From Data to Deals: Biscred's Innovative Approach to CRE Taxonomy

In commercial real estate business development, the advantage often comes down to one thing: finding the right prospect faster than the competition. That requires more than a large database. It requires structure, accuracy, and a taxonomy that reflects how the industry actually operates.

Under the leadership of Rob Armstrong, President of Biscred (who joined in June 2023), Biscred’s data team has refined a methodology that blends machine learning, artificial intelligence, and human review to build a prospecting platform designed for CRE professionals. The focus is deliberate: high-quality data, thoughtful categorization, and filters that match how owners, operators, developers, and service providers think about their markets.

Think of Biscred’s database less as a list and more as a well-zoned city. Every building has a purpose. Every tenant is categorized correctly. Every parcel is mapped with intention. The data team doesn’t just collect information; they organize it so users can move efficiently from broad searches to highly specific opportunities without sorting through irrelevant leads.

The result is a platform built for precision prospecting — where technology handles scale, human oversight ensures accuracy, and taxonomy does the heavy lifting behind the scenes.

How does Biscred approach database taxonomy?

When Biscred launched in 2022, the platform included 8,600 companies and just over 96,000 contacts. Today, it covers more than 584,000 companies and over 6 million professionals. The underlying philosophy, “start slow so you can move fast,” remains, but the execution has evolved.

According to Rob Armstrong, three lessons shaped that growth.

First, speed to market matters. “We learned that perfect can be the enemy of the good,” Rob says. Biscred now releases new data segments in a high-quality Beta phase, even if market coverage is not yet complete. Clients gain early access, provide feedback, and help refine classifications in real time. That shorter feedback loop has accelerated iteration and improved time to value.

Second, early adoption of AI tooling has been critical. Advances in machine learning and automation have improved how companies and contacts are classified, enriched, and validated. Being first movers with new tools has allowed the data team to expand coverage efficiently while maintaining standards.

Third, human oversight remains essential. CRE businesses rarely fit neatly into a single category. Ownership structures, operating models, and service lines often overlap. “The human element still has a vital role to play,” Rob notes. The team’s ability to respond to bespoke client needs — and adjust taxonomy accordingly — has proven valuable as the database grows.

Refinements have followed client demand. One example: adding more than 47,000 residential single-family home builders after identifying demand within CRE-focused cohorts. Residential coverage now includes appraisers, mortgage lenders, and HOAs, supporting a broader range of prospecting strategies.

And, in early 2026, Biscred expanded data center coverage — over 28,000 data center–related construction and service firms were added — reflecting strong activity in a highly specialized sector.

What sets Biscred apart from other CRE data providers? 

Data quality and taxonomy remain foundational. Biscred continues to invest heavily in validation, classification, and structured organization. But strong data alone is not enough. The focus has shifted toward activation and accessibility, ensuring clients can translate information into targeted sales and marketing action.

“We are fully focused on driving outcomes for our clients,” Rob says. Biscred brings together companies, contacts, and property-level data in a single platform built specifically for CRE prospecting. That alignment allows users to move from identifying a building or asset type to pinpointing the right decision-makers tied to it, without stitching together multiple tools.

Customer feedback has reinforced how varied CRE prospecting needs can be. Segmentation often extends well beyond company size, generic industries, and job titles. The addition of building data revealed how nuanced client strategies truly are. Teams segment by asset specialization, capital stack focus, geographic footprint, and development pipeline — not just standard B2B filters.

At the same time, product evolution presents its own challenge. New datasets and features do not automatically translate into adoption. Biscred continues to invest in a dedicated customer success and support team to ensure clients are using the platform to its full potential.

Listening has also led to product innovation. After observing clients build ideal customer profiles using long lists of job-title keywords, the team developed proprietary, CRE-specific functional area filters, such as asset management, acquisitions/investments/dispositions, and capital markets. These classifications reduce manual filtering and help clients reach relevant decision-makers faster.

The result is a platform built not just to inform, but to convert insight into action.

Data quality

For Biscred, data quality starts with accuracy and contactability. If a professional appears in the database, clients need confidence that the individual works where listed and that the contact information is current.

To maintain that standard, Biscred combines automation with ongoing human review. Machine learning models monitor changes in employment, company structure, and contact data, while the data team validates complex or ambiguous records — common in commercial real estate, where professionals often operate across multiple entities or roles.

Validation is continuous. Companies evolve, titles shift, and email domains change. The platform is built to identify and correct those updates proactively.

