Скачать или смотреть AI in Learning & UX Design: Augmenting, Not Replacing

AI is already changing how we work, but too often people assume it’s about full automation. Barclays have been running a major AI across their internal learning teams and exploring how AI tools can genuinely improve efficiency and quality.

This talk by Mike Brown, Head of Design at Barclays will go through the structured testing done, the real measurable gains seen, and the unexpected ways AI has fallen short. Mike will also touch on the risks and challenges of integrating AI in a highly regulated environment like banking, where zero risk is the goal, not just moving faster.

Key take aways:
AI is great for small process improvements, not full automation
Real, measurable savings exist—but only with human oversight. AI can massively reduce hours spent on certain tasks, but nothing we’ve tested can just be handed over and left to run on its own.
People often think too big, too early. The real benefits come from integrating AI into existing processes, rather than trying to revolutionize everything overnight.
The ‘AI will catch up’ mindset is delaying work. There’s a growing trend of teams deprioritizing tasks because they assume AI will automate them in six months—this can be a dangerous gamble.
Regulated industries have unique challenges. In banking, AI adoption requires absolute certainty, which limits its use cases in ways other industries might not face.

00:00 Welcome back to UXDX
01:22 Mike’s background and why learning design at Barclays is different
03:05 Big companies, big change: why AI is the next transformation wave
05:18 The Jakob Nielsen reframing: it is about who uses AI better
07:40 The real unlock: measurable gains in a business context
09:05 How Barclays started testing AI with workshops across the UK
12:05 Building an AI innovation squad and prioritising use cases
13:40 Use case deep dive: test question generation
16:10 Results: time saved and quality impact at scale
17:55 Use case deep dive: summarising content and the counting problem
20:10 The bigger issue: AI can be confidently wrong, so you must verify
23:10 Personalisation and generative UI, and what it will take to get there
24:35 AI agents and what “agentic orchestration” could mean
26:10 Why timelines are unclear: the virtual company experiment
30:38 Closing thoughts and thank you

👉 Upcoming Events: Don't miss out on our vibrant live-online and in-person community events! For the latest schedule and to see if there's a free event near you, visit our events page: https://www.uxdx.com/community/

Connect with UXDX on Social Media:

📚 LinkedIn: Follow us for updates and insights:   / uxdx  

🔔 Stay Informed: Subscribe to our newsletter for the latest curated news on Product, UX, Design, & Development: https://www.uxdx.com/subscribe/.

🌍 Join the UXDX Community: Explore more about our community here: https://www.uxdx.com/