Diligence and deal flow: running a first fund in New Zealand

The volume problem nobody warns you about

Once a first close is announced, inbound arrives faster than a two person team can read it. The failure mode is not missing a good company. It is responding inconsistently, so founders and co-investors cannot predict how you behave, and your own partners cannot reconstruct why something was passed on.

The fix is boring and it works. Write the thesis down before the fund is live. Publish it. Then treat every inbound as a test against that written thesis rather than a fresh judgement call.

• .One page written thesis: stage, sector, cheque size, geography, ownership target.
• .A fixed response window for every inbound, published on the website.
• .A single intake path so nothing arrives only in one partner's inbox.
• .A standard pass note that gives the founder one concrete reason.

A triage system that scales past your headcount

Triage should split inbound into three buckets in minutes, not days: outside thesis, inside thesis but not now, and take a meeting. The first two need a written reason attached to the record, because the "not now" bucket is where a first fund's best follow-on opportunities usually sit.

Capture the reason at the moment of the decision. Reconstructing it a year later, when the company has raised a strong round and an LP asks why you passed, is not possible from memory.

• .Structured intake capturing stage, sector, round size, geography and referral source.
• .Thesis fit scored against the published criteria, not against vibes.
• .Every pass carries a written reason and a revisit trigger.
• .Referral source tracked so you can see which parts of your network actually produce.

Sourcing without a research team

Proprietary deal flow in a smaller market is a function of being genuinely useful before there is a round. Operator networks, sector communities, university spinout offices and angel groups produce more first looks for an emerging manager than any outbound programme.

Track sourcing the way you track a portfolio. If you cannot say which relationships produced first looks last quarter, you cannot invest in the right ones next quarter.

• .Named relationship map: who introduces, how often, and what happened.
• .A cadence for staying in touch with companies that are twelve months early.
• .Sector notes published openly, which convert into inbound over time.
• .Co-investor mapping so you know who leads and who follows in your market.

Founder diligence at seed, without a data room

At the earliest stage there is rarely a data room worth the name. Diligence becomes reference quality, evidence of customer pull, and clarity about what the founders believe that the market does not.

Write the diligence memo as a set of falsifiable claims. Each claim gets evidence and a confidence level. The point is not certainty. The point is that in three years anyone can see exactly what you believed and what you believed it on.

• .Reference calls logged with who was spoken to and what was said.
• .Customer evidence captured directly, not summarised second hand.
• .Cap table and prior instruments reviewed before terms are discussed.
• .Founder background verified, including any prior regulated roles.
• .Key person and single point of failure risks stated plainly.

Investment committee discipline from fund one

A two person fund still needs an investment committee, and it still needs minutes. LP due diligence questionnaires ask how decisions are made. Auditors ask how valuations were arrived at. Regulators, including the FMA, expect a manager to be able to evidence its own process.

The standard to hold yourself to is simple. For any investment in the fund, you should be able to produce the memo, the dissent, the decision, the date and the people present, without searching anyone's email.

• .A written IC memo template used for every decision, including declines.
• .Recorded dissent, because an unrecorded dissent is a governance gap.
• .Conflicts declared and minuted before the discussion, not after.
• .Decision, date, attendees and rationale stored against the company record.
• .A single immutable trail so the reasoning survives staff turnover.

Where an AI investment platform actually helps

Reuben AI is built for this part of the job: intake and triage against a written thesis, diligence evidence captured against the company record, and decision provenance that holds up years later when an LP or auditor asks. The point is not speed for its own sake. It is that a small team can run an institutional process without adding headcount.