Interview: Inside AI Automation team

Five questions to Mara: personalising the student experience with reliable AI.

February 3, 2026

By Ann Steinbach

What it’s about:

✔️ You’ll meet Mara and her path into our AI Automation team.

✔️ You’ll peek into her team rituals: goals, stand‑ups, fast unblockers.
✔️ You’ll learn how data turns into personal support for students.
✔️ You’ll discover why IUG is a place to grow.

Reading time: ~4 minutes

Our AI Automation team builds reliable, data‑driven solutions with the overarching goal of delivering the best possible, personalised student experience and turning feedback into direct, everyday solutions. Mara Coellar, Senior AI Automation Manager at IU Group, explains how we partner with engineering and service teams, test and compare models, and turn student feedback into everyday improvements.

Hi Mara, could you tell us a bit about yourself – professionally and personally?

Hi 🙂! I work in the AI Automation team, bringing AI automation and data-driven solutions to higher education in close collaboration with other teams. We build solutions for the teams who are in daily contact with our students. The central question is always: how can we measurably improve student experience through more efficient processes, better support and smart automation? To do this, we analyse student enquiries, experience data and support tickets, and pinpoint the topics with the greatest potential impact. Right now, we’re working particularly close with service teams for our distance-learning programmes, where impact is highly measurable. In parallel, we’re developing solutions for Upskilling and our International unit.

My professional background is in Customer Experience (CX) Management. Through various roles (for example, an internship at BMW, as product specialist at Qualtrics and my previous role in the CX team at IUG), I learnt to evaluate feedback systematically, identify patterns and prioritise the levers that drive the biggest effect. And that’s exactly what I now combine with AI.

And personally? Outside of work, I love being outdoors with long walks and bike rides, often along the Isar or in parks. I also go cold-water swimming with friends from time to time; it’s incredibly energising. Part of my family lives in Mexico, so I try to visit about once a year. At home, I enjoy reading and spending time with my husband and our cat.

What does a typical day look like in the AI Automation team?

There’s rarely such a thing as “typical” – and that’s what makes it exciting. We’re a cross‑functional, data-driven team spanning engineering, data and business that has grown quickly from three to nine colleagues, bringing a wide range of perspectives. And the question that drives us day in, day out is how we can enhance the service experience for students and the day-to-day work of service teams through AI and automation.

On Mondays, we set goals, deadlines and any potential blockers; daily stand‑ups keep us closely aligned; and at the end of the week we reflect on progress, learnings and development priorities. We address blockers openly and resolve them quickly. We work in parallel on several projects, often in small groups, and stay in close contact with the service teams to make sure we’re tackling the right topics.

Working with AI means testing, validating and challenging. We benchmark LLMs, check for consistency and track which model is more prone to hallucinations. For analyses or reports, we switch between models depending on the task and compare outputs. Once a prototype becomes a product, clear guardrails and governance apply: strong data grounding, explicit rules and full traceability.

What, in your view, is crucial to ensure new technologies don’t remain hype but deliver real value?

The starting point should be real problems, not the technology itself. We look beyond symptoms in the “outer loop”: where in the process can we intervene to make things sustainably better? AI is a tool, not an end in itself. Clear goals, realistic expectations and close collaboration with future end users drive adoption. Sometimes “not automating” can be the best decision. Ongoing testing, clean data and regular stakeholder dialogue ensure technology saves time, reduces frustration and enables better, measurable decisions.

How do you stay up to date on AI and tech?

Mostly through hands-on experimentation in everyday work. That way I notice straight away when models (LLMs) gain new capabilities or when their quality changes.

“At IUG, we can experiment extensively and trial new developments, and I make use of IU’s AI upskilling offers – like our AI Barcamp and the annual IUG Tech Conference – to challenge ideas, exchange with peers and stay sharp.”

I read specialist articles, follow discussions on LinkedIn and listen to podcasts such as “AI & I”. If something is free to test, I’ll try it in my own time and ask AI tools for recommendations on which model suits a particular goal. In this way, I’m not only building solutions but continually refining my own way of working. I also value learning from people outside IUG who work with AI. It often opens up new angles. My mum is a German teacher, and since I showed her how to use AI, she has been surprising me with use cases I would never have thought of.

Is there a tool or a piece of technology you couldn’t do without in your day-to-day work?

I rely on Claude every day. Integrated with our ticketing and knowledge base, it speeds prioritisation and decisions. Switching models and comparing outputs improves quality.

Thank you, Mara!

From left to right:
1. At IU Tech Conference in Summer 2025.

2. + 3. In her favourite personal pursuits.

Come build with us

Team Tech