I've worked across the full spectrum of company maturity. I started in a startup from zero, then lived through several iterations and stages of growth until that business became Hitachi Solutions Global, with thousands of people. I also helped take a large midsize business into multibillion-revenue territory. That arc — startup to global leader, midsize to multibillion — is what shaped how I see technology investments today. It informs my perspective, and it shows I can build and manage across the spectrum. The right move depends on where a company actually is, not where a slide deck says it should be.
I've also worked with large enterprises, including Fortune 1000 companies. But I'll be honest: I especially love working with small and mid-sized businesses.
Here's what each world has taught me, and why it matters for AI-led engineering and transformation right now.
Company maturity continuum from startup to midsize to global or multibillion, with the levers that matter most at each stage.
- StartupFocus & speedFew AI bets; measure, then expand
- MidsizeBoth at onceProcess to scale, without losing the special sauce
- Global / multibillionOne windowStandards that travel; build on what already exists
Lesson 1: In SMBs, the right move shows up fast
In a small or mid-sized business, there's nowhere to hide. When you make the right move, you see it quickly in the work and in the numbers. When you make the wrong one, you see that too.
That's what I love about it. At Iteration2 we grew without venture or debt funding. Every decision had to pay for itself. That kind of discipline is healthy, and it's exactly the mindset I'd bring to AI in a smaller company: pick one or two places where AI changes the work, measure them, and expand from there.
Lesson 2: Enterprises need it to scale and to build on what they have
Big enterprises want AI too. They just need it to scale across the organization and build on what they've already invested in. That's a very different problem. You're not starting from a blank page; you're starting from years of systems, contracts, processes and people.
In enterprise work, integration and architecture matter as much as the idea. You need cross-functional pods of architects, engineers and designers who can work at the level of complexity enterprises bring, and you need engineering pillars that hold up at scale: quality, performance, feasibility, scalability, security and maintainability.
Lesson 3: Maturity changes the problem, not the standard
Building from a startup to Hitachi Solutions Global, and later helping take a large midsize business into multibillion-revenue territory, taught me the same lesson from different angles. Early-stage companies need focus and speed. Global organizations need one window for the customer and engineering standards that travel. Midsize companies in the middle often need both at once: enough process to scale, without so much process that they lose the special sauce that got them there.
That range is why I don't preach one playbook. The maturity of the business tells you which levers matter most. The standards — quality, honesty about scope, and a team that can ship — stay the same.
Lesson 4: The special sauce is the same in both
Whether it's a 50-person firm or a global company, the question I ask is the same: what's your special sauce, the thing that makes you great? Then, how do we amplify it with technology?
A lot of small and mid-sized businesses are ripe for that kind of change right now. They often know their customers better than anyone, and they move faster than a big company can. AI lets them turn that closeness into speed and scale they couldn't afford before. Enterprises have the special sauce too. It's just often buried under layers of process, and the work is helping it come through.
Lesson 5: Don't let price run the conversation
With smaller companies especially, conversations can drift toward cost. I understand why. But in product engineering, what matters most isn't how much you're spending. It's what you can build that grows the top line. If the conversation is all about price, you end up competing with every low-cost provider out there, from big offshore shops to small boutiques. I'd rather talk about what we can build together.
Lesson 6: Scope honestly, whatever the size
In enterprises, the risk is committing to a complex technical solution before anyone has assessed feasibility and the capabilities required. In SMBs, the risk is trying to do too much at once with a small team. The fix is the same in both: scope honestly, prototype quickly, and tell the truth early about what's ready and what isn't.
Lesson 7: Partnerships look different, but they matter to both
I've spent years building partnerships with the hyperscalers, Microsoft first and then Google and AWS. For an enterprise, those relationships often come with co-sell programs and big shared accounts. For a smaller business, they can be a growth channel that would otherwise be out of reach. Either way, the platforms can be your best growth channel or a very expensive distraction. The difference is whether you build real value on top of them.
Lesson 8: People decide the outcome
At either size, the technology is rarely what decides whether a project works. People do. In a small company, a single great hire changes everything. In an enterprise, the challenge is getting thousands of people to adopt something new. Either way, I want people checking the work and owning the result, especially with AI.
What I'd tell leaders on each side
If you run a small or mid-sized business: you're probably closer to an AI advantage than you think. Pick a few places where it changes the work, prove them, and grow from there. Your speed is your edge.
If you lead an enterprise: build on what you've already invested in. Start with a strong lead pod, a repeatable win formula, and honest scoping, then scale it across the organization.
For both: know your special sauce, tie AI to the P&L, and build the team that can ship it. Meet the company where it is on the maturity curve, then help it grow from there.
The tools keep changing, now faster with AI. What makes them work hasn't. Good people, honest partnerships, and the patience to build something that lasts, at any size.
I write about AI-led engineering and transformation for companies of every size. More at simon-chan.com.