AI Is Closing the Gap. Open the Door.
As both a contractor and a GS-15, I’ve written Authority to Operate packages. I’ve built and maintained the systems those ATOs protect. I’ve designed interfaces for federal users, led engineering teams inside and outside of government, and shipped production software that real people depend on. I’ve sat in the rooms where architecture decisions get made, where six-billion-dollar IT budgets get finalized, where compliance meets delivery, where the work actually happens.
And yet, when my company pursues a contract, the path forward narrows quickly. The evaluation criteria weight past performance at dollar thresholds we haven’t hit as a prime, and the company’s history over the team’s experience. The result is predictable — even with a strong technical approach, you’re rated lower because the entity is young. And the guidance is consistent: go build your past performance through a partner.
I understand why. Zero tolerance for failure is the right instinct when you’re delivering services to veterans or managing national security infrastructure. But over time, that instinct has been translated into a proxy: if your company is new, you will fail. The company is not the CEO, the evaluation reads. It doesn’t matter that your team has decades of federal experience — the corporate entity hasn’t done it at this dollar threshold. Go partner.
The partner pipeline
So you partner. We’re doing exactly that — building a mentor-protege agreement and joint venture with [PARTNER PLACEHOLDER], who has been the kind of partner this system is supposed to produce. They invest in our growth, treat us as equals at the table, and genuinely want to see us compete independently. That’s how the program is designed to work.
But not every partnership looks like ours. Too often, the larger company dictates the tools, the processes, the staffing models, and the delivery culture — because they hold the contract and can revoke access at any time. You don’t get to challenge the approach or propose a fundamentally different way of working. You adopt the prime’s playbook, or you lose your seat.
The result: by the time a small business “graduates,” it has been shaped into a version of the company it was supposed to be an alternative to. The government pays small business rates but gets large business culture. Meanwhile, the small business designation in IT services covers companies up to $34 million in annual revenue. Some have been “small businesses” for twenty years, with deep incumbent positions and past performance that makes them nearly unbeatable. We actually support recent efforts to investigate and reform the 8(a) program — the integrity of these set-asides matters to companies like ours who rely on them to compete fairly.
But reform alone doesn’t solve the deeper problem. When a solicitation is set aside for small business, the competition still isn’t between new approaches — it’s between established players and new entrants whose team’s experience doesn’t count because the entity is too young.
This matters more now than ever. AI is fundamentally changing the relationship between team size and delivery capacity. At our company, this isn’t an add-on — it’s how we operate:
- Every person has an AI budget and is encouraged to integrate it into how they work — engineering, design, accounting, operations
- Our engineers run coding agents across multiple terminals for hours at a time, producing higher quality code at a pace that would have required a team three or four times the size two years ago
- I rebuilt a production federal case management platform serving fifty-plus users with ten terabytes of data in two weeks
The acquisition system still uses team size and corporate revenue as proxies for delivery confidence. Those proxies were already outdated. AI is increasingly making them obsolete.
A path forward
I’m not asking for the bar to be lowered. I’m asking for it to be recalibrated.
Evaluate for potential, not just history. When a company’s leadership has built the systems, written the ATOs, and led the teams — weight that. Create evaluation criteria that distinguish between a new company with no relevant experience and a new company whose team has been building and shipping federal software for years.
Demand higher output and expect lower cost. Higher cost doesn’t mean better delivery — not anymore. AI is driving faster, higher-quality output from smaller teams with lower overhead. That’s not a lesson learned from past failures; it’s the new reality. Evaluate accordingly: expect more for less from firms that have embraced it.
Recognize that the disruptors haven’t been through the pipeline yet. The companies that will deliver something genuinely different are not the ones who’ve spent a decade conforming to a prime’s delivery model. They have low overhead, high motivation, direct experience, and now — with AI — the capacity to match incumbents on delivery while moving at a speed incumbents can’t.
Build competitive disruption into the contract lifecycle. Waiting three to five years for a recompete isn’t competition — it’s renewal. Google has killed over 290 products — and those are just the public ones — because sustained internal pressure is what keeps the work from going stale.
Carve out $2 to $5 million task orders during a large contract and let new companies build alongside the incumbent — even in sandboxes. With AI, a small company can deliver a meaningful proof of concept at a fraction of what it would have cost five years ago. If the result is better, integrate it. If it isn’t, the cost was small and the incumbent was kept honest. If a company misrepresented its capabilities — remove them and don’t bring them back.
The risk isn’t giving a new company a chance. The risk is another contract cycle where nothing changes.
To the SES leaders setting acquisition strategy, the political appointees shaping agency priorities, and the contracting officers writing the evaluations: give companies like ours a real shot. A right-sized opportunity to prove that different teams, with different approaches, deliver different results. The capability is there. The question is whether you’re willing to let it through.
Ari Perez is the CEO of Coa Solutions, an SBA-certified SDVOSB and 8(a) small business delivering federal digital services and healthcare AI. A West Point graduate and combat veteran, Ari has stayed focused on delivery across administrations — whether deployed overseas as an Army officer or building federal systems at Deloitte, Accenture Federal Services, and the Executive Office of the President.