The trust-first turnaround that 4×’d adoption of PwC’s internal audit platform
Stuck at 100 engagements. Everyone said it needed more features.
It didn’t. The moment the product tripped them up, auditors quietly fell back to Excel.
And the auditors’ reports land in front of the shareholders of Fortune 500 firms. The product had zero tolerance for failure.
So I made the contrarian bet: reliability first, ahead of the features everyone was asking for.
When I took over PwC’s internal audit platform, it was stuck at about 100 engagements, and the easy assumption was that it needed more features. The real problem sat underneath. Releases went out unprepared, changes landed unplanned, and quality slipped, so the audit teams meant to use it couldn’t rely on it. The product had also been shaped from the builder’s point of view rather than the auditor’s, so it didn’t fit how they actually work.
I learned that the slow way. I sat on the troubleshooting calls where users hit real problems, and I stayed close to the Business Product Manager to line up what I was hearing with where the business needed to go. The pattern was clear: trust had to come before features. An audit team won’t adopt a tool they’ve watched break.
- Role
- Technical Product Manager · owned research, delivery structure, and roadmap, partnered with the Business PM
- Customer & adoption
- PwC audit and assurance teams · drove internal adoption from 100 to 500 Fortune 500 engagements
- Team
- Cross-functional: 6 engineers, 5 QA, 1 designer, plus US release teams
- Timeline
- Jul 2022 – Dec 2024 · Belfast, then Gurugram
- Surface
- PwC’s internal audit platform · SOC 1 / SOC 2 work on Fortune 500 engagements
The problem & who had it
- The platform was stuck at ~100 Fortune 500 engagements, and the assumption was that it needed more features.
- Two things held it back, and they compounded: unreliable delivery (unprepared releases, unplanned changes, slipping quality) and a product shaped without the auditor’s workflow in mind.
- The users were auditors on SOC 1 / SOC 2 engagements at Fortune 500 institutions, including global investment banks and card networks. Demanding users on regulated work.
- When the setup or the sequencing tripped them up, auditors did the rational thing: they fell back to Excel and shared documents, the levers they already trusted. Every fallback was adoption lost.
- And the work left zero tolerance for failure. These reports ship to the shareholders of Fortune 500 firms after publishing, so a product that wobbled was one teams couldn’t risk.
What I found
- I didn’t trust the “needs more features” assumption, so I went to the source.
- Troubleshooting calls: I sat on the calls where users hit real problems and watched where the workflow actually broke.
- Close partnership with the Business PM: I lined up what I heard from users with where the business needed the product to go.
- The read: trust had to come before features. A team won’t adopt a tool they’ve watched break.
The fix
Trust before features sounds clean on a slide. In practice you can’t freeze a roadmap. Stakeholders still want features, and a real org keeps shipping them. So there was no freeze. The work was winning the same argument over and over: this reliability fix goes ahead of that feature, and here’s why. The auditors had already voted with their feet. The moment the tool tripped them up they fell back to Excel, so a new feature on a shaky base wouldn’t get used anyway. The incident and release data said the same thing, that the platform couldn’t safely take on more until it stopped breaking. And the stakes settled it: these reports land in front of the shareholders of Fortune 500 firms, where a tool that wobbles is not survivable. Feature by feature, reliability kept winning the slot.
- Stabilised delivery first. Release planning, QA, testing and telemetry, and clearer prioritization, so the platform stopped breaking and teams could rely on it.
- Reframed the product around the auditor. The roadmap got rebuilt around the real workflow the research surfaced, away from the builder’s original assumptions.
- Automated the grind, once trust was restored. With the platform stable, I drove the control-sequencing engine and Smart Prefill / Guided Roll-Forward to kill the repetitive manual setup.
Fixing trust before features took the platform from 100 to 500 engagements, which is the ~$100M a year the firm now saves.
The outcome
- Adoption grew from 100 to 500 Fortune 500 engagements, a 400% increase.
- The platform delivered roughly $100M in annual audit efficiency value.
- Automation removed 40,000+ person-hours of manual work a year and cut audit setup from 6 hours to 1.
- Production incidents fell 20%, post-release issues halved, and platform NPS rose 2 points.
- The work earned PwC’s Over and Beyond award in 2023.
What I'd do next time
- Make the troubleshooting-call research a standing cadence, so the next workflow break shows up before it dents adoption.
- Instrument reliability and adoption per engagement, so the next trust problem surfaces in telemetry rather than in a quarterly review.
- Package the stabilise-then-automate playbook for other internal PwC products facing the same adoption wall.