At SAP NOW AI last week, two conversations from the main stage stood out for us. The first was SAP's vision for Business AI and how it intends to turn the promise of the Autonomous Enterprise into tangible business value. The second was Naomi Weir from the CBI's assessment of the UK business landscape, and the challenges facing organisations trying to invest, grow and remain competitive.
On the face of it, two quite different topics. But we came away struck by the parallels between them. In both cases, there is a gap between the opportunity being articulated at a strategic level and the more immediate, practical concerns of businesses on the ground. And in both cases, the challenge is increasingly one of execution.
From AI opportunity to business value
At Sapphire in May, SAP set out its vision for the Autonomous Enterprise. Unsurprisingly, Business AI dominated the main stage at SAP NOW AI too. This time, however, SAP spent more time addressing the practical challenges customers face in turning that vision into reality.
Manos Raptopoulos, SAP's President for Customer Success EMEA, explained the challenges holding back enterprise AI adoption, and why SAP is taking the approach it is with Joule. He referenced research from Boston Consulting Group suggesting that 74% of organisations remain stuck in AI pilot mode, unable to scale their initiatives into meaningful business outcomes.
That feels consistent with what we're seeing. AI is already delivering real value at an individual level. Most of us are using tools like ChatGPT, Claude or Copilot in our daily working lives. But translating that individual productivity into measurable, enterprise-wide business value is proving considerably harder.
Manos highlighted some fundamental reasons for this. Large language models are probabilistic, whereas enterprise systems of record require deterministic, reliable outcomes. The cost of AI consumption, measured in tokens, doesn't necessarily align with the value of the business outcomes it delivers. And general-purpose language models lack the semantic understanding of complex enterprise data structures and processes.
These challenges also help explain why SAP has invested so heavily in developing Joule, rather than simply relying on third-party AI technologies. SAP's advantage lies in the depth of its understanding of enterprise processes and data. By grounding AI in SAP's business data models, metadata and knowledge graph, SAP is seeking to deliver outcomes that are more accurate, contextualised and, crucially, governed.
He went on to make the case that enterprise AI needs traceability, observability, human oversight and appropriate controls over what autonomous agents can actually do. Particularly when those agents are operating within systems that manage financial transactions, supply chains and other business-critical processes.
SAP also presented a pragmatic view of the value opportunity. Rather than promising wholesale transformation overnight, the potential to improve an established process such as inventory or asset management by even 1–2% can have a disproportionately positive impact on profitability.
The reality for SAP customers
That gap between opportunity and reality became even clearer during the discussion with the CIOs of HMRC and Centrica.
Both offered useful perspectives on the balance between innovation and the responsibilities of running complex organisations. Daljit Rehal from HMRC was particularly clear that anything which compromises trust is simply not acceptable. AI-driven outcomes need to be explainable, auditable and, where necessary, defensible in a court of law.
There was also an interesting debate about whether AI represents the next evolution of automation or something so fundamentally different that businesses need to rethink their operating models from first principles. HMRC's experience demonstrated that sophisticated automation and algorithmic decision-making are hardly new, while Centrica offered a perspective on the potential to redesign business processes around AI.
But the contrast between SAP's vision and the immediate priorities of its customers became most apparent when we left the main stage.
It was striking that not once in the first 90 minutes of the keynote did anyone mention S/4HANA. Only 18 months ago, it was dominating the agenda. Yet in conversations with customers throughout the day, the questions were still very much about their existing SAP landscapes. What should they do with ECC? How and when should they move to S/4HANA? What can they achieve with the systems they already have?
For many customers, S/4HANA and RISE remain significant milestones on the journey towards the Autonomous Enterprise. SAP is right to articulate where the technology is heading, but there is an execution gap between that vision and the position of much of its installed base.
UK business: resilience before transformation?
Naomi Weir's contribution from the CBI offered an interesting parallel.
She described a UK business environment in which ambitions for growth and investment are being constrained by rising operating costs, taxation, energy prices and workforce pressures. She also highlighted the gap between government ambitions around infrastructure, energy and AI, and the ability to deliver against them.
For many businesses, the more pressing priorities are managing costs, improving productivity and strengthening resilience in an uncertain economic and geopolitical environment. The parallels with SAP's AI conversation were difficult to ignore. There is plenty of ambition for the future, but investment decisions are being driven by the realities of running a business today.
We're seeing this in our own conversations with SAP customers. Rather than committing to large, multi-year transformation programmes, many are prioritising more focused projects that can deliver results in months rather than years.
That might mean reorganising company codes or plant assignments in response to tariffs, restructuring supply chains to deal with geopolitical disruption or making changes to legal entities following acquisitions and divestments. These are practical changes that businesses need to make now, often within their existing SAP landscapes.
And resilience isn't just about responding to today's challenges. It's also about having the flexibility to respond to whatever comes next. For SAP customers, that means understanding how quickly they can make significant changes to their systems and data when business circumstances demand it.
Making progress today
We came away from SAP NOW AI encouraged by SAP's approach to Joule, particularly its recognition of the challenges involved in delivering enterprise AI at scale.
But Naomi Weir's perspective was a useful reminder of the environment in which customers are making their investment decisions. Few businesses have the luxury of pursuing a long-term technology vision without also addressing more immediate operational and financial pressures.
These priorities needn't be in conflict. Improving data quality, simplifying financial structures or making existing SAP systems more flexible can deliver immediate business value while also preparing the ground for AI. After all, autonomous agents will be of limited use if the underlying business data is inconsistent or poorly governed.
SAP has set out a compelling vision for the Autonomous Enterprise. For many of its customers, the important question is how they can move towards that vision given the systems, budgets and immediate priorities they have today.
.png)
