Rebuilding engineering trust: the 90-day playbook.

Turnaround work is mostly about restoring a single thing: predictability. Stakeholders don't need miracles; they need to believe what you tell them.

When I joined the team responsible for our most critical supply chain platform, the situation was familiar: instability, eroding confidence, and a business that had stopped expecting much from IT. The technical problems were real, but they weren't the root cause. The root cause was a broken relationship.

Here is what the first 90 days looked like.

Days 1–30: Diagnose before you cut

The first month is not for fixing. It's for understanding — and resisting the temptation to act like you already know the answer.

I came in as an outsider, which is actually an asset in this situation. No political history. No pre-formed alliances. No prior wounds. That neutrality gives you access that insiders rarely have: people will tell you things they wouldn't say to someone already inside the system.

I used that window to listen. Engineers, product managers, business stakeholders, operations teams. Not to collect grievances, but to build a picture. Technical signals — incident frequency, deployment cadence, test coverage, mean time to recovery — give you one lens. The conversations give you the other. Together, they show you where the real fault lines are.

By day 30, I had a clear diagnosis. The technical debt was significant but manageable. The deeper problem was organizational: the team had been isolated from the business for so long that each side had developed a distorted view of the other.

Days 30–60: Tear down the wall

The business team had a story about IT, and it wasn't flattering. Every time they tried to engage on a technical challenge, they were pushed back. They were kept at a distance, given filtered information, and handed decisions rather than brought into them. Over time, they stopped trying.

What looked like apathy was actually accumulated frustration.

I made a deliberate choice to change that dynamic entirely. I brought the business onto the journey. We showed them exactly where the platform stood — the problems, the risks, the unknowns. Full transparency, no managed messaging. That's a risk, and people told me so. But there is no trust without transparency. You cannot build a real working relationship on a curated version of the truth.

The response surprised some of my team. The business didn't punish us for the honesty — they leaned in. They had been eager to help all along; they just hadn't been invited. When we started making decisions jointly, when their operational context began informing our technical choices, something shifted. They stopped treating our mistakes as evidence of incompetence and started treating them as the inevitable cost of hard bets taken with shared understanding.

That shift was the bedrock. Everything that came after depended on it.

The benefit wasn't only external. When engineers understood the operational consequences of their decisions — when they could see the downstream impact on the business — they made better technical choices. The environment became psychologically safer. Taking a calculated bet felt less like exposure when the people affected by the outcome were in the room when you made it.

Days 60–90: Install the habits

With trust established and the team no longer operating in survival mode, we could load up on the technical work that had been waiting.

We split the team into two focused streams. One stream held the line: keeping the platform stable, handling incidents, protecting the business from further disruption. The other stream drove modernization: addressing the most critical areas of technical debt, extending test coverage on the highest-risk components, improving deployment confidence.

We hardened the operational process. Observability had degraded — we restored it. Incident response was ad hoc — we made it rigorous. Quality gates that had been bypassed became non-negotiable again.

None of this was glamorous. All of it was necessary.

Two years later

The 90-day playbook doesn't fix a platform. It creates the conditions under which fixing is possible.

Two years on, the platform is fully stable. We deliver new capabilities — including AI-based use cases — at a pace the business couldn't have imagined when I arrived. Performance has improved by a factor of three. Scalability is no longer a concern for the projected demand curve.

The turnaround is not finished. We are still modernizing, still deepening the work. But it is no longer a question of survival. It is a long-term investment in future prosperity — and the business is making that investment alongside us, not watching from a distance.

That is the real measure of whether the 90 days worked.