How to Implement AI in Your Real Estate Organization

The following is an excerpt from a final project that I completed as part of an MIT course on AI and Machine Learning at MIT:

The Proposed initiative is to launch a Project to embed Artificial Intelligence throughout the organization. This project will touch all departments in the organization – Finance, Asset Management and Acquisitions, and also the external 3^rd Party property managers, which are currently outsourced.

The project will be organized into Phases, with the first phase identifying quick wins that will enable the project to gain traction with a successful start that will assist with employee engagement and help with people’s reluctance to accept the AI impact.

Phase 1 – Finance – Property and Fund Accounting

This part of the project will be focused on Cost Differentiation in helping drive down the cost of processing the large volume of transactions that occur on a daily basis.

This phase of the project will utilize Machine Learning and Robotic Process Automation to process the Accounts Payable received daily throughout the organization and also to process Journals and record receipts of funds from Tenants in payment of Rent, CAM charges and other items.

Phase 2 – Asset Management

The second phase will utilize Machine Learning and Natural Language Processing to improve the processes in the Asset Management department, in areas such as Lease Abstraction – the process of entering a new or amended lease onto the ERP system when is has been signed and also in reviewing large reports – either new Contracts or Environmental reports on properties. These reports tend to be large, dense, time consuming and difficult to read – leading to the risk of errors in omission or in entering the information onto the ERP – utilizing NLP to identify the areas that the Asset Manager can focus on can significantly decrease the risk of an issue being missed and reduce the time spend on these tasks – allowing the asset manager to spend more time on managing the assets and dealing with tenants.

Phase 3 – Acquisitions

The acquisitions process is both one of the most time consuming and riskiest areas in real estate – missing an issue with a property or making the wrong investment can cause significant problems with performance of the underlying portfolio.

The project will utilize machine learning in building the underwriting models, taking information from brokers and other data from research – and utilizing the information to enable the team to underwrite more deals than they currently have time to – at present the team is given 1000 deals per annum – of which they quickly discard around 900, then drill down further on the remaining 100 – with the utilization of Artificial Intelligence to underwrite all of the prospective deals, and perhaps more, there is an increased change of finding deals that were previously discarded, or rejected.

Once properties are under contract, there is a significant amount of time spent poring over legal documents – Sale and Purchase Agreements, Property contracts, Environmental reports and others – with the development of Natural Language Processing apps, these documents can be reviewed much quicker, enabling decisions to be taken on whether to proceed with the purchase or renegotiate the price.

Phase 4 – Artificial Intelligence in the Property

The first 3 phases of the project will embed Artificial Intelligence throughout the operating company, this phase, which can run concurrently with the first 3 phases, will deploy technology at the property level – utilizing robots for basic tenant interaction at properties – this will have more impact in multi-family and self-storage properties than in Office or Industrial properties, but when deployed will help increase rents as there will be increased tenant satisfaction – particularly in multi-family as millennials make up a large portion of the customer base, and are known to want to live in buildings that are technologically advanced. In addition, the adoption of AI at the property level will be a cost differentiator, reducing the number of staff required at the property level.

Phase 5 – Data Analysis

The final phase will be to utilize the increase volume of data that will be collected from the properties, acquisitions and other data sources to utilize machine learning to analyze performance and identify paths to increase performance – either by acquiring better and different assets, or by identifying strategies to help increase rents and drive value.

Summary

This entire project will help primarily drive differentiation of the Asset Manager – by being the earliest adopters of Artificial Intelligence and embedding it throughout the organization. A by produce of this will be the company being able to also benefit from cost leadership. The successfully completion of this project will ensure the real estate manager is ahead of the competition and able to benefit from artificial intelligence, with happier tenants (from improved service) and investors (from improved returns).

It’s important to define a criteria for success in any project, but for this.