More Governance Isn’t the Same As More Reliable
Hi all 👋
I spent two days at Energy Disruptors last week. The speakers came from different industries, at different stages, with different stories. But almost every one of them came back to the same thing at one point: team culture as the foundation of high performance.
It got me thinking about when culture actually gets tested. It isn't when things are moving. Momentum makes almost any team look good. The real test is the slower stretches, when deals take longer, timelines slip and the wins are further apart.
That's when you find out whether your team performs because of the culture you've built or because of the conditions around them. High performing teams don't wait for things to pick back up. They use the quiet to get sharper.
Most things take longer than we'd like. The question is what your team does with that time.
What's Happening in AI
Sinch put out some research worth sitting with: 74% of companies that launch an AI customer-service chatbot have to shut it down or roll it back.
The part that actually surprised me was that companies with the most mature AI governance had a higher rollback rate, 81%.
That's backwards from what everyone assumes. It's worth understanding why.
To be honest, governance and reliability solve two different problems.
Governance catches a failure after it's already happened: more policy, more review steps, more sign-offs. Reliability is what was supposed to stop the failure in the first place: the data, the integrations, the actual infrastructure AI gets built on top of.
Most companies are only investing in the first one. That's exactly why the failure rate went up instead of down.
So if your AI keeps breaking and the response has been to add more process around it, know that the process is actually only buying you a faster alert when something breaks. Not a reason it stops breaking.
This Week's Case Study
We built one project around closing that exact gap: an AI whose only job was to prove our own numbers wrong.
The client runs a large casual workforce on loaded rates, one flat number that stops being accurate once the shifts stop being even. Across 10,000+ timesheets a year, nobody was catching that by hand. So we set a second AI on the first one: break it if you can. It found a six-figure compliance exposure sitting in a handful of roster patterns nobody had flagged.
If you're running AI on anything that touches pay, compliance, or risk, the question worth asking isn't whether it's accurate. The valuable question is ‘what's checking its output?’.
→ Read the full case study.
A Useful Resource
One question I get quite often is: "How do I know if a process is actually worth automating?"
That's why we built the Administrative Tax Calculator. It shows you how much you're paying premium salaries to do low-value work, chasing updates, manual entry, bridging gaps between systems. Before you hire another person, see how much you're already losing to broken processes.
That's it for this week.
— Robbie
Get the latest tech news in your inbox
Hi all 👋
I spent two days at Energy Disruptors last week. The speakers came from different industries, at different stages, with different stories. But almost every one of them came back to the same thing at one point: team culture as the foundation of high performance.
It got me thinking about when culture actually gets tested. It isn't when things are moving. Momentum makes almost any team look good. The real test is the slower stretches, when deals take longer, timelines slip and the wins are further apart.
That's when you find out whether your team performs because of the culture you've built or because of the conditions around them. High performing teams don't wait for things to pick back up. They use the quiet to get sharper.
Most things take longer than we'd like. The question is what your team does with that time.
What's Happening in AI
Sinch put out some research worth sitting with: 74% of companies that launch an AI customer-service chatbot have to shut it down or roll it back.
The part that actually surprised me was that companies with the most mature AI governance had a higher rollback rate, 81%.
That's backwards from what everyone assumes. It's worth understanding why.
To be honest, governance and reliability solve two different problems.
Governance catches a failure after it's already happened: more policy, more review steps, more sign-offs. Reliability is what was supposed to stop the failure in the first place: the data, the integrations, the actual infrastructure AI gets built on top of.
Most companies are only investing in the first one. That's exactly why the failure rate went up instead of down.
So if your AI keeps breaking and the response has been to add more process around it, know that the process is actually only buying you a faster alert when something breaks. Not a reason it stops breaking.
This Week's Case Study
We built one project around closing that exact gap: an AI whose only job was to prove our own numbers wrong.
The client runs a large casual workforce on loaded rates, one flat number that stops being accurate once the shifts stop being even. Across 10,000+ timesheets a year, nobody was catching that by hand. So we set a second AI on the first one: break it if you can. It found a six-figure compliance exposure sitting in a handful of roster patterns nobody had flagged.
If you're running AI on anything that touches pay, compliance, or risk, the question worth asking isn't whether it's accurate. The valuable question is ‘what's checking its output?’.
→ Read the full case study.
A Useful Resource
One question I get quite often is: "How do I know if a process is actually worth automating?"
That's why we built the Administrative Tax Calculator. It shows you how much you're paying premium salaries to do low-value work, chasing updates, manual entry, bridging gaps between systems. Before you hire another person, see how much you're already losing to broken processes.
That's it for this week.
— Robbie
