Building AI capability: early lessons from inside the industry
Artificial Intelligence (AI) is quickly becoming part of daily work across investment management. Many organisations have moved beyond whether to engage with AI, and are now focused on how to build capability across teams in a way that is useful, safe, and scalable. Use cases are already real, from personal assistants and extracting key clauses from lease documents, to streamlining ESG data collection across portfolios.
The INREV Technology Committee’s recent discussions suggest that differentiation lies in learning approach, guardrails shaping usage, and how well organisations connect training to real workflows.
This article shares early observations from INREV members.
Literacy as the starting point
Most organisations begin with foundational AI literacy: where to access approved tools, how to write effective prompts, and how to validate outputs. This stage is less about mastery and more about building confidence and safe habits, reducing the risk of false confidence, incorrect outputs, or misuse of sensitive data.
Approaches to AI education varies, but common elements include company-wide programmes on safe and responsible use, role-specific training, and tiered learning paths that allow employees to progress at their own pace.
“There’s a central, firm-wide baseline covering safe and responsible use, complemented by more role- and function-specific sessions”, explained one member.
Some are investing in dedicated training for managers and teams building use cases. Tools are often rolled out in waves, starting with early adopters and expanding as value is proven. Learning journeys range from beginner to expert, delivered through self-paced platforms.
“Tailoring learnings to user roles and responsibilities is key. Personalisation is also a real driver for adoption”, noted another member.
What’s working: short formats, real workflows, peer learning
Across organisations, several patterns are emerging.
Short and frequent over long and intensive
A mix of short live sessions, practical demos, recorded materials, and regular refreshers appear the most effective.“Shorter, more frequent touchpoints work better than long, one-off training. Momentum builds fastest when training feels immediately useful rather than abstract”, observed one member.
Grounded in real tasks
The approaches gaining the most traction use internal documents and realistic workflows rather than generic examples. “Sessions increasingly use internal use cases, real documents, and realistic workflows (e.g. emails, papers, meeting prep) which makes the value immediately tangible,” one member observed. Another introduced fortnightly “follow-along” sessions specifically for real estate teams.“We always start with the challenge, not the tool,” highlighted another. “People sometimes come to us excited about a specific solution, but the question should be: what are you trying to achieve?”
Peer-driven
Several organisations are building champion communities, where power users share best practices, build reusable workflows, and support less experienced colleagues. “Peer-to-peer learning has been one of the most effective approaches. People are more willing to try something when they see a colleague in their own team doing it successfully.”Similar roles like “catalysts” or “learning ambassadors” also promote AI literacy and encourage experimentation.
Governance as a continuous thread
Rather than treating responsible use as a standalone compliance module, the most effective approach is to weave it into every session.
“Responsible AI principles are reinforced consistently rather than treated as a one-off topic,” one member noted. This includes clear guidance on data use, output validation, and maintaining human accountability for AI-assisted decisions.
Some organisations have developed internal guidelines aligned with external regulation, including the EU AI Act¹, and require new use cases to pass internal IT security and risk reviews. Others focus on clear rules on approved tools and their use cases, with all tools undergoing data privacy and information security checks before deployment.
That said, building AI capability is not without challenges. Keeping pace with rapidly evolving tools, maintaining momentum and ensuring consistent adoption across seniority levels are recurring issues. However, starting small, staying practical and reinforcing responsible use consistently appear to be the most effective ways forward.
Emerging patterns
While every organisation’s journey is different, some factors are clear: design training for busy teams with short, repeatable formats; ground learning in real workflows; build peer networks; and embed responsible use throughout.
“Training is only one part of the journey. The key is to actively start using AI: experimenting, applying it, and learning on the job. Start with real work, and create space for people to share practical examples.”
This article is based on discussions within the INREV Technology Committee and individual conversations with member organisations. Contributors include a Head of AI & Technology at a large European investment manager, a Head of Operations & Digital Strategy at a large global real estate investor and manager, a Head of Real Estate Digital Strategy at a large European real estate asset manager, and a Head of Research & Intelligence at a large European real estate investor.
Want to learn more about technology in the real estate industry? You can explore a range of practical resources on the INREV Technology page.