For the last nine years, Basking has been focused on one central question: how is space actually used?

That question led us to build occupancy analytics for some of the largest workplace portfolios in the world. We learned how to normalize complex data sources — Wi-Fi, badge data, sensors, IT telemetry, and other real-time signals — into utilization insights that workplace and real estate teams can actually trust.

But over time, our customers kept bringing us back to a bigger problem.

They had great visibility into how their offices were being used. But the contractual side of the portfolio — the leases, obligations, options, critical dates, rent schedules, and legal conditions – often lived somewhere else. Sometimes in a lease administration system. Sometimes in spreadsheets. Sometimes still buried in PDFs.

And by the time the occupancy data and the lease data were manually reconciled, the decision had often already been made.

That is the problem we built LeaseOps to solve.

Occupancy data is only half the equation

Real estate decisions are rarely made from one dataset:

• A team deciding whether to renew a lease needs to understand utilization, attendance, demand patterns, cost, lease obligations, notice dates, termination rights, sublease options, and business context.

• A team considering whether to right-size a location needs to know not only how much space is being used, but also what the lease allows, when the next decision window opens, and what financial impact each scenario creates.

• A team preparing for a negotiation needs clean data, fast answers, and a clear workflow that brings the right people into the decision at the right time.

This is why we decided to couple occupancy analytics with lease administration.

Basking started with how space is used. With LeaseOps, we added what the leases actually contain. And with AI abstraction, we are now helping teams turn lease documents into decision-ready portfolio data.

The real bottleneck is not storage. It is clean data.

Many companies already have a place to store lease information.

The harder problem is getting accurate, complete, usable data out of the lease documents in the first place.

A single lease agreement can run 60 pages. A full package with amendments can be much longer. Teams need to extract dates, rent schedules, escalation terms, renewal options, termination rights, obligations, legal conditions, and many other fields.

Traditionally, this is manual work. Someone reads the document, pulls the information into a spreadsheet or lease administration system, checks the work, and repeats the process across hundreds or thousands of documents.

That work is slow. It is expensive. And because humans can only review so much detail at scale, important fields may be missed or entered incorrectly.

In lease administration, a small typo can become a real financial problem.

So our goal with AI abstraction was not simply to “use AI on PDFs.” The goal was to remove the manual data entry burden while keeping expert review exactly where it belongs.

Our approach: PDFs in, decision-ready data out

The workflow is intentionally simple.

First, a lease document is uploaded into Basking LeaseOps. It can be a clean executed agreement, a redlined document, or a full amendment package.

Second, our AI abstraction engine reads the document and extracts the relevant data based on a standardized lease schema. This includes rent, escalations, options, obligations, key dates, legal conditions, and other fields required for portfolio management.

Third, the results are delivered into a human-in-the-loop review interface. Users can see where the data came from, review confidence levels, accept or edit values, and then push the approved information into the portfolio system.

This last step is critical.

We do not believe lease abstraction should be a black box. Real estate teams, lease administrators, and partners need to understand where each value came from and which fields deserve closer attention.

AI should reduce the manual burden. It should not remove accountability.

Why “AI as a judge” matters

One of the most important parts of our system is what we call an AI-as-a-judge approach.

Instead of relying on a single model output, multiple models review the same lease independently. They extract answers field by field and compare results.

When the models agree, confidence is high.

When they disagree, the system flags the field for human review.

This changes the review process. Instead of asking a person to reread the entire lease, we help them focus on the fields where expert attention matters most.

In our testing with abstraction experts on real documents, we are seeing 90–98% accuracy depending on document type, language, and field complexity. We are also seeing approximately 30 minutes of processing time per document, compared with the 3–6 hours often required in traditional manual workflows.

At enterprise scale, that difference matters.

For a portfolio with 500 leases and 1,500 related documents, this can mean thousands of hours returned to the team — and cleaner data available much earlier in the decision process.

Human review is not optional

I want to be very clear on this point: we do not think lease abstraction is the place to be cavalier with AI.

Lease data is critical business data. It affects financial reporting, compliance, renewals, negotiations, exits, and obligations. Our customers and partners expect human review to remain part of the process, and we agree.

The role of AI is to do the heavy lifting: read the documents, extract the fields, identify the source language, compare model outputs, and surface the areas that need attention.

The role of the human reviewer is to confirm the data that matters before it becomes part of the system of record.

That combination is where the real value is.

From abstraction to action

The bigger opportunity is not just faster abstraction.

The bigger opportunity is connecting lease data to decisions.

Once the lease data is structured, teams can use it across the portfolio. They can review critical dates, track obligations, understand upcoming options, prepare reports, create workflows, and ask natural-language questions about a lease or a group of leases.

They can ask:

Should we renew this lease?

Do we have a sublease option?

What is the next exit date?

What are the financial commitments over the next few years?

What should I tell the CFO about this lease?

What should I tell the CFO about this lease?

This is where LeaseOps becomes more than a lease administration tool. It becomes part of the operating system for corporate real estate decisions.

The missing link between utilization and lease obligations

For us, the most exciting part is connecting lease abstraction back to occupancy reality.

A lease may look fine on paper. But if the space is consistently underused, the decision changes.

A building may appear expensive. But if utilization is high and the lease terms are favorable, the answer may be to renew or even expand.

A location may have low attendance today. But the real question is whether the lease gives the company enough flexibility to act on that insight.

This is why lease data and occupancy data should not live in separate worlds.

The best real estate decisions come from both sides of the equation: what the contract says and how the space is actually used.

That is the reason we built LeaseOps on the same platform as our occupancy analytics.

Built for enterprise workflows

Abstraction is only valuable if the data can be used.

That is why LeaseOps includes workflows, approvals, reporting, and integrations. Teams can route decisions to the right stakeholders, create task flows around renewals or approvals, and maintain the audit trail required at enterprise scale.

For example, a company may require one approval path for contracts above $2 million and a different path for contracts above $6 million. These workflows can be configured so that the right people are involved at the right time.

The data extracted from the lease becomes part of the decision process, not just a static record.

LeaseOps also supports reporting and API-based integration, including outputs for financial reporting such as IFRS 16 calculations, commitment reports, lease reports, and event tracking.

The goal is not to create another isolated application. The goal is to help real estate, finance, and workplace teams work from the same trusted data.

Why this matters now

Corporate real estate teams are under pressure to move faster.

Portfolios are changing. Workplace patterns are still evolving. Cost discipline is high. Leadership wants better answers. And many teams are still trying to make major decisions with lease data that is incomplete, outdated, or difficult to access.

AI abstraction can help, but only if it is applied in the right way.

For us, that means:

AI that reads the full document, not just obvious fields.

Confidence scoring that shows where review is needed.

Human-in-the-loop validation for critical data.

Structured outputs that connect to lease administration, reporting, and workflows.

Integration with occupancy analytics, so teams can make better portfolio decisions.

This is the difference between extracting clauses and enabling decisions.

The future of lease administration is decision-ready

The promise of AI in corporate real estate is not that every process becomes fully automated overnight.

The promise is that teams can spend less time on manual entry and more time on the decisions that actually shape the portfolio.

That is what we are building with Basking LeaseOps.

PDFs in. Decision-ready data out.

And from there, better answers to the questions that matter most: renew, sublease, right-size, renegotiate, or exit.

For teams managing complex portfolios, this is where AI abstraction becomes more than a productivity tool. It becomes a foundation for faster, cleaner, and more strategic real estate decisions.