Rob Gwyther, UK Managing Director and Country Lead at Apexon, shares insights from over 300 conversations with senior executives on how leaders are really thinking about data, AI, cloud, cost, and transformation today…
Theme 1 – The Unscalable Experiment
“We experimented. It worked. We tried to do more. It didn’t work. We’re resetting.”
This pattern of experimentation with data and AI tools but quickly hitting a barrier to scaling what had worked was common across all organisations. It also wasn’t limited to specific industry types, size or maturity. In fact, the larger the organisation the more limiting some of these data barriers became and the solution exponentially more difficult.
Multiple data platforms, undocumented governance, poor data quality, just not having the organisation in place to use data or generate the return on investment from the business were the most common themes that had caused a ‘reset button’ to be pressed.
What seemed to be the answer?
A common data strategy covering consolidation of platforms, data governance, organisational support and integration with the business. Executives admitted they knew they needed it, but they didn’t know how badly they needed it until they had tried to move without it.
Theme 2 – The AI Tug-of-War
“Tech-first means safety is a bolt-on, but safety-first means we don’t move fast enough”
Organisations were split evenly into two camps.
- Experiment with AI and work out what is safe later
- Define what AI safety means first before experimentation.
Unlike Theme 1 it wasn’t surprising that organisations in regulation-heavy industries or where there was a high impact of the AI getting it wrong (e.g. healthcare, pharma, finance) sat firmly in the safety-first camp; whereas organisations in those in industries in which falling behind the innovation curve (e.g. technology, retail) has financial impact, were far more likely to throw caution to the wind.
Regardless of the approach, the overwhelming feeling from both camps was the same.
Frustration.
The experimentation-first camp expressed frustration that they couldn’t move forward with pace, not having the ‘trust’ guardrails in place. These organisations said they were in a continual loop of great experimentation, but limited impactful work at scale.
The safety-first camp expressed frustration that they couldn’t move forward due to the continual cycle of debate on the definition of ‘AI safe’. All great ideas, but no tangible view as to whether those ideas could have the expected impact.
The answer? Do enough of one to inform the other. The feedback we got from the customers we spoke to was that many organisations had not linked safety with the business return. Many had carved out mini governance organisations that had open ended dates to define safety governance or was decided in isolation away from those experimenting. The solution must be found through closer interaction between internal teams and a fail-faster learn quicker philosophy.
Theme 3 – The Great Uncluttering
“We want to, but it’s all just too expensive.”
Multiple platforms, competing tools and separate support organisations inflating the cost of change was common. Many organisations were focusing on simplification, consolidation and rationalisation. Data Platforms were the most common, particularly with larger organisations who had competing platforms across multiple business units, or multiple hubs of data analysis and reporting occurring across an enterprise.
In addition to data platforms, cloud rationalisation was also a key topic, such as moving to single or specific cloud environments. Customers talked about “cool projects” that took them on tangents, investing in specialised cloud platforms and SaaS products that had lured them away from the main hyperscalers and never delivered. These decisions were now being reversed.
FinOps has become a key agenda item again. Focusing on repeatable cost optimisation while new projects are limited, defining leaner and transparent cost structures for future projects.
Theme 4 – The Productivity Mirage
“We’re not chasing the ‘extra hour’ in a workday, we’re chasing the lost percent on the balance sheet.”
The customers who are achieving returns on AI projects are focusing on efficiency programmes that don’t require substantial levels of human productivity improvement and are banking on future headcount reduction to make the business case successful.
Contract and commercial compliance, finance optimisation, auditing efficiencies and broad access to information are all driving a return on their investment. Customers who are trying to measure the productivity improvements of broad GenAI tooling or assess widespread head count reduction due to a myriad of broad AI capabilities are struggling to generate credible business cases and sell the ROI to senior leadership.
Takeaway. Focus AI on hard cost improvements. 10% improvement on your worst controlled contract, or a 15% improvement in cash management can have an easily bankable saving.
Theme 5 – The Echo Chamber
“Everyone says they’re different, but they all provide the same ideas”
Almost every conversation involved a form of “Covid supplier hangover”. Customers said that during a post-covid surge in projects they onboarded suppliers and resources at scale to fulfil their demand. Many had reduced the volume of resources but were still dealing with a selection of resource driven suppliers offering little in innovation, just more resources.
The common theme that was expressed was that despite the narrative of suppliers being “AI native” and offering new and innovative ways to tackle problems, much of the same ideas were coming back from the market. Customers had despair at playbook driven proposals and obviously AI generated materials that showed little focus or understanding.
Takeaway. You can’t just say you’re different, you have to be. Listening to customers and forming quality approaches, tailored to problems is still the objective.
If any of these themes resonate with your organisation’s journey, feel free to get in touch to continue the conversation: rob.gwyther@apexon.com.
