How AI Is Expanding Underwriting Capacity Across Commercial Real Estate

Contributed perspective

Editor's note: This is a contributed perspective by Keyway's co-founder and CEO. InterCapital Group and 12ten Capital are Keyway clients. The performance figures come from Keyway-produced customer case studies supplied for this article. They have not been independently verified, and the case studies do not specify a measurement period, sample size or methodology. The examples describe investment and acquisition workflows, rather than measured lending outcomes.

Underwriting sits at the center of many of commercial real estate’s most important decisions. Whether a firm is evaluating an acquisition or a lender is considering a new financing opportunity, teams need to turn large amounts of fragmented information into a clear view of the deal.

That information may be spread across offering memorandums, rent rolls, operating statements, appraisals, leases, borrower materials and internal investment or credit requirements. Before a decision can move forward, it needs to be found, reviewed, reconciled and turned into usable context.

For many CRE teams, that still requires underwriters and analysts to spend significant time searching across documents, reconciling figures, checking assumptions and transferring information between systems and materials. Those steps are necessary, but they take time away from the analysis and judgment where professionals add the most value.

This is where AI can have a much larger impact on CRE underwriting than simply making individual tasks faster. By taking on more of the work surrounding underwriting, AI can enable teams to evaluate more opportunities, accelerate analysis and dedicate more time to the decisions that require professional judgment.

From Document Processing to Underwriting Capacity

Consider what happens when a new deal reaches an underwriting team.

The relevant information may be spread across dozens of documents, and those documents do not always agree. An updated rent roll may contain figures that differ from an earlier offering memorandum, an appraisal may introduce new assumptions, or new information may arrive separately as the deal progresses.

Before an underwriter can properly evaluate the opportunity and its risks, all of that material needs to become usable context.

AI can handle much of that preparatory work: organizing incoming documents, extracting relevant information, reconciling figures across sources and identifying missing or conflicting inputs. The specific output will vary by workflow. An acquisitions team may use that context to evaluate pricing and market positioning, while a lender may use it to prepare the credit materials required to evaluate a financing opportunity.

When discrepancies arise, AI can surface them while maintaining a clear link to the underlying sources. Underwriters can then review the evidence, adjust assumptions and determine how the information should inform the decision. The technology streamlines the work leading up to that point; professional judgment remains fundamental to what happens next.

Turning Capacity Into Results

Real estate firms are already applying this approach across recurring underwriting workflows.

InterCapital Group, a Keyway client and multifamily investment firm, uses its own underwriting criteria, including geography, asset type, amenities and unit mix, to generate customized comp sets and analyze market positioning. Its investment team reviews the resulting comps and determines which are relevant to the deal before incorporating them into the underwriting analysis. A Keyway case study dated April 2, 2025 reports more than 20 hours saved per underwriting and a two-week reduction in property evaluation time for InterCapital.

12ten Capital, another Keyway client and real estate investment manager and operator, applies AI earlier in the acquisition process, combining owner behavior, portfolio patterns and market signals with parameters based on the firm’s investment strategy to identify potential opportunities and validate pricing. The acquisition team then determines which opportunities warrant further diligence and a site visit. A Keyway case study dated April 2026 reports a 30 to 50% reduction in time spent identifying and qualifying opportunities, 2-5 hours saved per property, and two to three times as many high-conviction opportunities identified per market for 12ten.

In the same Keyway case study, Luis Felipe San Martin, Managing Partner at 12ten, says the approach has enabled the team to identify the right opportunities faster, better understand ownership and validate deals before visiting a property, allowing the team to focus on the opportunities that matter most.

The workflows differ, but the underlying opportunity is similar: reducing the manual work required to turn fragmented information into something an underwriter can evaluate.

For CRE lenders, that opportunity is particularly significant. A financing decision can require information from borrower materials, rent rolls, operating statements, appraisals, leases, loan documents and internal credit standards. AI-enabled workflows can help organize and reconcile that information, surface discrepancies and prepare a first version of a loan narrative for review.

The underwriter can then validate the underlying information, adjust assumptions and determine how the findings should inform the credit decision. By reducing the work surrounding that process, lenders can evaluate more deals, move viable loans through underwriting faster and dedicate more time to risk analysis and credit judgment.

Building AI Around the Underwriting Process

Underwriting decisions depend on judgment, making traceability and human review essential to AI-enabled workflows. Extracted information should remain connected to its source, discrepancies should be surfaced for review, and professionals should be able to validate assumptions and intervene wherever their expertise is required.

This requires more than a model that can read a document or generate a summary. Commercial real estate is a document-heavy industry, with critical information distributed across different files, systems and stages of a transaction. Effective AI needs to understand how that information relates, accommodate firm-specific processes and preserve the controls that govern important decisions.

This is where vertical AI and industry specialization become particularly important. The value of an AI partner increasingly depends on its understanding of the industry it serves: its terminology, document structures, workflows and decision-making context. In commercial real estate, that domain knowledge can turn AI capabilities into systems that operate within the way teams actually work.

For CRE leaders considering where to begin, the most useful starting point is the work consuming their teams’ time today. Where are professionals repeatedly searching for information, reconciling figures or rebuilding the same materials? Which internal standards and previous decisions could provide context for the next transaction? And where does professional judgment need to enter the workflow?

Lending provides a particularly compelling example of what this can unlock. Banks and lending firms that reduce the manual work surrounding underwriting can evaluate more opportunities, respond to borrowers faster and move viable loans through the credit process more efficiently.

Greater underwriting capacity can ultimately translate into greater lending potential, allowing institutions to grow and compete at a different scale while keeping human judgment at the center of every credit decision.

Source materials

• Keyway InterCapital Group case study, April 2, 2025

• Keyway 12ten Capital case study, April 2026

About the author

Eglae Recchia is Co-Founder and CEO of Keyway, a company developing AI software for commercial real estate.