Sell CRM data and business data to AI

Why CRM data and business data are in demand

AI labs are increasingly buying workplace data to train AI agents – systems meant to carry out real work tasks. Sales, account management, purchasing and finance are typical areas where such agents are expected to work. To learn, they need to see what the work actually looks like: how a deal moves through a sales process, how quotes are followed up and how transactions are booked.

CRM history and exports from business systems contain exactly that structure. Read more in the guide on why AI agents are trained on company data.

Selling CRM data – what can be included?

CRM history

From systems such as HubSpot, Salesforce and similar:

• opportunities and how they developed step by step
• activities, notes and follow-ups
• quotes and reasons for won and lost deals
• communication history linked to deals

ERP and accounting data

From ERP and accounting systems:

• journal entries and account coding
• order, purchasing and inventory flows
• invoices and payment patterns
• project and time reporting

Structured data with clear fields and a long history is often more useful than large volumes of unsorted exports. Notes and free text where salespeople or finance staff explain why something was done can carry particular weight.

What makes business data valuable?

Value is determined by the same factors as for other operational data, with a few particular features:

• Complete sequences. A deal you can follow from first contact to invoice says more than isolated entries.
• Long history. Several years give seasonal patterns, economic cycles and more unusual cases.
• Links between systems. If CRM data and accounting data can be linked, the picture of how deals are actually done becomes more complete.
• Explanatory text. Notes, comments and justifications are often what separates useful training data from plain statistics.
• Local markets and languages. Business communication and accounting flows in Nordic and other European languages and jurisdictions are rare in global training data.

We describe what different types of data can be worth in more detail on what is your data worth.

Personal data and customer details

CRM data is fundamentally information about customers and contacts. That means you are not "selling your customers" – the goal is to license the patterns in the business, not information about individuals. That requires careful groundwork:

• Contacts. Names, emails and phone numbers of people at client companies are personal data, even in B2B.
• Sole traders. Information about sole traders can be personal data.
• Salaries and employees. Payroll details and staff costs in the accounts need to be removed or de-identified.
• Free text. Notes often contain more than you think.

European data protection authorities stress that reusing data for AI training must be compatible with the original purpose, and the European Data Protection Board requires a documented balancing test if legitimate interest is relied on. That is why we assess rights and personal data first and prioritize data without personal data. Read more about GDPR and anonymization and in the guide Can you sell customer data?. What we write is general information, not legal advice.

Trade secrets and contracts

Business data can reveal prices, margins and customer relationships. It is up to you to decide what may be included. A few things to review:

• Agreements with customers and suppliers may contain confidentiality clauses that affect what can be licensed.
• System terms. Check what your agreements with system vendors say about exporting and using the data.
• Limiting by time. Older data is often less sensitive than current data, and can be a natural first selection. You can also start with a limited part, such as one business area or time period.

By default the license is non-exclusive and limited in scope, time and use. Nothing is shared until a non-disclosure agreement and an accepted offer are in place.

How it works

1. You describe the data in our form – which systems, how many years, approximate volume and whether there is personal data. No files needed.
2. We assess rights, personal data and demand, and get back to you within 48 hours on business days.
3. You approve a concrete offer – or say no.
4. Delivery and payment take place after de-identification.

It costs nothing to have the data assessed. We get paid through a share of the revenue, and only when a deal goes through.