Manual review also serves a second purpose: strengthening the system over time. When patterns emerge, that logic is incorporated into future automation. This approach allows Biscred to scale its coverage while maintaining the precision required for effective sales outreach.

Database taxonomy

Biscred’s taxonomy is designed to reflect how commercial real estate professionals actually segment their markets. The platform moves well beyond broad asset class filters.

Users can refine searches across company, individual, industry, asset experience, and geographic levels, along with seniority, job function, and company size. That structure allows teams to move from a high-level market view to highly specific decision-makers tied to particular asset types or specialties.

One meaningful distinction is contact-level asset specialization. Rather than assuming everyone at a firm shares the same focus, Biscred enables users to identify individuals aligned with specific asset classes or business lines. This is particularly important within large organizations that span multiple verticals — industrial, multifamily, retail, data centers, and more — under one corporate umbrella.

Traditional lead lists typically stop at the company name and a roster of contacts. Biscred’s taxonomy goes deeper, mapping people to functions and specializations within the firm. That added layer of classification reduces guesswork and limits the need for manual filtering.

The result is a prospecting database structured to surface relevant contacts more efficiently, particularly in an industry where roles, asset types, and investment strategies often overlap.

How AI Fits into Biscred’s Data Production

AI is now embedded in Biscred’s data production pipelines. Rather than replacing human validation, it strengthens how classifications and tagging rules are developed.

“AI is increasingly central to our data production process and pipelines,” says Rob Armstrong. He points to the development of CRE functional area categories as one example, where AI tooling was used to help develop and refine the business rules behind tagging.

At the same time, he is clear about its limits. “AI in 2026 still has gaps in reliability and accuracy.” For that reason, outputs are reviewed before being implemented at scale.

Biscred’s approach has been to extend AI tools beyond engineering. “Our goal has been to put AI tools in the hands of not just our engineering team but across the organization to expedite innovation and productivity,” Rob says. When a workflow proves especially effective, it is deployed more broadly across the company’s cloud infrastructure for production-grade processing.

The model pairs experimentation with oversight, allowing AI to accelerate scale while maintaining control over data quality.

Related: How Biscred Built the Biggest CRE Prospecting Tool in the U.S.

A Look Ahead: Quick Q&A with Rob Armstrong

What Does 2026 Look Like for Biscred’s Data Team?

Rob Armstrong: 2025 was the year we discovered how useful AI tools can be at upleveling our productivity and ability to meet customer needs in a timely fashion. The year’s efforts include frontier model advancements and improvements in token economics and increased context windows. 

What trends in client use cases or market demands have emerged, and how have you adapted your taxonomy to meet them?

Rob: The expansion into categories like HOA — something we never saw coming — and having flexibility and humility to follow customer demand and allowing the platform’s vision to evolve (previously it was strictly CRE and anything outside this would be thrown out if gathered by accident).

It’s become less about taxonomy and more about great coverage, great recency and freshness of data (this is so critical) and making the data easy to use and generate results from. 

As your user base expands across different verticals — for instance, adding residential builders — how do you ensure your taxonomy remains intuitive for all of them?

Rob: There is significant internal discussion and healthy debate on how to continually expand coverage while maintaining a clean and organized product experience. In the future we will provide a way for customers to customize their UI experience so that data that is not relevant to their use case can be hidden / deprioritized. 

Anything else you want to add?

Rob: The future is very bright. We have been talking about adding building-level coverage for several years and we finally have this data in the product. In less than 90 days of release we have learned an enormous amount of insight from the market about where to take this Beta product next and we are very excited. 

---

The Quantitative Revolution: Forensic Analysis of Data Scaling (2022–2026)

Biscred’s growth has moved through three distinct phases.

1. From late 2022 to mid-2023, growth was linear and researcher-led. The database expanded steadily through manual curation, with output directly tied to human effort. This ensured depth and accuracy but limited scale.

2. In late 2023 and throughout 2024, the company shifted to algorithmic ingestion through AI partnerships. Large batch integrations — most notably a major data jump starting in November 2023 — dramatically increased coverage. Growth followed a step-function pattern, rapidly expanding the top of the funnel and moving the organization from research-led to technology-led sourcing.

3. By 2025 and into 2026, Biscred transitioned again, this time to internally built, continuous AI-driven pipelines. Batch jumps gave way to sustained exponential growth powered by proprietary infrastructure. At the same time, contact density recovered to levels comparable to the manual era, demonstrating that the system had learned to map organizational depth, not just breadth.

The following table summarizes the dramatic acceleration in data acquisition velocity, highlighting the impact of the AI transformation.