Having just completed a successful multi-site cloud migration for The Compleat Food Group across three sites in the UK, we sat down with Janindu De Silva, COO Europe & UK. We spoke to him about what’s shaping the tech landscape in the region, how he engineered a blueprint for a successful multi-site CloudSuite rollout, Fortude’s recent introduction of agentic AI into Infor deliveries and ongoing support, and his view on what businesses need to prioritize to introduce agentic AI successfully into their enterprise systems.
1. What are the standout trends you are seeing across the EU and UK digital landscape right now?
Something I keep hearing in conversations with customers across the EU and UK this year is how ambition has shifted from ‘let’s add AI’ to ‘let’s fix our foundation so AI actually works.’
So, the standout trend I’m seeing isn’t AI itself, it’s the standardization and simplification work happening underneath it. Organizations are consolidating platforms, retiring redundant systems, and rebuilding a single, governed data layer before they push AI further. It’s a lot less exciting to talk about than agentic AI or copilots, but it’s the difference between a pilot that impresses in a demo and a capability that actually changes how the business runs.
We are seeing this first-hand in our own work. ERP is the digital foundation most businesses build on, and getting that foundation right – clean master data, standardized processes, a stable platform – is what determines whether the AI layered on top delivers real value or just adds another layer of complexity.
2026 feels like the year ‘AI-ready’ stopped meaning ‘we bought a license’ and started meaning ‘our data platform can actually support it.’ The organizations getting ahead are the ones investing in that unglamorous groundwork now.
2. You have been doing a lot of work with customers on multi-site cloud implementations. What does that work look like in practice?
Multi-site cloud migrations are where I spend a lot of my time right now, and the pattern is consistent: get the blueprint right once, then repeat it with discipline.
What I’m hearing from customers, more than anything, is a push for standardization, a global template they can roll out site by site rather than reinventing the implementation each time. That’s really about predictability and repeatability: once the template, the processes, and the governance model are proven on the first site, every subsequent rollout should carry a reduced risk profile and benefit from continuous improvement, rather than becoming its own project. We built exactly that for The Compleat Food Group, a UK chilled prepared foods manufacturer: a standardized global template with the right processes built in from day one, applied first at Palthorpes and Nottingham (with Clitheroe and Barnsley to follow), backed by a rigorous governance structure and a custom Finance Data Bridge to keep legacy and cloud systems in sync during the transition. Customer service levels stayed above 98% throughout.
The part that’s easy to miss is that the template shouldn’t stay static. The real value shows up when each rollout gets more efficient than the last, the team tightening the process, feeding lessons learned back into the template, and shortening the timeline site over site. That’s what turns a standard template from a one-off deliverable into a genuine, compounding capability.
3. What are the key challenges or points of learning you'd point to from that work?
If I’m honest about where the real friction sits in this work, it’s not the cloud platform or the AI – it’s data, and its discipline.
Data migration and data quality are the challenges that show up on every single site, every time. Years of local workarounds, manual adjustments, and tribal knowledge get baked into legacy systems, and none of that surfaces until you try to move it. On the Compleat Food Group program, that’s exactly why we built a custom Finance Data Bridge rather than assuming a clean cutover. Legacy and cloud systems needed to run in sync for a period so the business never lost visibility mid-transition. The lesson: budget real time and real ownership for data cleansing before go-live, not as a task you squeeze in alongside everything else. It matters even more now, because the same fragmented, low-trust data that derails a migration is exactly what derails an AI agent; it just does it faster and at scale.
The second challenge is holding the standard. Every site believes it’s the exception, and there’s usually a genuine reason; a supplier relationship, a regulatory quirk, a process someone’s proud of. Let each one in unmanaged, and the template erodes site by site until you have lost the whole point of building it. What’s worked for us is treating template changes as a formal decision, not a local one, a steering group that can say yes to a genuine local requirement and no to a preference, so the template stays a template rather than becoming ten different implementations wearing the same name.
Neither of those is a technology problem. They are both about whether the organization has the discipline to protect the thing it built, the first time it gets tested.
