How Should HR Leaders Use AI Tools?

Warning: Undefined variable $count in /srv/users/circus/apps/ec2022/public/wp-content/themes/electriccircus2022/universal-template/flex-banner.php on line 134

The Innovation Day That Changed Our Thinking

We recently ran an experiment at Electric Circus. The brief: everyone on the team had an afternoon to prototype a solution for a real HR problem using AI tools. No coding experience required, no lengthy development cycles – just identify a problem and see what AI could build.

Within a few hours, our team had created:

A career guidance concierge that conversationally guides employees through internal mobility options
An absence management tool that consolidates scattered employee records and company policies to support difficult conversations
A gamified training retention system using a building rocket ship mechanic
An HR problem solver that recognises issues and surfaces relevant company guidance
An adaptive learning system that identifies learning styles and reformats content accordingly

These weren’t sketches or wireframes – they were working prototypes addressing genuine pain points. One problem we tackled, absence management, costs a client an estimated million pounds annually. Real potential.

But as we refined these prototypes, we started hitting walls.

The Reality Behind the Magic

Leon, who built the training retention prototype, captured it perfectly: “AI kind of knows what it’s talking about, but putting it into practice doesn’t seem to do it.” He’d tried implementing specific tweaks – adjusting the rocket’s appearance, fixing question feedback – and found the AI struggled to execute them consistently.

Jamie, working on the career concierge, had a similar experience. They’d prompted the AI with specific character details (moustache, big nose, the works) to create a friendly career attendant. Whilst it generated something functional, getting it to feel right was another matter. As Jamie put it: “I’d prefer to do the work myself, but I acknowledge the utility of AI for quick results.”

Perhaps most tellingly, we noticed a pattern: initial AI-generated prototypes were often the best versions. Subsequent attempts to refine them frequently made things worse, not better. AI could generate ideas quickly, but the nuanced improvements that turn a prototype into a polished solution? That’s where it struggled.

What Makes a Prototype vs. What Makes a Solution

Let’s take the absence management tool as an example. The prototype was clever – it could pull together hypothetical absence records and present generic guidance. Impressive for an afternoon’s work.

But to make it genuinely valuable, you’d need:

Company-specific content. Generic HR advice isn’t useful. Managers need their actual policies, escalation procedures, and organisational approach. The prototype needs proper content strategy to work in the real world.

Contextual intelligence. A first-time absence conversation needs different support than managing long-term sickness. The prototype lacked L&D expertise to understand how people actually learn and apply information under pressure.

Security and compliance. AI prototypes don’t consider data protection, GDPR compliance, or security protocols. Real HR tools handle sensitive employee data – they need robust security architecture, audit trails, and compliance frameworks.

Design and brand quality. The prototype might function, but does it reflect your organisation’s brand? Does it feel professional enough for HR teams to trust? Design quality, visual consistency, and brand alignment require expertise that goes beyond functional prototyping.

UX and accessibility. Was the information hierarchy right? Could managers find what they needed quickly in a stressful moment? Was it accessible across devices and for users with different needs? These considerations require design thinking that goes beyond getting something to work.

Production polish. Prototype rough edges that are charming in an innovation session become deal-breakers in production. Error states, loading behaviour, mobile responsiveness, edge cases – the thousands of small decisions that make software feel finished.

The Professional Gap

This is where the real insight landed. AI has genuinely democratised idea generation. Anyone on our team – regardless of technical background – could prototype an HR solution in an afternoon. That’s remarkable.

But there’s a substantial gap between that prototype and a solution that delivers real value. Closing that gap requires:

L&D expertise to understand the actual problem properly. Our prototypes improved dramatically when we stopped thinking about what AI could build and started thinking about what learners actually need. The career concierge became more valuable when we considered Gatwick’s specific internal mobility challenges – making career pathways conversational and fun addressed a real barrier we’d discovered through proper user research.

Content strategy that goes beyond generic advice. Ben realised this during the session – the HR problem solver became genuinely powerful when you considered uploading actual company content like codes of conduct, security policies, or onboarding materials. The prototype was just a shell until you filled it with real, contextual information that employees would actually trust and use.

