By Stephen Ledwith April 21, 2026
A few years ago, I wrote about scaling an engineering team from 6 to over 100 — the lessons learned, the hard pivots, and the patterns that held up under real pressure. That article was grounded in a specific kind of scaling: more people, more process, more structure.
The game has changed.
Today, the most important scaling question for engineering leaders isn’t “how do I grow headcount to match our ambitions?” It’s “how do I multiply the output of the team I already have?” AI tools have fundamentally shifted what’s possible — and leaders who understand that are building leaner, faster, and more resilient organizations than I thought possible even two years ago.
This is the follow-on article I didn’t know I’d be writing. Here’s how AI is reshaping every dimension of the team-scaling challenge.
1. Building a Strong Foundation — With AI in the Stack From Day One
In the original piece, I talked about establishing core values and culture before you scale. That still holds. What’s changed is the expectation you need to set from the start: AI fluency is now a baseline team competency, not a specialty.
When you’re hiring or onboarding, the question is no longer just “can this person code?” It’s “can this person work effectively alongside AI tools?” The engineers who thrive in this environment aren’t necessarily the most technically brilliant — they’re the ones who’ve learned to use AI as a thought partner, know how to critically evaluate AI-generated output, and don’t waste cycles doing work that a well-prompted model can do in seconds.
Build that expectation into your culture early. It’s much harder to retrofit.
2. Onboarding: From One Day to One Hour
At eXp Realty, we compressed onboarding from 30 days to one through automation and documentation. I was proud of that at the time. Now, teams I work with are doing it in hours.
AI-powered onboarding is the lever most organizations are underusing. Tools that can answer questions about the codebase in natural language, generate contextual documentation on demand, and walk a new hire through the architecture of an unfamiliar system have removed one of the biggest friction points in scaling: the knowledge transfer bottleneck.
When a new engineer can ask “how does our authentication service work, and where are the edge cases?” and get a well-sourced answer in seconds — rather than waiting two weeks for the right senior engineer to have time — your onboarding ceiling goes away. The constraint shifts from knowledge transfer to judgment development. That’s a much better problem to have.
3. Communication at Scale — AI as the Signal Filter
One of the hardest things about scaling a team is that communication complexity grows faster than headcount. A team of 10 has manageable coordination overhead. A team of 50 can be crippled by it if you don’t architect the communication layer carefully.
AI is changing this in practical ways:
- Meeting summaries and async-first workflows: Teams that used to rely on real-time syncs are using AI to summarize discussions, surface decisions, and flag action items — reducing meeting load without losing alignment.
- Context delivery: Engineers no longer need to dig through Slack history or ask around to understand why a decision was made. AI-assisted search and summarization tools surface that context on demand.
- Escalation filtering: AI can triage noise — support tickets, monitoring alerts, code review queues — so that human attention goes where it’s actually needed.
The teams scaling best right now have figured out how to use AI not just for execution work, but to reduce the coordination tax that kills velocity as organizations grow.
4. Leadership Evolution: From Gut Feel to Informed Judgment
Delegating to team leads and giving them real authority was one of the central themes in the original article. That principle hasn’t changed. What’s changed is the quality of information available to those leaders.
AI-driven engineering dashboards now give team leads visibility into cycle time, deployment frequency, PR review patterns, and team health signals — in real time, without manual reporting. Leaders who used to manage by intuition and informal check-ins now have data to back those instincts up, or challenge them.
The leaders I see struggling with this aren’t the ones who distrust the data — they’re the ones who let the data replace the conversations. AI gives you better questions to ask. It doesn’t replace the act of asking them. The one-on-ones, the skip-levels, the informal pulse checks — those still matter. AI just makes sure you walk into them better prepared.
5. Process Optimization: AI as the Default First Pass
Agile methodologies gave us the framework. AI is now handling a significant chunk of the execution within that framework.
Code review is the most obvious example. AI-assisted review tools catch issues that human reviewers miss — not because humans aren’t capable, but because humans are reviewing dozens of PRs and cognitive load is real. AI doesn’t get tired. It doesn’t skim. It applies the same rigor to the hundredth PR of the week as the first.
Testing, documentation, sprint planning, architectural diagramming — all of these have AI-assist options that are genuinely useful today. The teams capturing the most value aren’t trying to do everything with AI at once. They’re identifying the highest-friction steps in their development cycle and systematically cutting that friction with the right tools.
Start there. The ROI is immediate and it builds team confidence in AI as a force multiplier rather than a threat.
6. The New Headcount Math
Here’s the implication that I want to be direct about, because I think a lot of leaders are dancing around it: AI is changing the relationship between team size and team output.
A well-structured team of 15 engineers with strong AI tooling and clear AI-assisted workflows can deliver what a team of 40 would have delivered three years ago. I’ve seen this. It’s not theoretical.
This creates a genuine leadership dilemma. Do you maintain headcount and dramatically increase output? Or do you right-size the team and invest the savings elsewhere? There’s no universal right answer — it depends on your growth stage, your market, and your people strategy. But the question has to be on the table.
What I’d caution against is the approach I’ve seen some organizations take: adding AI tooling without adjusting any expectations. If you give your team AI capabilities and still plan and staff as if nothing changed, you’re leaving enormous value on the table. Worse, you’re going to confuse your team about what “good” looks like.
7. Continuous Feedback — Better Input, Same Conversations
Feedback mechanisms were a key pillar in the original article — check-ins, retrospectives, anonymous surveys. AI can improve all of these, but it cannot replace them.
What AI does well here is pattern recognition at scale. Sentiment analysis across team communications, recurring themes in retrospectives, correlation between process changes and velocity shifts — these are things AI can surface that a manager simply can’t track manually across a large team.
What AI cannot do is build the trust that makes feedback honest. It can tell you that a third of your team’s Slack messages contain frustration signals. It can’t make people feel safe enough to tell you why. That’s still a human leadership job.
Use AI to find the signal. Use your leadership skills to act on it.
8. Scaling Infrastructure With AI Ops
Infrastructure scaling was always the unsung hero of team growth. As your team scales, your systems have to scale with them — and historically, that’s required significant DevOps investment.
AI ops tooling is changing the economics here. Anomaly detection, self-healing infrastructure, AI-assisted incident response — these capabilities are reducing the toil burden on platform and infrastructure teams meaningfully. Teams that used to need a dedicated SRE for every 15 engineers are operating with leaner on-call rotations and faster mean-time-to-recovery.
This doesn’t eliminate the need for strong infrastructure leadership. It raises the bar for it. When AI is handling the routine monitoring and first-response, your senior infrastructure engineers are freed up for the architecture decisions that actually matter at scale.
The Through Line
Looking back at the original article, the core argument was that scaling a team is fundamentally about people, culture, and process — not just headcount. That’s still true. What AI has done is raise the ceiling on what a well-led, well-structured team can achieve without adding bodies.
The leaders who are going to win the next phase of this aren’t the ones who adopt the most AI tools. They’re the ones who stay clear on what AI is for — accelerating execution, reducing friction, surfacing information — and what it isn’t for: replacing judgment, skipping culture, or avoiding the hard conversations that make teams great.
The playbook has a new chapter. The fundamentals still apply.
Stephen Ledwith is a technology executive and consultant with over 20 years of experience scaling engineering organizations across real estate, fintech, and enterprise software. For more on building and leading high-performing technology teams, subscribe to the blog or reach out directly.
