# Enterprise software fund data referrals: a brief | SourceX

Canonical: https://refer.sourcex.si/guides/the-data-supply-conversation-for-funds-focused-on-enterprise-software
Last updated: 2026-11-04

An enterprise software portfolio holds two kinds of records, and usually only one belongs to the portfolio. The data each product hosts is the customers'; a vendor processing it on their behalf cannot offer it. What a software company may own is the record of running itself: support case handling, implementation projects, incident reviews, QA and release histories. Start the data supply conversation there, with companies that reached 50 or more full-time employees at peak, then add two lanes: customers whose CEOs opt in through the vendor, and the enterprise operators the fund already convenes.

## Turn the enterprise software thesis into a question about records

Most enterprise software funds already hold a view on AI. The data supply version is narrower: teams training and testing agents on office work need realistic records of how a company actually operates, and researchers go to some length to get them. The authors of TheAgentCompany benchmark built a self-hosted, simulated software company, with internal tools such as GitLab, Plane, RocketChat and ownCloud, simulated colleagues and 175 professional tasks; the best baseline agent in their study completed about 30% of those tasks autonomously.

If researchers simulate a software company to get workplace context, the internal operating records of real software companies are worth a question. The question for a portfolio is not "which company has the most customer data", which points at the wrong owner. It is this: which of our software companies has years of internal operating records created by its own staff, reached 50 or more full-time employees at peak, and has a CEO who would want to discuss licensing them?

| Record inside a software company | Example | Usually belongs to | Can the portfolio company discuss it? |
|---|---|---|---|
| Hosted product data | Tickets, documents and transactions customers create in the product | The customer | No; the customer's CEO decides |
| Support operations | How the company's own team triaged, escalated and resolved cases | The company, though case text can carry customer content | Possibly, after rights review |
| Implementation history | Project plans, change requests and go-live checklists for each rollout | The company, subject to customer contracts | Possibly |
| Engineering operations | Issue tracker history, incident reviews, release and QA records | The company | Usually the clearest case |
| Sales engineering | Discovery notes, solution designs, questionnaire answers | The company, with customer confidential material inside | Case by case |

The legal line on hosted data is firm where EU personal data is involved. Article 28 of the GDPR requires a processor to process personal data only on the controller's documented instructions, and a processor that decides the purposes and means itself is treated as a controller for that processing. SourceX's screening guidance reaches the same practical answer from another direction: material that belongs to a company's clients is usually not a fit.

## The enterprise-network activation brief

Write the brief once, before anyone contacts a company. It prevents three failures: approaching the wrong owner, approaching through the board seat, and continuing after the evidence has turned.

| Field | What to write | Illustrative entry |
|---|---|---|
| 1. Thesis | One sentence about records, not markets | Growth-stage portfolio companies hold years of support and engineering records their own staff created |
| 2. Disproof | The result that ends the brief | Four of the first six companies say their support records are mostly customer content |
| 3. Lanes | Which of the three lanes below are in scope | Lanes 1 and 3; lane 2 only if a portfolio CEO proposes it |
| 4. Fit facts | Peak full-time headcount, years of records, rights, CEO | Six portfolio companies above 50 full-time employees, six to twelve years old |
| 5. Record families | Named at metadata level only | Support cases, incident reviews, implementation projects |
| 6. Relationship owner | Who asks each CEO | Board partner for four companies, platform lead for two |
| 7. Limits | What the fund will not do | No customer data, no files or logins, no request made in a board meeting |
| 8. Pilot | Size and end date | Ten companies, ending 2026-12-18 |
| 9. Review | Who reads the result and decides | Platform lead and two deal partners |

**Method note.** The fields come from SourceX's published screening profile and program terms and from Article 28 of the GDPR, checked on 2026-10-26. The brief records ownership defaults, fit facts and relationship paths. It does not estimate demand for any record family, record quality or payout; those are assessed only in SourceX's review. Counts, dates and descriptions in the right-hand column are invented for illustration.

## Three lanes, three different decision-makers

| Lane | Whose records | Who decides | First move | Main risk |
|---|---|---|---|---|
| 1. The portfolio company's own operations | The portfolio company's | Its CEO | The board partner raises it one to one | Customer content mixed into support and implementation records |
| 2. The portfolio company's customers | Each customer's | Each customer's CEO | The vendor tells its customers; interested CEOs come forward | Customers reading it as the vendor trading on their relationship |
| 3. Enterprise operators the fund convenes | Each operator's | Each operator's CEO, not the executive who attends | Ask the attendee to sponsor an introduction to their CEO | Treating a functional executive as the authorizer |

**Lane 1** is easy to overlook, because a fund tends to think of its portfolio as sellers of software rather than holders of records. A company with ten years of [issue tracker histories](/datasets/software-engineering/issue-tracker-histories), incident reviews and support escalations has a decade of its own operating history. Early-stage portfolios are likely to fail the size test here; growth-stage portfolios may not.

**Lane 2** belongs to the portfolio company, not the fund. The vendor can tell customers about the program and introduce those whose CEOs ask, while never touching the data it hosts; the [customer opt-in flow for software companies](/guides/how-a-software-company-can-introduce-customers-to-sourcex) sets out the steps, and the guide to [aggregated-data clauses in SaaS agreements](/guides/aggregated-data-clause-saas-agreement) covers what a vendor's own contract may say about customer data. Whether to make this a fund-wide program at all is weighed in [opt-in versus mandatory portfolio programs](/compare/opt-in-vs-mandatory-portfolio-company-programs).

