Innovation needs a “portfolio mindset”
The room explored why AI initiatives stall after pilots: there’s a gulf between “low‑cost experimentation” and the true cost of productionising. Leaders often demand ROI too early, and teams struggle to justify investments when outcomes are uncertain - even though that uncertainty is inherent to innovation.
A compelling reframing was discussed: treat innovation as a sunk cost or strategic debt - like financial leverage - where the problem is not debt itself, but uncontrolled, ungoverned, untracked debt.
There was also discussion of structured experimentation models (e.g., small teams testing ideas rapidly and only moving proven outcomes into production pipelines), helping organisations sustain learning without betting everything on a single “big win”.
Executive takeaway: build a visible innovation portfolio with clear stages, decision gates, and transparent costs - rather than forcing every initiative to justify itself as guaranteed ROI from day one.
Power skills are now “data skills”: storytelling, collaboration, prioritisation
A consistent conclusion: the skills most organisations under‑invest in are the ones that make data and AI usable at scale -
storytelling, communication, collaboration, and prioritisation.
Participants discussed practical methods that work because they reflect how humans actually learn:
- Gamified learning and problem‑solving (including “escape room”‑style exercises)
- Presentation constraints like PechaKucha to force clarity and narrative discipline
- Cross‑functional sessions that bring “technical + business + storytellers” into the same room, rather than persona‑siloed training
Executive takeaway: if enablement is only technical, it will always lag the pace of change; invest in power skills that compound across tools, platforms, and waves of AI capability.
The workforce impact is real - and hiring signals must change
The discussion addressed concerns about AI reshaping career pathways, especially for junior roles, and what that means for future capability-building. There was frank debate about the consequences of automation removing early-career “learning through doing” work - and the risk of hollowing out pipelines if organisations don’t redesign development routes.
Recruitment practices were also questioned: CVs and applications are increasingly AI‑optimised on both sides, reducing their value as a signal. The group highlighted a growing premium on human interaction, judgement, and demonstrated thinking, suggesting interview processes and assessments need to evolve accordingly.
Executive takeaway: rethink hiring and progression around evidence of thinking (context, judgement, communication), not just polished artefacts.
Closing reflection: AI may make organisations more human - if leaders choose to lead
If there was one unifying message, it was this: the organisations that succeed with AI won’t be those who deploy the most tools fastest - they’ll be those who deliberately shape culture, decision quality, and capability while technology evolves around them.
Next Steps
These conversations reinforced why leaders value space away from the hype - to think clearly, challenge assumptions, and learn from one another.
If you’d like to explore these questions with peers who are grappling with the same realities -innovation trade‑offs, workforce shifts, and leadership in the AI era - we’d love you to join our
leadership community so that you can be among the first to know of future events.
If these themes resonate, the question isn’t whether to adopt AI - it’s how deliberately you choose to lead through it.
The most resilient organisations are
building portfolios, power skills, and human‑centred decision‑making alongside technology - not after it.
If you’d like to continue this conversation with our
Director of Consulting or explore what this looks like in practice,
let’s talk. You can also take a look at our
value proposition to learn more about Eden Smith and how we work.
