The human impact of AI; quicker, better or happier?

The human impact of AI; quicker, better or happier? 

Three leaders from the BBC, National Grid and Hiscox told me how they're building data and AI capability at scale. None believed the tech is the hardest part. 

By Tiffany Manteaw – Data and AI Solutions Director at Corndel 

Last month, I had the pleasure of hosting a panel discussion on The Human + AI Era, with expert stakeholders from several Corndel clients, at Big Data LDN.  

Near the end of our conversation, Louise Blythe, AI Adoption Lead at the BBC, asked a question about upskilling which has been on my mind since:  

"How do we continue to hold the person, when we're trying to move so fast?" 

Though we were coming together at a data conference, where much of the conversation was about platforms, pipelines and tooling, almost everything my panel said about building data and AI capability came back to people. 

Alongside Louise, I was joined by Sarah Blundell, Head of Data Capability at National Grid, and Jason Bell, Global Head of Learning and Development at Hiscox.  

I opened the floor to their views on the gap between the tools organisations have in place, and the skills and confidence to use them well. It’s both a tech challenge and a people challenge. Here’s what I took away: 

Usage is running ahead of capability 

Sarah measures two things at National Grid: 

1) Are we increasing capability?  

2) Are we improving AI adoption and application?  

This addresses a known gap. In Deloitte's 2026 GenAI Workforce Survey of 25,000 UK workers, 63% said they knowingly use generative AI for work. Around half of users said they'd had no formal training.

People are already using these tools. The question for leaders is whether they're also building the confidence and judgement to use them well. 

Be honest that the fear exists 

Sarah segments her workforce into “curious, comfortable and confident”. The confident group is running ahead, and leaders don't want to dampen that. But she estimates 75% of the workforce is still at the curious stage, and some "might actually be terrified". 

She encourages honesty and openness around this fear, which exists in most organisations. Making room for questions and concerns is part of building a constructive, healthy culture of AI adoption. 

Louise added something I think many leaders miss. Adopting AI means you're working with "experts in their field who, for the first time, aren't experts". The tech and the environment are changing around them. That can hit people hard. Some get defensive, others dive straight in. Both need looking after. 

For me, this theme of compassion shone through. Organisations need to build a culture that addresses fear through transparency, because the experience of the workforce is the foundation everything else rests on. 

One route won't fit everyone 

It’s important to get the right fit of individuals to programmes, when it comes to apprenticeships. This ensures that positive word of mouth expands that learning mindset across the business. 

Jason's approach at Hiscox shows what this looks like at scale. Around 3,500 people took an enterprise-wide data and AI fluency and confidence assessment. Each level has a persona and a tailored learning pathway. This allows learning initiatives to, in Jasons’s words “understand where people are, and meet them there.” 

This is why the blend of different training models available today is important to understand and navigate effectively. Apprenticeships, the new apprenticeship units and shorter-form training each do a different job. Targeting the right intervention at the right population is how you mobilise a whole workforce rather than an isolated few. This is where Corndel supports the businesses we partner with, to co-design solutions that fit their unique workforce. 

Structure and coaching make it stick 

Jason emphasised that learning should not be thought of as a one-time event; it sticks through application, coaching, feedback and, often through making mistakes. 

The structure and support built into a good apprenticeship has benefits beyond the participant themselves. It can show organisations how to design their own in-house upskilling. There's something to learn there for everybody. 

At Corndel, we take measuring that impact seriously. Our data demonstrates the knock-on effect of learning; people on our programmes typically impact 14 others in their organisation by sharing knowledge.  

Sarah saw the same ripple effect, but culturally. While apprenticeships can often be misunderstood and therefore underutilised, once one person in a team does an apprenticeship, it becomes a normal, accepted route for that team and even the whole organisation. 

But what about providing models for that applied learning that work for everyone? National Grid also operates in the US, where there's no Levy to fund apprenticeships. So Sarah’s team built its own version, with a learning platform, in-house coaches and buddies. "We basically just stole the model from apprenticeships," she said. 

Build it with your experts, and measure it in plain words 

The theme of partnership arose again and again throughout our panel. From Jason’s perspective, it’s clear-cut that if L&D and HR alone create the data and AI learning, "it’s not going to land". At Hiscox, data scientists and AI engineers help create and curate programmes, giving the learning more relevance, validity and gravitas. 

When it comes to measurement, Louise and the BBC use a simply question that I believe can enhance every leader’s mindset around AI. They ask whether AI is making work "quicker, better, or happier". 

As Louise put it: "Any one of those three is pretty good. Two is brilliant." Those words are so useful because people can relate to them, at an individual, team or whole-workflow level. 

What comes next 

As we heard in the room, it's hard to keep up with what might happen next. But on human skills, the evidence is pointing one way. In IBM's 2026 global study of HR leaders and employees, 57% of chief HR officers and 49% of employees said critical thinking and problem framing are among the skills that matter most in the AI era. 

That matches what I see in action, and Corndel’s stance: human skills will become more important. Critical thinking, judgement, and the ability to see a process end to end and know how to improve it matter more in the age of data and AI. 

So be transparent about the fear. Offer a blend of structured learning, not a single route. And as you move fast to make work better and quicker, don’t forget about making individuals and teams happier, too. They’re still the most important part.