CONFIDENTIAL CLIENT
How a US real estate asset manager began forecasting revenue more accurately with AI
Forecast real estate revenue in a volatile US market
To make financial planning faster and more accurate, a US real estate fund manager adopted a BlueMetrics solution that combines generative AI and machine learning. Once dependent on manual, subjective analysis, the company now forecasts revenue with a margin of error under 5%, directly through a conversational interface. In doing so, it cleared operational bottlenecks, sped up strategic decisions, and made artificial intelligence a partner in asset management.
Overview
In a market as dynamic as US real estate, forecasting revenue accurately is essential for making strategic decisions quickly.
That is exactly what led a US real estate fund manager, part of one of the country’s largest business groups, to bring artificial intelligence into its financial planning, with striking results.
BlueMetrics had already delivered several data engineering and analytics projects for this client, building and safeguarding a solid foundation of governance and information quality.
That track record was decisive in making the new solution possible to build quickly, letting the predictive models be trained on structured, reliable data that fit the business context.
Market context
- High volatility and competition in the US real estate market
- The need for more accurate, more reliable financial forecasts
- Reliance on manual, subjective analysis for revenue planning
- Demand for solutions that combine data, automation, and fast decision-making
The problem: manual forecasts with little standardization
Even with business intelligence tools and a large volume of historical data, the company still relied on the individual experience of its managers to estimate each asset’s future revenue. The process was manual, slow, and open to subjectivity.
With dozens of properties spread across different regions and many variables in play, forecasting results in a practical, reliable way was a constant challenge, and a clear drag on operational efficiency.
Main challenges:
- Operational limitations
- Business limitations
- Technological limitations
The solution: generative AI and machine learning for automated forecasting
BlueMetrics built a complete solution that pairs generative AI for interaction with machine learning for forecasting. Through an intuitive conversational interface, managers now work directly with the company’s financial data in plain language, without leaning on complex dashboards or specialized technical support.
Questions like “What will fund X’s revenue be over the next 6 months?”, “What was the average revenue by region over the last few quarters?”, or “What is the projected impact of a 10% vacancy on portfolio Y’s properties?” can now be asked directly, by chat. The AI agent understands the request, interprets the context, and automatically calls the time-series prediction models to return clear, accurate, and actionable answers.
This approach opened up access to data and advanced analysis, letting people in different roles, even without a data science background, make faster, better-grounded decisions. The team gained more autonomy, financial planning became faster and more reliable, and the organization reduced its dependence on manual processes and spreadsheets.
The solution’s architecture was built on scalable AWS technologies, such as Amazon Bedrock and Amazon SageMaker, delivering performance, security, and native integration with the company’s existing systems and data. That made for quick adoption and steady, ongoing use of the tool as part of the manager’s day-to-day strategy.
Key components:
- A natural-language conversational interface for querying data
- A predictive model trained on historical revenue time series
- An AI agent that connects the predictive model to the user interaction
- Integration with the company’s existing database
Technical differentiators:
- Generative AI combined with machine learning for automated forecasting
- Use of scalable AWS technologies (Amazon Bedrock and SageMaker)
- An integrated architecture with high availability and native scalability
- A simplified user experience that needs no technical knowledge
Immediate benefits:
- Revenue forecasts with a margin of error under 5%
- More autonomy for managers when making decisions
- Less reliance on manual, subjective analysis
- Financial planning that is faster, more reliable, and more accessible
- Greater analytical maturity across the organization
Results
With the new solution, the manager now produces forecasts with a margin of error under 5%, removing the subjectivity from its analysis and enabling faster, better-grounded decisions.
Revenue estimates, once based mainly on the intuition and experience of individual managers, are now grounded in statistical models trained on the company’s own historical data, which brought more confidence to the planning process.
The ability to model scenarios directly by chat, with no spreadsheets, manual cross-referencing, or technical teams involved, significantly increased managers’ autonomy and sped up financial planning. Through the conversational interface, direct questions about the real estate portfolio now get instant, in-context answers, opening up access to analytical intelligence.
The impact landed squarely on the organization’s analytical maturity. AI stopped being just a support tool and took on a strategic role in day-to-day management, guiding everything from short-term tactical calls to scenario analysis for setting targets and allocating resources. With this solution, the company gained more predictability, speed, and accuracy to compete in a market as competitive as real estate assets.
Operational efficiency:
- Automated revenue forecasts with a margin of error under 5%
- Less reliance on manual analysis and complex spreadsheets
- Faster financial planning, with immediate answers through the conversational interface
- More team autonomy in making strategic decisions
Technical progress and integration:
- Generative AI and machine learning combined in a single integrated solution
- Use of time-series prediction models that run on demand
- Native integration with the manager’s existing database
- A scalable architecture built on AWS technologies
Technologies used
The solution was built on a range of AWS technologies, including:
AWS services
- SageMaker
- Bedrock
- S3
- DynamoDB
- MemoryDB
Languages, libraries, and frameworks
- Python
- JavaScript
- Node
- React
Conclusion
This case shows how combining generative AI and machine learning can turn data into decisions, with precision and speed. By automating revenue forecasts, the company not only raised its operational efficiency but also took an important step toward intelligent asset management in real estate.
A decisive factor in the solution’s success was the manager’s well-structured database, built with support from BlueMetrics. That data maturity made for a smooth integration between the predictive models and the generative AI interface, delivering a fast, reliable experience from the very first tests.
“Projects like this only scale and create real value when there is a well-tuned data structure behind them. Our data engineering expertise is a differentiator that ensures not just speed of delivery but also technical accountability in building the foundations of any GenAI or machine learning application. Or, as in this case, exactly at the intersection of the two.”
More than a technology solution, the project marks a strategic step forward: by putting artificial intelligence at the center of decision-making, the company strengthened its competitiveness in one of the world’s most dynamic markets, with decisions that are faster, safer, and driven by data.