Alex Christou
2026 Saltire Scholar
AI Project Intern, ProsperoHub
Economics & Business Analysis Technology, University of Strathclyde
When I started this placement in May, I knew I'd be working on AI and automation, but I didn't fully appreciate how much of the job would be about people. Three months in, that's probably the biggest thing I'm taking away.
The brief
I joined ProsperoHub, a HubSpot Elite Solutions Partner recently acquired by the Siloy Group, as an AI Project Intern. The core of my placement was an AI Opportunities and Operational Efficiency initiative: looking at where context gets lost as a customer moves from sales, into delivery, and through ongoing support, and then designing and building AI-enabled processes to close those gaps. ProsperoHub itself became the test case, which meant I wasn't just proposing ideas on paper, I was helping build and govern the systems the business actually runs on.
What I actually built
Early on, the work was diagnostic: talking to people across sales, delivery, and operations to understand where handovers broke down and where information disappeared. That research turned into a formal paper on AI opportunities and a phased schedule of works, which gave the project some real structure and got buy-in from leadership.
From there I worked on:
- A HubSpot demo environment showing how AI could read a live deal record, draft next actions, and trigger legal workflows (NDA and DPA generation) automatically at the right pipeline stages.
- Mapping out where integrations knowledge was falling through the cracks, and proposing a proof of concept to make that knowledge easier to find rather than just writing more documentation nobody reads.
- Digging through account and task data to spot commercial opportunities, like flagging which clients might be a good fit for a product cross-sell based on real usage signals rather than guesswork.
- Helping shape a "process handbook" that could eventually split into clear guides for sales and for operations, so the lifecycle from first contact through to renewal has one consistent source of truth.
What surprised me
I expected the technical side to be the hard part. It wasn't. The harder skill was learning how to bring people along: translating a workshop's worth of scattered notes into something a busy stakeholder could act on in two minutes, or knowing when to flag something to leadership rather than quietly building it myself. Working with people much further into their careers than me, and having them take my recommendations seriously, was genuinely one of the most useful things I'll carry forward.
I also underestimated how much governance matters with AI work. Every output needs a human approval gate before it goes anywhere near a client. That's not a constraint I found frustrating, it's a principle I now think is just good practice, and I try to build it into anything I recommend.
Working virtually
My internship was fully virtual, which shaped the experience in ways I didn't fully anticipate. It meant being deliberate about communication: no corridor conversations to fall back on, so getting the tone and clarity of a message right the first time mattered more than it might in person. It also meant building trust with people I'd never met face to face, which took longer, but happened. By the middle of the placement I was getting pulled into conversations and asked for opinions in a way that felt like proper involvement, not just observation.
Looking ahead
After finishing the internship I am moving on to a master's degree in Data Science and AI, and I can already see how directly this placement will feed into that. I've gone from reading about AI governance and efficiency to building, testing, and defending real systems in front of people whose jobs depend on getting it right.
If I had to sum up the internship in one line: I came in expecting to learn about AI, and I am leaving having learned at least as much about how organisations actually work.