Retail Leasing Using AI

Perhaps the most interesting application of AI in Indian real estate lies in the field of Retail Leasing, traditionally a very difficult, high-risk effort, with many layers of data which need to be analyzed for retail chains to get the sizing and the location right on their first attempt. Of course, there is no perfect method to predict store sales in India, and many stores have to close down if the unit economics don't meet expectations. But the advent of AI can make a large difference in the ability of retailers to open the right size of store in the right locations.

Logically, an AI can generate a list of possible tenants for any given location, provided it has enough information to act upon. With the digitization of the Indian economy, it is not impossible to conceive that an AI will generate such a list in the very near future. The most important part of this exercise would consist of gathering enough data for the AI to reach a meaningful decision, and hence reduce the effort required to find tenants, or vice versa for brands to find great locations.

Focussing for the time being on the information required to make such a decision possible, I am listing what I would consider as necessary information for a successful AI model.

1. Location Details: The specific address or area where the commercial space is located, including the city or region.
2. Local Market Information: Information about the local real estate market in the area, including current demand for commercial/retail spaces, rental trends, and recent lease agreements.
3. Surrounding Businesses: Knowledge of nearby businesses and their types (e.g., retail stores, restaurants, offices) can help identify potential tenants that complement or avoid direct competition. This can also serve as a negative list for brands that already have operational spaces in the vicinity.
4. Demographics: Information about the demographics of the area, including population density, income levels, and consumer preferences, can help target suitable businesses.
5. Transportation and Accessibility: Details about transportation options, nearby public transport stations, parking facilities, and overall accessibility to the location.
6. Local Regulations: Any specific zoning laws, permits, or restrictions that may impact the types of businesses allowed in the area.
7. Property Features: Additional details about the property, such as amenities, infrastructure, and any unique selling points, as also ceiling height, frontage, parking availability, number of floors, etc.

The purpose of running such an AI-based decision-making model would be to increase the productivity of property consultants on the one hand, who currently have a one-in-ten hit ratio, and reduce the vacancy cost for landlords, who have to currently rely on an army of consultants who individually reach out to the clients in their circle of influence. If the AI is able to predict the eventual success of the retail operation and its breakeven point is reached faster, the number of mistakes made by expanding retailers can be greatly reduced.

Such an AI is unlikely to replace human consultants, but those consultants who use such an AI are likely to make much more money as they are likely to get a high repeat clientele.

Next, imagine that you could figure out which brands would lease the space, before the start of construction. This would then allow you to design each commercial shop as per the highest paying brands, based on the location details. Imagine, how high rentals could go if AI were to be involved.

The AI real estate party has just begun.