From pilots to practice: How organisations can harness AI
Our Think Ahead panel explores how firms can move beyond pilots to build trusted and practical human-centred adoption.
In 30 seconds
AI adoption is being held back less by technology than by trust, governance, and organisational readiness.
Our panels argues businesses should focus on verifiable, accountable AI systems, not just rapid experimentation.
As AI capabilities become commoditised, human skills like judgement, empathy, and critical thinking will become more valuable.
- Spotify
- Apple Podcasts
Listen to the full podcast on Spotify:
Although AI’s ever-advancing capabilities are increasingly influencing working lives, many organisations are struggling to move beyond pilots and proofs of concept. In our latest Think Ahead podcast, Sergei Guriev, Professor of Economics and Dean at London Business School, Keyvan Vakili, Associate Professor of Strategy and Entrepreneurship at London Business School, and Riham Sattie, co-founder and CEO of MeVitae explore this conundrum. They discuss how AI is being used in real organisational contexts, and the actions that businesses need to take now in order to get the best out of AI whilst operating accountably and transparently.
Can we trust AI?
A lack of trust in AI seems to be the key factor holding back many companies from wider experimentation and adoption. How to deal with this? Riham suggests a practical framework containing the factors needed to build trust, including aspects such as transparency, universal access, testing security, accountability, and data integrity. Keyvan adds a critical dimension: verifiability.
Using the example of building an AI recruitment tool, Keyvan cautions against automatically reaching for generative AI tools without fully thinking things through. He draws a sharp distinction between building a system to save time versus one designed to identify the best candidate – two very different goals requiring very different approaches, needing to be verified according to a different set of standards.
"With GenAI in particular, it's so easy to build and jump into these things that there is now a tendency of not taking that extra pause and thinking: why am I building this tool?"
Make or buy?
A further dilemma facing organisations today is whether to develop AI capabilities in-house or partner with specialist providers. This is an ever more pressing question given the ease and low cost of building, though Keyvan is clear that this should not necessarily be the default: “If there is another company that can do this properly, and there is a competitive market for providing these services, then there's almost always a contractual mechanism that works out best for all parties."
Riham adds that the real value of specialist providers lies in proprietary data, deep integration capabilities, and hard-won reputational trust – assets that take years to build. Creating a trusted system on a bespoke basis and dealing with all the complexities of maintenance, accuracy, and dealing with security flaws may take far longer than buying in a solution.
Where does this leave humans?
What should remain distinctly human in an AI-augmented workplace? For Riham, the answer lies in soft skills – authenticity, empathy, and courage - qualities that are hard to quantify but are increasingly important as the pace of technology change and innovation speeds up.
"With GenAI in particular, it's so easy to build and jump into these things that there is now a tendency of not taking that extra pause and thinking: why am I building this tool?"
Keyvan discusses the importance of another soft skill – judgement, the ability to take an AI output and convert it into a meaningful decision within a specific context. As technical capabilities become commoditised, he argues, the premium will shift firmly to those who can contextualise and apply AI outputs wisely. His advice to students worried about the job market is that judgement, organisational knowledge, management skills, transformation and change management are all going to be valuable skills for the future, and that people who invest in these skills will continue to be in demand.
Preparing the next generation
The panel offers reassurance and practical advice to junior professionals wondering how to develop these skills in a world where AI handles much of the analytical groundwork. Critical thinking and having the knowledge to challenge AI will be important. Riham advocates for internships, work experience, and mentorship – surrounding yourself with people who challenge your thinking.
Keyvan points to the fact that in a world where everyone has access to sophisticated AI tools, the real challenge is how to improve on it – and that this ability to take AI output and make it better is the skill that will separate people in the years ahead.
Reasons for optimism
While AI is undoubtedly changing our landscape, the panel concludes that there is room for optimism. Keyvan notes that organisational complexity, politics, and misaligned incentives mean the pace of real-world AI adoption is far slower than headlines suggest. "These are things that span over time. That gives us, and organisations, time to adjust," he said.
Riham agrees: “AI is the buzzword. It doesn't mean it's actually being applied right now in a lot of organisations.” She’s also excited about how roles are evolving and is optimistic about the new roles being created on the back of AI.
The panel's closing advice to leaders is consistent: invest in the complementary human assets that AI cannot replicate, ensure strategic alignment between AI investments and organisational goals, and above all, start with the problem, not the technology. As Keyvan puts it, the most exciting question may not be how AI fits into what we already do, but what entirely new possibilities it opens up that we simply haven't imagined yet.