Three Perspectives on Trustworthy AI: The IONOS AI Model Hub Webinar
Last week I joined Jawad Alaoui from NORMA and Ashwin Nair Anilil from IONOS for a live webinar on The AI Model Hub — a platform that makes it easier for businesses to access, deploy, and manage AI models without building everything from scratch.
Three Perspectives on the Same Problem
What made this panel interesting was the different angles we each brought. Jawad, as founder of NORMA, talked about the data and machine learning architecture — how models are selected, trained, and validated. Ashwin represented IONOS’s infrastructure perspective — the hosting, scaling, and cost management side. And I brought the practitioner’s view: what it actually looks like to integrate AI into real business applications when you’re building software for clients.
My Contribution
My part of the discussion focused on the gap between AI capability and business reality. We’ve now integrated AI into multiple production applications at Exe Squared — document verification in our Ad-MOTO rider management portal, intelligent search in our CRM platform, automated reporting across several client projects. Each of these taught us something about what it takes to move from “AI demo” to “AI in production.”
The biggest lesson? Cost management. AI models are expensive to run, and the costs scale with usage in ways that can surprise you. We’ve learned to be disciplined about which tasks justify AI and which are better handled by traditional code. A well-written SQL query will always be cheaper and more predictable than an LLM call for structured data retrieval. AI shines when the task genuinely requires natural language understanding, pattern recognition, or dealing with unstructured data — and we talked through specific examples of each.
The AI Model Hub Proposition
The AI Model Hub concept addresses a real pain point we’ve seen: businesses wanting to use AI but struggling with the infrastructure complexity. Which model do you choose? How do you handle versioning? What about data privacy — especially for European businesses dealing with GDPR? The platform provides a managed environment where you can deploy models on European infrastructure with clear data residency guarantees.
For our clients, this matters. When we build an application that uses AI for document processing or data extraction, the client needs to know where their data is being processed and who has access. Having IONOS’s European infrastructure as the foundation makes that conversation significantly easier.
What’s Next
AI adoption in business is moving from the “should we use AI?” phase to the “how do we use AI responsibly and cost-effectively?” phase. That’s a much more interesting question, and it’s where practical experience matters more than theoretical knowledge.
The webinar recording should be available through IONOS’s channels. If you have questions about integrating AI into your business applications — or if you want an honest assessment of whether AI is even the right approach for your specific problem — we’re always happy to talk.
