What happens when a non-engineer gets invited to an AI hackathon?

I had just landed at Gatwick after a week in the sunshine when an email from our CTO popped into my inbox. “It’s finally here – the Nous AI Hackathon! Briefing at 9am tomorrow”. 

Surely he invited me by mistake, I think. But no. Turns out, it’s a cross-functional thing—the full Nous salad. Hackathon? I’ve got some Googling to do while I unpack my suitcase. 

Just after 9am, I’m meeting up with the team – engineers, designers, ops people, and other marketing folks – in the basement of our local coffee house. 

Our engineers are well clued up on AI already, but in the next 48 hours, we've got to hack (AKA design, build, and present) two new, functioning AI-powered features that could be game-changing for Nous. And that means bringing together all the brains in the business to collaborate.

We’re split into competing teams. My team is focussing on a new solution for a common customer problem, while our rivals will apply AI to solve an operational problem. 

After the allotted 48 hours, each team will present its functioning feature to be judged, and a winner will be picked based on three things: how well they’ve embraced GenAI, their style and presentation, and the overall business value of their idea.

With unlimited coffee and pastries in hand, we set to work. 

‍Team One: Using AI to solve an operational problem 

The role of any Ops team is to make sure a business runs as smoothly and effectively as possible. So what’s their latest challenge now that Nous is saving real people real money?

In order to find someone a better energy, mobile, or broadband deal you need to gather a set of information about their household setup. Our current signup flow captures lots of what we need to know, but in many cases, we ask some follow-up questions.

Nous is all about making life simpler and fairer. We want every interaction with a customer to meet that test. But developing a great relationship, understanding the intricacies of someone’s household, and knowing what’s best for them is time-consuming. After all, that’s why we’ve got a whole team dedicated to that work! When customer conversations move between different operators over days and shift patterns – a classic challenge of any service operation, it can quickly get complicated. 

Team One had a difficult problem to solve, but thankfully they’re smart cookies. Within a few hours, the team had a solid concept, design ideas, and a plan to make it all work. Within two days, they had built a new AI feature and a show-stopping presentation to show it off… 

‍

Team Two: Using AI to solve a customer problem

Modern life bombards us with bills. They’re stressful, they’re boring, and sometimes they’re downright baffling. I haven’t met many people who look forward to reading their bills, and many don’t read them at all. But burying your bills causes problems – payments get missed, errors go unnoticed, and anxiety lurks. 

But with AI by your side, perhaps it doesn’t have to be this way… 

Team Two’s mission: take the BS out of bills. After identifying the most important questions customers have about their bills, we began thinking of ways we could use AI to present this information to customers. The engineers started building, while the non-engineers got to work designing and naming our new product, and (of course) having a lot of fun filming the day for our TikTok channel. 

‍

But who won?  

Both teams presented their ideas, explained their thinking, and showed (in real-time) how their product worked. Nous CEO Greg Marsh, after much deliberation, took to the stage to announce the winner.

“For AI ninjitsu, i.e. to what extent you’ve embraced large language models, the teams are neck-and-neck, both scoring 4.5 out of a max 6 points.

“When I think about style and delivery, it’s a tough call. But due to John’s am-dram performance, I’m giving Team One 5 points and 4 points to Team Two. Even though Team Two’s slides and design treatment were better...

“But on to business value, AKA how useful the tool is today. Team One is the winner, with 4 points to Team Two at 3.5 points. The problem solved by Team One will get better and smarter and over time could radically change how we provide best-in-class customer service. Team Two, your idea has the potential to be game-changing – you just needed a bit more time.

So there we have it. Team One won with their time-saving, stress-beating approach to customer service. And the prize? A giant cheque for Team One with the promise of a free lunch (which I’m still hoping they’ll share).

‍

Funny to think that just 48 hours ago I didn’t even know what a hackathon was… 

And while it obviously wouldn’t be one without a gang of genius engineers, it was also great to learn that non-developers can play a crucial part as well. The team agreed there was an intensely satisfying buzz created by the cross-functional collaboration… or was that all the coffee?

Excited to see these features in action? Follow us on Twitter to be kept in the loop.