← Back to blog
Artificial Intelligence

“Your Data Is Your IP”: Takeaways from the Commercial Observer AI & Innovation Forum

Raj Singh
Raj Singh
Co-Founder & CEO · November 20, 2026 · 5 min read
Your Data Is Your IP: takeaways from the Commercial Observer AI & Innovation Forum, featuring Raj Singh, Lily Bowerman, Clayton Schuller, Ryan Botwinick and Abhinav Somani

Key takeaways

  • Foundation models and the tools built on them are available to every firm on broadly similar terms. A firm's own record of screening, underwriting, acquiring and operating real estate is what competitors cannot license or replicate.
  • Most firms are not yet able to put that data to work, because their institutional knowledge sits in email threads, offline spreadsheets, data rooms and the memories of experienced professionals.
  • AI has made software prototypes cheaper to build, but the prototype is a small part of the effort. Integrating, maintaining, securing and governing the data underneath it is continuous work that falls outside an investment team's focus.
  • As firms plan for 2027, we encourage them to ask whether their transaction and operating history is structured and accessible, maintained as the portfolio changes, and governed well enough to be entrusted to an agent.

On Wednesday I joined a panel at Commercial Observer's AI & Innovation Forum: CRE, alongside C-suite and senior executives from across ownership, investment, operations, technology and venture capital. We were asked to address two practical questions: which applications of AI delivered a measurable return for real estate firms in 2026, and where those firms intend to direct their attention and capital in 2027.

There is no shortage of AI announcements in our industry, and most real estate professionals are (justifiably) skeptical of them. A forum organized around return on investment rather than product demonstrations suggests that the industry's conversation about AI is maturing, and I was glad to take part in it. What follows is the position I put forward on the panel, along with the point on which I found the most agreement among my fellow panelists.

Your data is your IP

An idea raised in one of the forum's discussions, and one I strongly agree with, is that a CRE investment firm's data is its intellectual property. The foundation models, and most of the tools built on top of them, are available to every firm in the market on broadly similar terms. What a competitor cannot license or replicate is the record a firm has accumulated over years of screening, underwriting, acquiring and operating real estate: the opportunities it reviewed and its reasons for passing, the bids it won and lost, the assumptions it made in underwriting and how they compared with actual performance, the operating history of the assets it owns, and the relationships behind each of those transactions.

Some have argued that the advantage in this cycle will go to the firms that select the best model or adopt new tools first. We don't believe this. Access to capable models is quickly becoming standard across the industry, and a model working only from public information will produce the same analysis for a firm as it does for that firm's competitors. We believe the firms that outperform will be those that apply AI to their own proprietary data to make better investment decisions, generate alpha and, ultimately, deliver better returns to their investors.

In practice, most firms are not yet in a position to do this. Their institutional knowledge typically sits in email threads, offline spreadsheets, data rooms and the memories of experienced professionals, where it remains fragmented, unstructured and difficult to retrieve. Information that cannot be queried consistently cannot support an investment decision with confidence, regardless of how capable the model reading it may be.

Investing in real estate, not building software

Addressing that problem leads to the question of whether to build or buy, which I raised during the panel. Real estate firms are in the business of investing in real estate, not in the business of building software. At the same time, I understand why building is appealing. AI has substantially reduced the cost of developing software, and a capable internal team can now produce a working prototype in a fraction of the time it once required. For a firm that holds valuable proprietary data, building its own tools on top of that data can appear to be the logical next step.

We believe the prototype represents a small part of the total effort. Most of the work comes afterward: integrating with the many systems and documents in which a firm's data resides, keeping that data accurate as deals progress and assets change, securing it, governing who can access and use it, and maintaining all of it as personnel, models and requirements evolve. That work is continuous, it falls outside what an investment team is hired to do, and it competes for time and attention with the decisions that drive returns.

How we help firms build faster

For these reasons, we work with our clients to accelerate the development of the tools they want to build. Together, we build end-to-end agentic workflows that run from the initial screen of an opportunity through underwriting, investment committee and ongoing asset management, each designed around the way that client invests.

Each of those workflows rests on the same foundation: decision-ready data that is structured, maintained and governed in the Altrio platform. The usefulness of an agent is limited by the information it can access, and a firm can only rely on the agent's output if that information is consistent, up to date and appropriately controlled. Establishing that foundation first allows a firm to build on it with confidence, whether the workflow is one we build together today or one the firm designs for itself next year.

We believe this approach allows our clients to retain ownership of what distinguishes them from their competitors, namely their data and their investment judgment, while devoting their time to investing rather than to maintaining software.

Looking ahead to 2027

As firms decide where to direct attention and capital in 2027, I would encourage them to begin with their data, and specifically with three questions: whether the firm's own transaction and operating history is structured and accessible, whether that history is maintained as the portfolio and pipeline change, and whether it is governed well enough to be entrusted to an agent. Firms that can answer yes to all three are likely to get more value from each new AI capability they adopt, because those capabilities will be working from information their competitors do not have.

I am grateful to Commercial Observer for hosting the forum and to my fellow panelists for a candid and thoughtful discussion. We see a future in which the most successful investment firms are those that treat their proprietary data as a core asset and manage it with the same discipline they apply to their portfolios, and we look forward to helping more firms reach that point.