Loving every drop of AI-powered project intelligence - a revealing conversation with Anglian Water: part II

On last week’s episode of ‘Loving every drop of AI-powered project intelligence with Anglian Water’...

Towards the end of last year, Anglian Water’s Head of Delivery Portfolio Management, Euan Black, sat down with nPlan CEO Dev Amratia to record a discussion on AW’s use of AI-led assurance and risk management. The wider context for their conversation was the impending start of AMP8, the UK water regulator’s next five-year investment cycle (industry regulator Ofwat is recommending a record £96 billion allocation of public money to sixteen water companies for AMP8 - a significant increase on AMP7’s £51bn outlay). 

In the first post in this three-part series, I covered the highlights from Euan and Dev’s discussion that related to Anglian Water’s quest to ramp up project delivery without ramping up waste, and the various ways in which it is now using and getting value from AI. Today, I’ve curated a selection of clips related to the cultural challenges Anglian Water has experienced in rolling out nPlan's technology.

Good luck with that confirmation bias

Midway through the interview, Dev asked Euan to explain how Anglian Water uses nPlan - a fairly standard question (you might think!) but Euan's answer went beyond process in a rather revealing way - check it out:

Hopefully you noticed that Euan didn’t restrict himself to talking solely about process in his answer; there were a few other tidbits in there. For example, Euan mentioned that its new AI tooling enables Anglian Water to foresee problems–such as those “with kiosk or pipe deliveries” two years(!) ahead of time. However, Euan also mentioned that AI has not (yet) done away with the influence of confirmation bias: “it’s amazing to see…if a bit of technology tells you and backs you up…as a human we seem to believe it more or we will at least accept it”. nPlan has repeatedly argued that neutralising biases including confirmation bias is one of the most valuable aspects of AI-led assurance; although the context here is positive, Euan’s comment shows that AI-powered technology alone will not protect against the cognitive biases that so often lead to major project delivery failure - there must also be a concerted organisational effort to recognise and combat them. 

“We got push-back pretty early on…”

The crucial role played by workplace culture in the adoption of - and value derived from - AI-powered solutions is also the focus of our next clip, in which Dev asks Euan about the steps Anglian Water has taken to shape attitudes around nPlan’s technology. Take a look:

“We’ve put a lot of effort, more effort into the behaviour side of things than probably the technology side…the behaviour bit is critical and we got push-back, open and honest. We got push-back pretty early on - and then we got some quick wins and the quick wins start to feed through the organisation, and then it’s ‘oh, can we see nPlan?’ instead of ‘we don’t want to see’...”

Euan makes it clear that even though PMO leaders put a huge amount of emphasis on preparing the cultural ground for the use of AI-led solutions, it was only when project deliverers saw AI make an impact (‘some quick wins’) that they started to embrace it. Crucially, the result of that change in attitudes has been incredibly marked, with project teams going from rejecting AI assessments to wanting to see those assessments even before the PMO team performed their own review.

Small insights that add up to big savings

So what were those ‘quick wins’ which turned Anglian Water project deliverers around on nPlan? Fortunately we don’t have to guess, as Euan gave an example of one of them later on in the interview. Have a watch: 

“The first [schedules] we put through [nPlan], the funniest thing that popped out was, ‘Leading biggest risk: emergency showers’...we looked back…we never actually installed an emergency shower–which we supply the water to–on time. It was such a small delay that we would never have picked it up…”

Firstly, this anecdote is a great example of salience bias: time and time again we see that project teams predominantly focus on the flashy/showy/sexy risks and miss opportunities to avoid delays and claw back time by addressing more ‘boring’ risks. This is certainly one of those cognitive biases which AI-led assurance is great at countering. 

Secondly, as Euan says, in the context of AMP8, and the huge number of projects Anglian Water is undertaking to meet its commitments for the upcoming investment cycle, there’s no such thing as a small win - actioning even seemingly insignificant AI insights such as this one unlock huge savings. 

“We also found projects that we were delaying as a client; our contractor had finished, but we weren’t doing our bit at the end of it…it’s a simple tweak. Saved multiple days, probably weeks on each project over five or six thousand projects in the next five years, that's a huge saving.”

But wait, there’s more!

As juicy as all of this is, there’s more! Check back here next week for the third part of this three-part series, when we’ll be finding out what keeps Euan up at night, what would happen to his work-life if AI-led assurance and risk management were suddenly switched off, and what he ultimately hopes to get out of using nPlan long-term. See you soon! 👋