Data challenges for AI: Woodhurst’s podcast debut

Data challenges for AI: Woodhurst’s podcast debut

Ben Nadel
by Ben Nadel

Last week I was interviewed by Kyle Winterbottom for the upcoming Data Leaders of the North podcast. The podcast series will cover a range of topics on and around data. From building a Data Eco-System to creating a Data Culture; expanding into large-scale Tech Change and productionising AI. I was there to talk about our recent AI white-paper that touches on many of these themes.

As a long time listener of podcasts like this one, it was fun to be behind the mic for the first time and chat to Kyle about what I’m seeing as some of the key opportunities and challenges with AI across Financial Services.

Part of the reason we wrote our AI whitepaper was to highlight the pitfalls of using AI in a practical way. It isn’t about evangelising a process or a rigid methodology, but appreciating that AI is a new beast that, with the right organisational direction, data strategy and technical landscape, can be tamed.

You can sell a story about AI changing the world and delivering millions upon millions in benefits, but the reality is that to unlock this you need to put in time and effort up front to truly analyse the solution, and you need to have a business that is conducive to AI technology (and digital transformation in general).

We talked about the challenge of ensuring that senior leaders are well versed in AI, as without the right level background on the technology it can be difficult to see the scale of the opportunities and the likely implementation challenges. This would make it much harder to secure budget approval for AI projects, particularly those with a long-term outlook.

For those that are suitably briefed on the opportunities that the technology affords, we discussed how to navigate leadership that are either too keen or too cautious to embrace AI. We feel that by following a rigorous upfront assessment that properly defines the business problem, challenges the suitability of AI or machine learning and identifies the key measures of success, you can control over-zealous enthusiasm or influence leaders that are holding back.

Kyle had some great insight on the people aspects of AI and particularly how banks can compete with tech companies to appear as “sexy” a place to work as possible to a data scientist wanting to play with the latest cutting-edge tools. 

This comes back to one of the key considerations that we explore in more detail in the whitepaper. You’ve got to blur the organisational and talent boundaries between data, analytics and technology so you can focus on the business problem – these aren’t separate functions, but core capabilities that should be shared across the organisation.

We touched on some insights that we didn’t fully explore in the white paper, such as the challenges of getting the most out of the different members of an AI team. For instance, the need to allow your Data Scientists to focus on creating the best machine learning solution possible, rather than architecting, implementing and maintaining the underlying machine learning infrastructure. Organisations need to embrace the concept of “MLOps” and build out appropriate skills in the team to ensure that hardware doesn’t create a major obstacle to implementing models into production.

If you get the chance to listen to the podcast when it’s out I hope you enjoy and find it useful, and in the meantime you can find more information in the white paper.

discuss this with us & find more insights on

Want to learn more?

let's talk