OPINION | How to put AI to productive use in a business enterprise

AI augmentation is now available at all stages of the customer acquisition funnel.

Artificial Intelligence is the new buzzword. Rather, it is the “old” new buzzword. AI has been around in various forms for decades now – indeed my graduation work 30 years back was in AI-led computer vision. What have evolved over time are the specific technologies and the capabilities to apply AI to larger problems.

When you have a hammer, however, everything can look like a nail, leading to the question how do we as business leaders orient the investments in AI towards business returns?

At Indifi, we have prioritized the use of AI towards specific business objectives and have seen compelling outcomes.

The first amongst these objectives is managing credit underwriting. Credit inclusion is the core purpose of our business. Our key innovation is in being able to identify borrowers who are disqualified by the mainstream banking system but represent acceptable credit risk.

AI has allowed us to discriminate credit risk better. As a result, we are able to qualify more customers without taking undue risk – this reflects in over 30% of our borrowers being first time business credit users, and about 25% being women led businesses.

We are now also using Generative AI – the newly-found language capabilities of AI – to listen into the conversations our underwriters have with customers. This helps surface new insights from those conversations and extend our classic “human-in-the-loop” approach to actually letting human expertise improve our models through a rapid feedback loop.

Within the risk management domain, the other key application of AI has been in fraud detection and avoidance. This is enabled by analysing customer information and transactions to uncover potentially fraudulent patterns and subjecting those customers to higher scrutiny.

The second key area of AI application has been in driving growth. AI augmentation is now available at all stages of the customer acquisition funnel.

At the first stage, marketing content generation – across text and video – has allowed us to tap into the diverse needs and aspirations of our customers. AI has also proved to be a useful tool in optimizing our marketing spends. Once the customer applies to us for a loan, we then use AI based conversations to qualify customers and help them complete their application and onboarding processes.

The impact of such applications is beyond cost reduction – they allow us to get higher returns from our investments and enable growth through channels that would otherwise have been unviable.

Enhancing customer journeys using AI also allows us to drive superior customer experience – something that is reflected in a net promoter score of over 60 in relation to our onboarding process.

The third key area we have applied AI to is in optimizing costs and hence enabling to scale in a cost-efficient manner. For example, our internal costs of customer acquisition have seen over a 20% drop in the last 12 months. Similarly, we have been able to drive higher efficiency in our product development process, thereby doing more with the same resources.

Aligning the use cases of AI to our business goals has helped us focus our efforts and invest adequately behind this emerging technology. The management process that drives this is based around starting with the most significant business outcomes we are chasing, and then figuring out the best toolset, rather than what is currently in the hype-cycle.

In case of technologies that are evolving fast, we believe in extensive experimentation and persistence. AI technologies are at a stage today where hardened enterprise solutions are not pervasive – it is then our prerogative to invest and experiment with a wide range of tools to figure out what works in our environment.

Similarly, the rapid pace of evolution implies that what doesn’t work today may work tomorrow. Hence it is important to not write off promising ideas that may evolve into reality in the next year or two. Balancing this focus on business returns from investments, as well as investing in building organization-wide capabilities is an important element of getting AI right.