4. Agentic AI is something everyone is talking about. Where are you actually seeing adoption take hold, and how is Fortude's work contributing to this picture?
As I mentioned before, laying the groundwork with the right data foundation is a key area of interest for most organizations. It’s the part that decides whether agentic AI adoption sticks or stalls after the first pilot.
Fortude’s approach has been to use it on ourselves before we ever put it in front of a customer. We follow a simple “adopt, learn, execute” approach: implement agentic AI into our own delivery process first, make it more efficient and cost-effective on our own work, then refine it into something we hand to customers as a more polished output. We recently launched a delivery model built on exactly that discipline; pre-built industry assets, AI agents, and human-led governance across the whole CloudSuite lifecycle. It’s live now for food & beverage, manufacturing, fashion, distribution and chemical customers.
The part I’d point to as genuinely different is what we are doing with the gaps standard ERP leaves behind. Every implementation eventually hits requirements the base functionality doesn’t cover, and the old answer was a custom build, effective on day one, then increasingly bulky and expensive to maintain, and usually the first thing that breaks on the next upgrade. Instead, we are building purpose-built AI agents for those specific outcomes: lighter to run, easier to evolve, and because we test and refine them across multiple customers rather than building bespoke for one, they get better with every deployment instead of just getting older.
That’s really the shift I’d point to. Agentic AI isn’t just automating what ERPs already do; in the right places, it’s starting to replace the customization layer altogether.
5. Looking ahead, what shifts in innovation or technology do you expect to reshape the region over the next 12 to 18 months?
If I zoom out from everything we have talked about, the next 12 to 18 months is when a lot of this stops being optional.
The clearest marker is regulatory. High-risk AI obligations under the EU AI Act – data governance, human oversight, technical documentation, full audit trails become mandatory from December 2027, and the UK has just created a dedicated AI minister, which tells you governance is tightening there too, not just in Brussels. That’s well within this window. Organizations that have already built human-led governance into how they use AI, which is exactly the discipline we have applied internally with our own adopt, learn, execute approach won’t need to scramble. Everyone still running AI as scattered pilots without that discipline will be doing compliance and capability-building at the same time, against a deadline.
The second shift is less about a policy date and more about sequencing, and it’s the belief I keep coming back to across a broad and complex customer base. You don’t get to clean, AI-ready data by buying a platform. You get there by structuring and standardizing the business processes first, then wrapping that in a standardized solution. The platform is what generates clean, actionable data once the processes underneath it are disciplined, not before. Europe’s push toward tech sovereignty – semiconductors, cloud, AI infrastructure adds urgency to getting that sequence right, because organizations will increasingly need to move platforms and data with confidence.
And the third is maturity in agentic AI itself. The conversation moves from ‘Can an agent do this?’ to ‘Has this agent been proven across enough customers to trust it?’ exactly the bar we have been holding our own agents to.
6. If a UK or EU business leader wants to get serious about their digital journey, what would you tell them to focus on first?
Don’t start with the technology decision. Focus first on the business transformation, the processes, the data, and the standardization underneath them. Get a genuinely honest view of how your business actually operates today, across every site and every team, and be ruthless about which differences are real (a regulatory requirement, a genuine market difference) and which are just habit or history. That’s the hardest, least glamorous part of any digital journey, and it’s the part most organizations try to skip because a new platform or a new AI capability feels like more visible progress. Standardizing the process, and the data that comes out of it, is what everything after depends on.
Wrap those standardized processes in one platform rather than a patchwork of systems and workarounds, and you turn your data from scattered and unreliable into clean and actionable. That’s the only thing AI can really build value on top of.
Only once that transformation is genuinely done would I look to leverage AI, and even then, start small and specific: pick the operational gaps where your process and data are strongest, prove the value with one team or one site, and only expand once you trust the result.
So, if I had to boil it down: business transformation, process, data, standardization first. AI follows all of it, not the other way around.
Request a 1:1 session with Janindu
