NTrust Infotech chief executive and president Sri Ramachandran dispels some of the myths surrounding AI and explains how it can benefit real estate. Richard Betts reports.

Artificial intelligence – AI – is already affecting real estate significantly. Not only is it being utilised by proptech suppliers to help property professionals to do their jobs more effectively, it is already widespread and in growing use by investors and occupiers.

Real Asset Insight’s Richard Betts spoke to NTrust Infotech chief executive and president Sri Ramachandran, who explains the potential of AI and dispels some of the myths around it.

First of all, he distinguishes between AI and generative AI. “AI has been in existence in different forms for over 10 years and it has been used in different forms. So you have different flavours of
AI, like national language processing which is used for translations and text studies. The typical AI solution works on a smaller data set and it is used in terms of predictive analysis and the decision-making process.”

This includes the purchase recommendations generated by online shopping sites based on evidence of a person’s preferences. “Generative AI takes this to the next level and is able to make new content, not just to answer questions based on existing content,” he explains. “Generative AI uses a larger data model and can create content from data sets. For example, generative AI could be used to write poems after you teach it English.”

Dispelling myths about AI

Ramachandran says there are several myths about AI to dispel. “One is that people think an AI solution is 100% right. That is not true. AI can be wrong because it is dependent upon how we have trained it, what it has learned and from there it is going to be giving you responses. If it is not trained right, if it is given the wrong set of data, that is what is going to come back in the response.”

The second myth is that AI will take people’s jobs away. “That is not true. AI is definitely going to help people. Generative AI will probably make you 30-40% more efficient in what you’re doing, it is going to give you the option to work on something else but it is not necessarily going to take away the entire job from humans. The human in the loop is going to be critical, but is going to be assisted by AI.”

The third myth is that AI is too expensive and only for big business. “AI can be used in small business and it actually gives small businesses a lot more value,” says Ramachandran.

Managing data is critical

There are, however, challenges, notably in the data with which it works. “You need to have a bigger data set for AI to be effective. It depends on what data you are feeding into the system for it to give you the right responses and managing data quality is going to be critical.”

There are potential cost considerations too. “AI needs a lot of machine power to process so it can become expensive. You need to be a little bit more prudent to see how well you’re going to be managing the infrastructure that is used to build it compared to the benefits it will give you.”

AI also throws up ethical considerations, making transparency and accountability critical. “Companies have to disclose what is AI generated – only then will people be able to trust AI. And they are going to make decisions based on AI so there needs to be accountability in terms of the AI solutions that are being used. If it is going to be driving business, that is going to be very important.”

Privacy and personal information

Another consideration is privacy and the amount of personal information that is going to be shared. “For AI to work you need a lot of data. When you’re trying to put in a lot of data you need to make sure that none of the personal information is being exposed. People have to be very ethical and say what kind of data is being fed. Once it is fed it is going to be there forever, so you need to be very cautious.”

‘AI will probably make you 30-40% more efficient in what you’re doing, it is going to give you the option to work on something else but it is not necessarily going to take away the entire job from humans.’

Sri Ramachandran, NTrust Infotech.

There are different emphases and interpretations of AI in the different geographical regions of the world. “When it comes to North America, the focus is more on innovation.” In Asia AI is being embraced by governments keen to use it to solve problems, so there is a drive to reduce red tape to support implementation.

But in Europe, the situation is different, notes Ramachandran. “What we see in Europe is very interesting. Here the focus is more on the ethical factors of AI and how it can affect humans. They want to make sure that they draw the boundaries to say how well AI has to be used in terms of ethics, transparency, GDPR compliance and all those factors. They want to make sure they draw the framework well before they start implementing things.”

He says this more cautious approach in Europe may partly be a reaction to lessons learned from the social media experience over the past 10 years, and a desire to create a more solid framework this time. But each region is also learning from each other and this is beneficial.

Influencing innovation

AI has influenced innovation and creativity both internally and on the client side at NTrust. Ramachandran explains: “When a new solution is needed for a customer, the first focus is to see how we can use AI to address this. And for creativity, across the company we want everybody to think about how they can use AI, either in their day-to-day work, or their personal lives and hobbies. We want people to see how creative they can be and in October we are going to have a digital display in which each of our employees shares how they are using AI in their day-to-day life.

“The need to come up with innovative solutions is also building from our customers. It is creating a lot of interest. People are coming back and saying ‘show us how you’re using this’, so we have brought a lot of new features into our software product, like ‘Ask me’ and ‘Lease-AI’.”

Routine functions

There are many routine functions that are now capable of being handled by AI in a business setting, he adds. “Previously, when somebody had a problem they had to run a lot of reports or go to subject matter experts to get responses, then process it and start working on that issue. Now with the features that are available they can ask intuitive questions and get responses back. So a lot of self-service is now made possible. It cuts down the time to make decisions, they’re able to handle questions by themselves very easily.”

Many human interactions, like customer support and help desks, are activities for which people are trying to use generative AI, he says. “For generative AI to work properly it needs to have good, relevant data available and the pertinent questions have to be there for it to give the right responses. Those things occur only with people who have the subject knowledge. So the human in the loop is very important to make sure of two things. One that the data quality is good, and two, what the pertinent questions to ask are. Random questions will bring bad responses, so the human in the loop is critical for effective AI solutions.”

Return on investment

The real estate sector is being somewhat cautious in its approach to AI. “AI is not cheap and especially after coming out of a tough time with higher interest rates people are cautious on how much to invest and where to invest in AI. In the short term, people are going to be doing a lot of proof of concept to see which portion of the business will get a real return on investment with AI. Then they will determine what it takes to build a good AI solution for that particular aspect of the business.”

It is different to the internet boom when the focus was less on profit and more on the level of adoption. “With AI it is going to be a combination of adoption controlled by a return on investment,” he says.

In the short term it will be a question of identifying the areas that are suitable and in the medium term the focus will be on building the right data model for those areas, determining what data is necessary.

“Everybody used to complain that they do not have good data in their system and so far people have hesitated to put all their lease data, for example, into a particular system so that it can be become effective because it was cost prohibitive and time-consuming. But now, with generative AI, companies can see that they can bring all this data together to make it much more effective in a generative data model, but at the same time be very cost-effective.”