UX and design thinking that makes solutions genuinely usable, not just functional. Jamie’s observation about preferring to do the work themselves highlights this perfectly. AI can generate interfaces quickly, but making them intuitive, accessible, and genuinely pleasant to use? That requires human design thinking and iteration based on how real people interact with systems.

Professional execution and the refinement that AI struggles with. Those tweaks Leon couldn’t get Claude to implement consistently? That’s the difference between a prototype and a product. It’s the thousands of small decisions about error states, loading behaviour, mobile responsiveness, and edge cases that make software feel finished rather than functional.

Why Prototype Anyway?

Given these limitations, you might wonder if AI prototyping is worth the effort. We think it absolutely is, but with clear expectations about what you’re creating.

Prototyping with AI is brilliant for:

Testing concepts quickly before investing heavily. We validated seven different HR solution concepts in an afternoon. Some had legs, others didn’t. That’s incredibly valuable information before committing development resources.
Getting stakeholder buy-in. A working prototype, even rough, is far more powerful than a written brief. When we showed clients our absence management prototype, they immediately understood the value proposition in a way that PowerPoint slides never achieve.
Identifying technical feasibility early. Some ideas that sound great on paper turn out to be technically complex or problematic in practice. Prototyping surfaces these issues before you’ve invested heavily.
Kickstarting creative thinking. AI’s suggestions often sparked better ideas than what we’d initially proposed. The adaptive learning system, for instance, evolved substantially as the AI offered alternatives we hadn’t considered.

But prototyping isn’t the same as building. The career guidance concierge might work beautifully with sample data, but making it work with Gatwick’s actual career pathways, their organisational structure, and their specific employee needs? That requires the expertise we’ve built over 16 years.

The Bigger Picture

This mirrors the choice we explored in our previous article about authoring tools versus custom development. AI prototyping is another tool in the kit – powerful for certain purposes, limited for others.

Where AI prototyping excels: rapid concept validation, stakeholder engagement, identifying opportunities, generating creative starting points.

Where professional development matters: contextual understanding, content strategy, user experience design, production polish, ongoing support and iteration.

The exciting part? Seeing how these approaches complement each other. AI generates the initial concept. Professional teams transform it into something genuinely valuable. The prototype kickstarts the conversation. The expertise delivers the outcome.

What This Means for HR and L&D Leaders

If you’re considering AI solutions for your learning and development challenges, here’s what we’d suggest:

Prototype liberally. AI tools have made it remarkably easy to visualise solutions. Take advantage of that to explore possibilities and get stakeholders excited about potential.
But be realistic about the gap. A working prototype is perhaps 20% of the journey to a production-ready solution. The remaining 80% requires expertise in learning design, content strategy, UX, and professional development.
Look for partners who understand both. The future isn’t choosing between AI tools and human expertise – it’s knowing when to use each. Partners who can prototype quickly AND deliver production-ready solutions give you the best of both worlds.
Focus on the problem, not the technology. Our most successful prototypes came from clearly understanding the HR challenge we were solving. The AI was just the tool to explore solutions. Start with the problem – absence management costing £1m annually, career pathways that people don’t engage with, onboarding that doesn’t stick – and work backwards to solutions.

The democratisation of prototyping through AI is genuinely exciting. But as we discovered in our innovation session, there’s a significant difference between proving a concept works and delivering extraordinary learning experiences that perform.

The first might take an afternoon. The second? That’s where professional expertise still matters.

Closing the Experience Gap

This all comes back to the challenge we see every day: the gap between the slick digital experiences your employees have at home and the clunky systems they’re forced to use at work.

AI prototyping can generate impressive concepts quickly. But closing that experience gap – creating HR technology that feels as intuitive and engaging as the apps your people love – requires more than clever prototypes. It requires learning science, storytelling, gamification, and 16 years of understanding what makes digital experiences genuinely memorable.

Whether it starts with a prototype or a conversation about what your people actually need, we’re here to help you close that gap.

Want to explore what’s possible? Let’s talk about the gap between the HR tech you have and the employee experiences you want to deliver.