurnit's exclusion list | Exclude the folder, ask each company who de-identifies |

Otterbrook ends the afternoon with two questions for counsel, one exclusion and one gap to explain. It gets no price and no promise of a deal. Being imaginary, it also found every admin login on the first try.

## Is a data readiness checklist the same as AI data readiness?

Partly. A readiness checklist for licensing asks whether an AI data company can use your records and whether you can show they are yours to license. Most guides use "AI data readiness" for getting your own data fit to run your own AI.

[Hiniduma, Byna and Bez's survey](https://arxiv.org/abs/2404.05779) (revised November 2024) treats it as whether a dataset is fit for AI training, judged by metrics still being standardized.

> **Readiness means a company can use your records and you can show they are yours to license.** It's like renting out a spare room. First you check the lease lets you, then you clear out what isn't yours.

## Who removes the personal data a readiness checklist flags?

Usually the company you license to, so ask who removes it, which method it uses and what it certifies. Handshake AI's [data partnerships page](https://joinhandshake.com/ai/data-partnerships/) says the de-identification work happens on its side, and that no process removes the risk entirely.

In an undated recorded forum talk, micro1 staff said personal data has to be transformed on micro1's side, because it swaps each name for the same stand-in in every tool. Data scrubbed beforehand would break that trail.

Goldberg adds that de-identification addresses whether a record links to a person. A pay complaint stays sensitive without a name on it. The preparation guide covers [whether to scrub first](/guides/how-to-prepare-company-data-for-ai-licensing#should-you-scrub-the-data-first).

## How does the Vaultfish fit check use the checklist?

The free Vaultfish [fit check](/business) asks about full-time employees, years of records, where the work happens, whether work ties to its outcome, how easily you can export, whether contracts allow this use and what personal data the records hold.

That makes seven questions, plus an optional industry question it never scores. No upload, no login.

You get a score, a band and a short list of what would raise it. Public reference points come with their source, and none of them estimates your payout.

The [methodology page](/methodology) shows how the points are built, and the weights are our own design choices, with no market data behind them.

Some answers set the band aside. If a contract bars this use, or the records are health, financial or education data, the result shows no band or range and sends you to counsel first.

If you haven't checked your contracts yet, it shows no band and sends you to the [rights check](/rights-check). micro1's published range stays hidden until that check flags no blocker.

The fit check doesn't review a contract, give legal advice or value your data. Our guide on [what your data is worth](/guides/how-much-is-your-business-data-worth-to-ai) shows why nobody publishes a price list.

## What comes after a data readiness checklist?

You choose. Take your rights notes to counsel and work through the gaps yourself, ask us for an introduction, or hand the job over. The [data licensing audit](/answers/data-licensing-audit) page sets out a deeper review in five parts, and the [preparation guide](/guides/how-to-prepare-company-data-for-ai-licensing) turns this list into a one-page packet.

Optional [managed readiness](/managed-readiness) builds a one-pager, a systems inventory, questions for your counsel, a shortlist of companies with current criteria and a term-question checklist. It works from descriptions and counts, never records, and it can end in a no.

Clearing a threshold doesn't promise a deal or a price.

[Apply to micro1's data partnership (paid link)](https://req.micro1.ai/post/74520e8b-4673-4da5-b855-53609aaf6710?referralCode=f82e519e-5d64-4e0e-9574-760c84371cf2&utm_source=referral&utm_medium=share&utm_campaign=data_referral) · [micro1's own data-partnerships page](https://www.micro1.ai/data-partnerships)

_Paid link: micro1 may pay us a referral fee if a company we introduce goes ahead with it. You pay nothing extra._

_Vaultfish helps you run the process. We are not lawyers and this is not legal advice. If you need counsel who knows this area, we can point you to one._

## Sources

- [data partnerships page](https://www.micro1.ai/data-partnerships)
- [referral page](https://joinhandshake.com/ai/data-partnerships/referral)
- [Frankfurt Kurnit](https://technologylaw.fkks.com/post/102o455/selling-company-data-for-ai-training-what-to-know-before-you-say-yes)
- [Troveo's business-data page](https://www.troveo.ai/business-data)
- [Enterprise Data Partnership post](https://req.micro1.ai/post/74520e8b-4673-4da5-b855-53609aaf6710)
- [data partnership estimator](https://joinhandshake.com/ai/data-partnerships/)
- [Hiniduma, Byna and Bez's survey](https://arxiv.org/abs/2404.05779)