**Lane 3** is the fund's convening power: the CIO councils, customer advisory boards and LP operating companies an enterprise platform team already brings together. The people in those rooms run large operations, but a CIO or VP of operations is rarely the person who authorizes licensing. Ask them to sponsor a conversation with their CEO, and only if they think the CEO would want it.

Credit follows the program rules in every lane. One partner per company gets credit, and only for the earliest valid introduction made while the company was still new to SourceX and inside the 90-day attribution window, settled once the company submits a verified application; if SourceX already had a lead, an account or email correspondence with that company from before, it is an existing relationship. Agree in writing, before outreach, whether the fund or the portfolio company submits a lane 2 introduction. Software roll-ups need one more note: a parent and its subsidiaries count as one company, so the USD 50,000 floor applies once, to the group's first qualifying transaction, and does not repeat for each subsidiary. What happens to credit when the parent or a sister company has already been in touch with SourceX, or another partner introduced a sibling, has no published rule; SourceX settles it in duplicate review, so promise nothing in advance.

## Ownership and relationship evidence, lane by lane

| Lane | Ownership evidence to gather (metadata only) | Relationship evidence | Unknown until SourceX reviews |
|---|---|---|---|
| 1 | Which internal systems, since which year, rough counts, how much customer text the records carry | The CEO has heard the idea from someone they trust and wants a conversation | Whether customer contracts or privacy terms limit reuse; demand |
| 2 | The customer's own description of its records, given by its CEO | The customer's CEO contacted the vendor, not the reverse | The same, plus whether the customer passes the size test |
| 3 | The operator's records as its CEO describes them | A named sponsor and a CEO who agreed to talk | The same |

Every request in every lane stays descriptive: what the records are, roughly how many, over how many years, in which system and administered by whom. Nobody at the fund asks for exports, samples or credentials, and nobody uses a board seat to obtain a portfolio company's records for this purpose. The [venture platform overview](/for/venture-platform-teams) covers the program from the platform side, and the [software partner canvas](/guides/the-partner-opportunity-beyond-selling-another-software-seat) is the vendor's own view of lane 2.

## A two-lane pilot, and the result that would retire the thesis

Run lanes 1 and 3 for eight weeks with ten companies: six portfolio companies above the size threshold and four operators from the fund's councils. Write the thresholds before the first conversation.

| Lane | Count to watch | Threshold written in advance | If it misses |
|---|---|---|---|
| 1 | Companies whose CEO can describe support or engineering records created by their own staff, with limited customer content | At least 3 of 6 | The thesis is wrong for this portfolio; stop lane 1 |
| 1 | CEOs who agree to a first conversation with SourceX | At least 2 of 6 | Ask whether the idea or the messenger failed before stopping |
| 3 | Attendees willing to sponsor an introduction to their CEO | At least 2 of 4 | Lane 3 is a relationship problem, not a data one; pause it |

Report counts with their denominators, including companies that declined or were never asked. One more result would retire it outright: CEOs of companies with strong internal records deciding, case after case, that the conversation is a distraction from product. Funding status plays no part in any of this; a well-funded company with 40 full-time employees fails the size test, and an older, quieter one may pass every gate.

## Where the enterprise software version of the thesis breaks

- **Simulation instead of records.** TheAgentCompany is itself a simulated company. If simulated environments keep improving, demand for real internal records from software companies could narrow.
- **Customer content everywhere.** Support cases, implementation notes and sales engineering files often quote customers. If most of a company's internal records cannot be separated from customer material, rights review may exclude them.
- **Size.** Many venture-backed software companies stay below 50 full-time employees for years, and contractors do not count.
- **Board-seat pressure.** A request that reaches the CEO as a board expectation rather than an option damages the relationship and fails the program's own expectation that the decision maker wants the conversation.
- **Changing programs.** SourceX applies the criteria of a specific program and checks relevance in review, so a company declined today may fit another program later, and the reverse.

## Next step

> **Refer a company. No cap on rewards.** Default reward on a company's first qualifying transaction: the greater of $50,000 plus 5% and 15% of Actual Supplier Payout, with no cap; further qualifying transactions first agreed within 12 months of the first Closing pay 15%. Eligibility and payment conditions apply, including the USD 100,000 Minimum Transaction, and a signed Schedule A or addendum can differ. A lead, meeting or signature alone does not trigger payment.

When the brief names a portfolio company with years of its own operating records and a CEO who wants to talk, [register for a partner link](/register), test it against [who qualifies](/who-qualifies) and read the [program terms](/terms) with whoever will submit. Related methods are collected in the [guides hub](/guides).

Activating a portfolio for data introductions is a referral decision; assessing SourceX as a possible investment is separate, and investor materials can be requested through the [SourceX investor overview](https://sourcex.si/investors).

This article is for information only and is not an offer to sell or a solicitation of an offer to buy securities. Any offering by SourceX is made only to verified accredited investors through its offering documents.

## Sources

- https://refer.sourcex.si/terms
- https://refer.sourcex.si/who-qualifies
- https://arxiv.org/abs/2412.14161
- https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex%3A32016R0679
- https://sourcex.si/investors
