The new bottleneck in projects: decision-making

• Project delivery costs have dropped by up to 80%, but governance costs haven’t changed in 20 years.

• Many projects take longer to approve than to build—governance is now the main cause of delay and waste.

• A distributed “proof of value” model empowers faster, lower-risk decisions where the benefit is clearest.

The cost of automation is falling fast

The cost of delivering automation has plummeted, thanks to AI-powered implementation teams.

We’re now routinely delivering projects that would have cost $300,000 two years ago for around $50,000—or even less, depending on the use case. The cost of testing the value of a solution at small scale is often under $10,000.

These projects are being delivered in weeks, not months. Where a 12-month timeline was standard, we’re now seeing completion in as little as 3 months—sometimes just a few weeks.

Governance is now the bottleneck

Despite this acceleration in delivery, our ability to realise value hasn’t kept pace. The cost of administering projects—funding approvals, business cases, governance processes—is still what it was 20 years ago. This was already an issue before AI came along, but now it’s glaring.

My honest answer to the question “how long will it take?” is: about a quarter of the time it takes to get a decision.

That’s the absurd reality. The funding and governance processes are now costing more than the actual delivery of the project. And the opportunity cost is even greater—we’re missing out on real, measurable improvement simply because we’re wasting time.

If we think financially: if implementation now costs a quarter of what it used to, we should be doing four times as many projects. But we're not. We're stuck in the same old decision-making cycles.

What we can learn from distributed governance

I worked with one of the world’s largest retailers in the 2000s that had already solved this problem—long before AI was in the picture. They had replaced their centralised assessment process with a distributed model.

Every employee, from the CEO to a retail assistant, could spend a fixed amount (equivalent to about $10,000 today) to run a “proof of value” initiative. Any manager could approve one, using a simple five-point checklist. That was the entire process.

Most of these proofs of value involved simple, human-emulated prototypes—often carried out by university students on break. The teams used paper checklists and phone calls to simulate what would eventually become software or automated workflows.

One such initiative demonstrated an 80% reduction in default rates for extended warranty underwriting. Students physically walked from checkouts to the back office, stapled paperwork together, and passed it to someone who phoned through the data to another team playing the role of the underwriter.

It wasn’t fancy, but it worked—and it was indisputable. No theory-based business case required. The opportunity was proven in action, not in a spreadsheet.

Let benefits drive the budget, not the other way around

This model turns traditional governance on its head.

Instead of building complex business cases around theoretical benefits and painstakingly estimating project costs, we flip the process. We run small, cheap experiments, and base funding decisions on demonstrated outcomes.

A conversation becomes: “You’ve shown you can save us $4 million a year in penalties, and our underwriter agrees. Can you deliver a solution for $4 million?”

That’s a far more grounded and practical approach than what most governance processes demand today.

Why distributed decision-making works

Distributed decision-making doesn’t just reduce cost and speed up delivery—it shifts responsibility to those best placed to own it. The people closest to the benefit are also closest to the information needed to make the decision.

This isn’t just logical—it’s supported by science. Research consistently shows that decisions are most cost-effective when made at the source of the relevant information.

Or, as they say in mining: decisions cost less at the coalface.

When you're ready, here's how I can help:

Andrew Walker 
10 years AI strategy & implementation for executives in staff-heavy organisations, often with mobile workforces.

Did someone forward this email to you? Want your own subscription? Head over here and sign yourself right up!

Back issues available here.