Written by Caroline Christie, Head of Risk.
AI in commercial real estate loan servicing has moved quickly from theory to practice, but its real impact is still being defined. The question is no longer whether the tools exist. It is where they can be used safely, and where judgement must stay firmly human.
At Solutus, we have begun bringing these tools into selected parts of our day-to-day work. The aim is not to replace expertise or decision-making, but to support it: improving efficiency in our most time-intensive processes while holding a strict line on accuracy, governance and accountability. That distinction matters, because commercial real estate loan servicing depends on interpretation as much as information. A system can process data, structure material and speed up routine tasks. It cannot bring the lived experience, commercial context and relationships that real servicing judgement is built on.
From hype to reality
Across the industry, the early enthusiasm around AI has given way to something more grounded. The challenges are no longer theoretical. They sit in data quality, governance, explainability, and the operational risk of bias and hallucination. The numbers tell the story: In 2024, 10% of European respondents to Deloitte's Commercial Real Estate Outlook Survey said AI had a transformative impact on their business. In the 2025 survey, that figure fell to zero. In a sector built on trust, these issues cannot be treated as secondary, or left to mature quietly in the background. They decide whether adoption is responsible and sustainable, or simply fast.
Precision is non-negotiable
Precision is one of our core values, and it sets a clear test. We cannot accept inaccuracy in our work, whether it originates from a person or a system. So the question is never only whether a tool can produce an output quickly. It is whether that output meets the standard required in an environment where servicing decisions carry real financial consequence.
That test draws a sharp line between where AI adds value and where it does not. It can process large volumes of information, support drafting, and structure complex material that would otherwise take significant time to prepare. What it cannot do is understand context the way an experienced professional does, recognise when something feels misaligned, or sense when a transaction needs a more cautious reading than the data alone suggests. In a judgement-led environment, those are not minor gaps. They are fundamental to how servicing decisions are made.
Shaping AI around who we are
This leads to a broader question we have spent real time on internally. As these tools become more embedded in how we work, is there a point at which they begin to shape not just how the work is done, but how we define ourselves?
Solutus has always taken a highly customer-centric approach. We handle complex loan servicing situations, adapt across jurisdictions and counterparties, and treat value-add support as a standard part of service rather than an exception. All of that rests on judgement, experience, and the ability to read context rather than simply process information. So we approach these developments carefully: not as something that will redefine our identity, but as something we must shape around it.
The research supports that instinct. In Imagining and Building Wise Machines, Johnson et al. (2026) describe human wisdom as the ability to draw on real-world experience, collaborate effectively, and weigh multiple future scenarios when making decisions. These are not qualities a system can reproduce on its own, and they are exactly what underpins good work in risk and servicing.
Augmentation, not substitution
Our approach is focused on augmentation rather than substitution. Over the past six months, our AI Working Group has explored practical applications across our most time-intensive workflows, with the aim of reducing friction and improving efficiency. The guiding principle is straightforward: AI can take on the heavy lifting where appropriate, but accountability and judgement must remain firmly human.
In practice, this should help make the strengths already present in the business more visible, more consistent and more valuable to our clients.
This is the balance we are working to maintain.
_________
Sources
Deloitte Commercial Real Estate Outlook Surveys, 2024 and 2025. Based on European respondents. When asked "How would you classify your overall experience with integrating new AI solutions into your organizational workflows?", in 2025, 28% reported challenges with implementation and 27% reported mixed results (compared with 16% and 23% respectively in 2024). 0% reported a transformative impact, compared with 10% in 2024. https://www.deloitte.com/us/en/insights/industry/financial-services/commercial-real-estate-dashboard.html
Johnson, S.G.B., Karimi, A.-H., Bengio, Y., Chater, N., Gerstenberg, T., Larson, K., Levine, S., Mitchell, M., Rahwan, I., Schölkopf, B. and Grossmann, I. (2026) 'Imagining and building wise machines: the centrality of AI metacognition', Trends in Cognitive Sciences. Available at: https://doi.org/10.1016/j.tics.2026.01.002
