• Data & Decision Making
• Productivity & Efficiency
• Workforce & AI Skills

Written by Neil Gladstone, Data & AI Practice Director, Sopra Steria

Across sectors, organisations are investing heavily in tools, pilots, and new capabilities, all with the aim of unlocking productivity gains and meaningful outcomes from AI. And  yet, for many, that impact remains frustratingly out of reach. 

I’ve seen some that are stuck in cycles of experimentation that never scale. Others that  are moving quickly, but without clear returns. Many are simply unsure what “good” looks  like beyond isolated use cases. 

But a small number of organisations are starting to lead the way. 

Not because they have better technology or because they’ve cracked some perfect  strategy. They’re ahead because they have created the conditions to not just adopt AI,  but to consistently turn it into meaningful outcomes. 

From what I’m seeing across organisations at different stages of their AI maturity, three  conditions consistently separate those who are realising value from those who are still  searching for it. 

1. Leadership that understands enough to make the right calls 

In many of the organisations I work with, AI is still treated as a technical topic – something to be delegated to specialists. 

But the decisions that shape successful AI outcomes aren’t technical, they’re strategic. They show up in: 

• Where investment is prioritised 
• How risk is understood, owned and acted on 
• What gets scaled, and what doesn’t  
• How people are engaged, supported, and brought along  

In my experience, when leaders don’t have a working understanding of AI, these  decisions become harder to make and easier to get wrong. If they move forward without  a clear view of where AI will create value too many use cases are pursued, too few are  prioritised, investment is spread thinly and then impact is diluted.

 C2 – Restricted use 

This doesn’t mean every leader needs deep technical expertise. But they do need  enough literacy to ask better questions, challenge assumptions, and make informed  trade-offs. 

The organisations moving forward are the ones where leaders are: 

• Actively building their understanding of AI in context 
• Focusing effort on a small number of high-value opportunities 
• Engaging with its implications, not just its potential  
• Making explicit trade-offs between speed, risk, and return 
• Taking ownership of decisions, rather than deferring them  

Without that clarity, even well-executed AI initiatives struggle to deliver meaningful  results. 

2. Data foundations that connect effort to outcomes 

There’s a growing urgency to adopt AI. But I’ve found that in many cases, the underlying  data simply isn’t ready. 

AI outputs are only as useful as the data behind them, and depend on data that is: • Accessible  

• Reliable  
• Well-governed  
• Fit for purpose  

Without this, even the most advanced tools will produce inconsistent or low-value  outcomes that don’t end up being adopted due to lack of confidence in the output. 

This is where many organisations hit friction. Data is often fragmented, poorly  structured, or difficult to access. Ownership is unclear. Quality varies. And efforts to fix  it can feel slow compared to the pace of AI innovation. 

But the organisations seeing progress are taking a different approach. 

They’re not trying to solve everything at once. Instead, they are: 

• Prioritising data improvements that directly support high-value use cases • Strengthening governance in ways that enable use, not restrict it 
• Aligning data investment to measurable outcomes, not abstract capability

 C2 – Restricted use 

• Building foundations that support both current needs and future scale  

They treat data as a value driver, not a background dependency. 

Because if AI outputs can’t be trusted or used in real decisions, they don’t create value – no matter how advanced the technology is. 

3. An experimentation culture that learns, measures, and scales what  works 

AI is evolving too quickly for long planning cycles or traditional delivery models to keep  up. 

What stands out to me is that the organisations seeing real impact aren’t waiting for  certainty. They’re creating environments where they can learn quickly, safely, and  continuously, and then act on it.  

This means moving beyond one-off pilots and towards structured experimentation. 

And crucially, these organisations are experimenting with outcomes in mind. Not just  creating activity without the direction – which leads to too many pilots, no clear  measure of success and no path to scaling. On these occasions, learning happens – but  the tangible outcomes aren’t realised.  

The organisations pulling ahead treat experimentation differently. 

Instead, they are: 

• Focusing on solving real problems, not hypothetical use cases. 
• Measuring value early, rather than waiting until scale. 
• Creating safe environments where teams can explore without fear of failure. • Building internal capability, not just relying on external solutions. 
• Stopping what isn’t working, and scaling what is. 

This requires discipline as much as speed. 

And it requires a culture where: 

• Teams are expected to test and challenge ideas. 
• Learning is tied to outcomes, not just insight. 
• Success is defined by impact, not activity. 
• Solutions are actually adopted and used in day-to-day work.

 C2 – Restricted use 

Because AI success doesn’t come from a single breakthrough. It comes from  organisations that can consistently identify what works – and scale it with confidence. 

Turning conditions into impact 

These three conditions are core to whether AI activity translates into real outcomes. Technology will continue to evolve. New tools will emerge. Capabilities will improve. But without: 

• Clear decisions about where value lies. 
• Data that supports real-world outcomes. 
• And a disciplined approach to learning and scaling. 

AI investment rarely delivers what organisations expect. 

The organisations pulling ahead aren’t necessarily moving faster. They’re moving more  deliberately – building the conditions that allow AI to deliver real, sustained value. 

Where to focus next 

For business leaders, the challenge isn’t just deciding whether to invest in AI. It’s  understanding where to focus to ensure that investment leads to impact. 

If you’re thinking about how these conditions show up in your own organisation and  where the gaps might be, we’re exploring each of them in more depth in our AI for  Leaders webinar series: Turning AI ambition into real impact. 

Across three short, focused sessions, we’ll look at: 

• How leadership literacy shapes better AI decisions. 
• What it takes to build data foundations that actually support AI. 
• How organisations are using experimentation to learn faster and scale what works. 

Each session is 30 minutes and designed to offer practical insight you can apply straight  away. 

Explore the series and register your place.

Read More Data & Decision Making

Comments are closed.