Capital allocation inside an AI venture studio.

We have written about why running a small number of brands deeply is the right shape for an AI venture studio. This essay is what happens underneath that decision every quarter: how build capital and senior engineering time actually get allocated across Lucidly, AdvisorKit, FlashGTM, RenewalPad, TempoDocs, Attribufi, and RecruitFractional, and across whatever the next brand turns out to be.

The two scarcest resources

Unlike a fund that allocates dollars into other people's companies, we allocate two things that matter more than cash: build capital, the actual money spent on infrastructure, tooling, and go-to-market for a given brand, and senior engineering time, the studio's most experienced builders, who can only be inside one brand's codebase at a time. Cash is rarely the binding constraint. People and attention are.

1. Fund the brand with the clearest near-term compounding opportunity

Our first priority in any quarter is the existing brand where an incremental dollar or an incremental senior engineer produces the clearest compounding return, an AI feature that closes a real gap against the workflow, a go-to-market motion that is already working and just needs fuel, an infrastructure fix that unlocks the next cohort of enterprise customers. This bar is high, because each brand already runs lean; the ask has to be for something the brand's own cash flow cannot yet fund on its own.

2. Fund a new brand's zero-to-one build

The second priority is originating and building a new brand from scratch, following the process we describe in the first 100 days of a brand you just started. New-brand builds compete directly with existing-brand reinvestment for the same senior engineers, so we only greenlight a new build when it has cleared the pre-build validation bar and has a specific operator ready to run it.

3. Invest in the shared studio platform

The third priority is the shared platform itself, described in our essay on the studio platform, the design system, engineering scaffolding, infra, and distribution layer every brand launches on. A dollar spent here pays back across every current and future brand, which usually makes it a better long-run bet than a single-brand ask, but it is slower to show up in any one brand's numbers, so it is easy to under-fund if we are not deliberate about it.

4. Let it sit rather than force it out

The fourth and final option is to simply hold build capital rather than deploy it. Some quarters, the right answer is that no existing brand has cleared the reinvestment bar and no new idea has cleared the validation bar. Forcing capital into a mediocre bet because it is sitting unused is how studios end up with brands nobody really wanted to build.

The scarce resource is never the money. It is the handful of senior builders who can actually make a brand's roadmap move.

The discipline of writing the argument down

None of the above is novel on its own. What actually keeps allocation honest is writing the argument down every quarter: which brand gets the next engineer and why, which new build gets greenlit and why, what we chose not to fund. That memo gets reviewed openly with the operators running each brand, so nobody is surprised by where the studio's attention is going next.

The annual review

We grade the prior year's calls honestly: did the reinvestment in a given brand produce the expected lift, did a new brand's build actually reach the milestones that justified starting it, did platform investment measurably reduce the time it took to launch the next brand. The grading is uncomfortable in the quarters where a call did not pan out, and those are exactly the quarters that teach us the most about our own judgment.

Common mistakes

• Funding the loudest brand instead of the highest-leverage one. The operator who asks most persistently is not automatically the one with the best use of the next engineer.
• Starting a new brand because capacity is technically available. Available engineering time is not the same as a validated idea.
• Under-investing in the shared platform. Platform work rarely wins a head-to-head comparison against a single brand's urgent request, which is exactly why it needs a protected allocation.
• Treating "hold the capital" as a failure to decide. Sometimes the most disciplined allocation decision in a given quarter is not making one.

The bottom line

Allocating build capital and engineering time across seven brands is the single most consequential decision the studio makes every quarter, more than any individual brand's roadmap call. The framework, fund the clearest existing opportunity, then a validated new build, then the shared platform, then hold if nothing clears the bar, is not exotic. The discipline of writing it down and reviewing it honestly is what keeps it from drifting.

If you're an investor curious how this discipline shows up in practice, start a conversation with our team.