Two seed-stage AI companies raise similar rounds in the same quarter. Both plan to hire ten people in the following twelve months.

Eighteen months later, one has a small team that ships and sells. The other has a bigger team that argues about priorities, rebuilds the same service twice and cannot say which customer segment is working.

The difference was rarely talent. It was sequence.

Hiring Is a Dependency Graph

Every early hire depends on decisions the previous hire made. The first engineer defines the architecture the second one inherits. The first seller defines what a qualified opportunity looks like for the second. Hire in the wrong order and the later people inherit decisions nobody made deliberately.

That is why a seed-stage hiring plan should read less like an org chart and more like a list of bottlenecks. What is currently stopping the company from moving faster? The next hire should remove exactly that, and nothing else.

A workable sequence follows four stages. The specifics change with the product, but the logic holds across most AI and infrastructure companies.

Hires One to Three: Make the Product Real

The first hires are almost always engineers, and the mistake is treating them as interchangeable. The first two or three engineers define what the product is technically capable of, what it costs to run and how it fails.

For AI companies this usually means one person with strong systems or infrastructure instincts and one with depth in the modelling, retrieval or agent layer the product depends on. A third is typically a generalist who can move across both.

• Bias towards ownership. At this stage the best engineers want a whole problem, not a ticket queue. Test for it by asking what they have built end to end.
• Hire for the architecture you will need, not the one you have. A wrong early technical choice is expensive to unwind after design partners depend on it.
• Keep the founders in the interviews. Nobody else can sell the scope of the role credibly.

Hires Four to Six: Make the Product Sellable

The founders have been selling. It has worked well enough to raise, and now the number of conversations is outgrowing their calendar. This is where the commercial hires begin, and it is where most sequencing errors happen.

The instinct is to hire a head of sales. Resist it. A leader who has scaled a motion is trained to install a system, and at this stage there is no system to install. The first commercial hire should be someone who can build a repeatable motion alongside the founders, not someone who expects one to exist.

Pair the first seller with a solutions engineer. In AI and infrastructure the buyer is usually technical, and the evaluation is a proof of concept rather than a pitch. A seller without technical support stalls at the second meeting. A solutions engineer without a seller has nobody to validate for.

We set out the commercial sequencing in more detail in Sequencing GTM Hires at Seed and Series A.

Hires Seven to Eight: Remove the Founders From the Loop

By this point there are usually one or two bottlenecks that still route through a founder. Deals that only close when the founder joins. Technical decisions that only settle when the other founder weighs in. Customer questions that only get answered by someone who wrote the original code.

These two hires exist to break those dependencies. They are rarely glamorous. Often they are a technical customer success or support engineer who takes customer escalations away from the engineering team, and a second engineer who owns a domain the founders have been holding together by hand.

The test for each is simple: after ninety days, which decisions no longer need a founder in the room?

Hires Nine and Ten: Add Leverage, Not Just Capacity

The last two of the first ten should multiply what the team already does. That might be a developer relations hire if the product is sold through engineers, a revenue operations lead once there is enough pipeline to measure, or the first product marketing hire when the message finally needs to work without the founders delivering it.

Which one depends on what the company has learned by now. The point is that this is the first stage where you have enough evidence to choose, and the earlier hires should have produced it.

Where the Sequence Usually Breaks

• The premature leader. A head of sales or head of engineering arrives before there is a team or a motion to lead. They spend the first quarter installing process and the second wondering why nobody is using it.
• The narrow specialist too early. A specialist hired before there is enough work in their area to keep them learning tends to leave within a year, and takes the context with them.
• Volume as a strategy. Four engineers hired in one quarter to move faster often produce less than two hired well, because the founders spend the quarter on onboarding instead of decisions.
• Skipping the brief. A role that has not been defined in terms of the first ninety days cannot be hired for properly. The searches that run slowly are usually the ones where this was never written down.

Selling the Role Matters as Much as Screening for It

The people most worth hiring at seed stage are typically employed and not looking. They will trade certainty for scope and equity, but only if the company can explain, with some precision, what they would own, why the timing is right and how the business intends to win.

A generic job description does none of that. A specific brief and a credible founder conversation do all of it.

This is the part of early-stage hiring that founders most often underestimate. The process runs in both directions. The candidate is assessing the company just as carefully as the company is assessing them, and at this stage the company has less to point at than it will ever have again.

A Simple Test Before Every Hire

Before opening any role, write down three things. The bottleneck this person removes. The decision they will own within ninety days. The hire that comes after them, and what it depends on.

If any of the three is unclear, the role is not ready to go to market. Fixing that takes an afternoon. Fixing the wrong hire takes a quarter.

If you are planning your first ten hires and want a second view on the sequence before anything opens, I would be happy to talk it through. You can see how we work with founders at this stage on our page for the recruitment agency for early-stage startups.