The technology is ready. Your organisation isn’t.
We’ve watched countless public sector AI initiatives stall at the same point. Not because the systems don’t work. Not because the data isn’t available.
Because nobody can actually come to a decision.
The Approval Labyrinth
Here’s what it looks like in practice: perhaps five layers of management. Probably more than ten different department heads. Each one needs to weigh in. Each one has their own priorities.
None of them are accountable for the final call.
When everyone’s involved, nobody’s responsible. The decision doesn’t move through the approval process. It dissolves into it.
Reaching consensus becomes impossible. Not because people disagree, but because they’re not empowered to agree.
The Knowledge Gap Nobody Mentions
Here’s where it gets worse. Few of those ten department heads are technical people. So they misunderstand the actual risks. They worry excessively about things that don’t need worrying about.
Data security becomes the conversation-stopper. Even when technical safeguards are already in place. Even when the protections exist.
Meanwhile, the genuine risks go completely unexamined.
The Paradox Creating Real Danger
Whilst committees debate data security protocols, your civil servants have already solved their problem.
They’re using ChatGPT on their mobile phones.
Shadow IT now accounts for approximately half of all workplace applications. ChatGPT has become the top unauthorised app in organisations worldwide.
One in five companies believe sensitive data has already been exposed to GenAI applications by employees without authorisation.
The excessive control you implemented to reduce risk? It’s creating far more dangerous exposures than properly governed AI tools ever would.
People aren’t waiting for your approval process. They’re working around it.
The Cost Nobody’s Calculating
The UK government spends £2.3 billion annually just keeping legacy systems running. That’s nearly half the technology budget.
NHS England reported 123 major system outages in 2024. Each one forced staff back to paper-based processes.
Yet organisations remain fixated on the theoretical risks of implementing AI rather than documenting the massive, measurable costs of maintaining the status quo.
The White House explicitly acknowledged in April 2025 that federal agencies face a “widening gap in adopting AI… largely due to unnecessary bureaucracy and outdated procurement processes.”
Even governments recognise the paralysis problem.
The Translation Gap
There’s a fundamental disconnect between IT teams making decisions in their tech suites and the rest of the organisation trying to do their jobs.
Tech people only know tech. Department leaders only know their departments.
They don’t speak the same language. They don’t understand each other’s constraints.
IT builds tools based on technical specifications. Departments need tools that fit how government work actually gets done.
Without translators bridging that divide, you get systems nobody uses and workarounds that create the exact risks everyone feared.
What Actually Needs to Change
The solution isn’t better technology. It’s better decision architecture.
Clear accountability. Fewer approval layers. Technical literacy among decision makers. Operational understanding among technical teams.
Most importantly: recognition that the current system isn’t protecting you from risk.
It’s guaranteeing it.
Your staff are already using AI. The question isn’t whether to allow it.
The question is whether you’ll govern it properly or keep pretending your approval process works whilst people route around it on their personal devices.
The technology problem is solved. The decision problem remains.