What operations research taught me: building AI systems that actually work

Senura Hatharasinghe is an MSc Business Analysis and Consulting graduate – here, he discusses what the MSc taught him when setting up his newly-formed business Ashtara AI Ltd. 

When I left Sri Lanka to study at Strathclyde, I didn’t fully understand what I was signing up for. I knew I wanted to understand how businesses worked at a deeper level and I knew I wanted to build things. What I didn’t expect was that a postgraduate programme in business analysis would give me the intellectual framework I’d use to start up a company. 

That framework is operations research. And right now, it might be the most valuable skill in the AI industry, even though almost nobody is talking about it that way. 

The AI implementation problem nobody is talking about 

Businesses globally are rushing to implement AI. They buy tools, run pilots, hire consultants, and then more often than not, projects quietly get shelved six months later. They either say that the technology didn’t work, or the timing wasn’t right. 

However, having spent time working with professional services firms on exactly this problem, I’d argue the technology is almost never the issue. The failure happens much earlier, in the moments before a single line of code is written or a single tool is deployed. 

The failure happens because nobody properly defined the problem. 

What Strathclyde taught me 

Across the MSc in Business Analysis and Consulting, from Strategy Modelling and Risk Analysis through to Spreadsheet Modelling, Business Analytics, and Managing Business Operations, I kept encountering the same underlying discipline which is to undertake “the rigorous, structured analysis of how systems actually work before you attempt to change them”. 

The module that built the foundation was 'Foundations of Operational Research and Business Analysis'. It introduced the discipline of structured problem framing, the kind of rigorous thinking that refuses to jump to solutions before the system is properly understood. Constraint modelling, quantitative reasoning, process analysis, problem structuring methods which were tools that sound academic until you realise they are exactly what separates AI implementations that work from ones that don’t. 

But where it became real was in the module 'Becoming an Effective Business Analyst'. Rather than theory, this module put us in front of actual client problems where we were given real consultancy projects from different organisations and sectors, requiring us to draw on everything else we’d learned and apply it under genuine commercial pressure. That combination of OR thinking and applied consultancy practice turned out to be exactly what building AI systems requires. 

Why OR is the missing layer in AI 

With my start up, Ashtara AI Ltd in Aberdeen, I set out to build AI systems for professional services firms such as law firms, consultancies and growing SMEs. What I am finding is that the organisations struggling most with AI aren’t lacking technology, they are lacking problem structure. 

They don’t know which processes are genuinely inefficient versus which ones just feel that way. They can’t articulate the constraints that make certain automations impossible in practice. They have no model of how work actually flows through their organisation versus how they assume it does. 

These are OR problems; capacity modelling, process mapping, constraint analysis, demand forecasting were all part of the toolkit Strathclyde gave me. Applied before any AI implementation begins, they transform the odds of success entirely. 

That’s the thesis behind Ashtara: fix the process first, then automate it. It sounds obvious. In practice, almost nobody does it. 

What this means if you’re studying at Strathclyde right now 

The AI industry is producing an enormous number of technically capable people. Developers who can build models, engineers who can deploy infrastructure, analysts who can interpret outputs. What it is not producing in sufficient numbers is people who can walk into a complex organisation, understand how it actually operates, identify where the real friction is, and design an intervention that will hold up in the real world. 

That is what an MSc in Business Analysis and Consulting trains you to do: that is what operations research thinking when properly applied, makes possible. 

If you’re sitting in a lecture right now and wondering when you’ll ever use this, the answer is sooner than you think, in contexts that didn’t exist when the module was written. 

The organisations that will get the most from AI in the next decade won’t be the ones with the best technology. They’ll be the ones with the clearest thinking. Strathclyde teaches you how to think clearly about hard problems: that turned out to be exactly what I needed. 

About Senura Hatharasinghe

Senura Hatharasinghe is the founder of Ashtara AI Ltd, an operational intelligence consultancy based in Aberdeen that applies operations research methodology to AI implementation for professional services firms. He completed the MSc in Business Analysis and Consulting at Strathclyde in 2025 on an academic excellence scholarship, graduating with distinction.