AI has found its way into the entire value chain of the financial services sector in the past 3-4 years.

At large, AI (artificial intelligence) for fintech enterprises is all about crafting valuable customer trajectories and improving UI/UX experiences. The unstructured data extracted through AI is modelled in augmenting deep learning techniques and understanding consumer preferences. Specifically, large language models, speech recognition are helping to offer bespoke customer support and tailoring the outcomes. According to a research report by Mordor Intelligence, a leading market intelligence and advisory firm, the AI-powered Indian fintech market is estimated to hit $1.3 trillion by 2025 with more startups bringing in more and more exciting suite of products.

Most of the enterprise data insight is used in expanding the premise of applications/use cases.

Angel One, the online stock broking and trading house combines static attributes like names, slow-moving variables like addresses and fast-moving ones like clickstream to generate a digital persona for the customer. In a recently concluded experiment where a specific customer query was handled by both man and machine, the firm recorded 70% accuracy in just 15 days of training.

"It needs augmentation of data learnt from a knowledge base combined with near-realtime and real-time signals. We are working to further improve the accuracy specifically with temporal data," says Jyotiswarup Raiturkar, CTO, Angel One Limited.

"We combine user-consented work and income data with behavioural data derived from digital footprints to build a credit score for the new-to-credit or credit-invisible population. AI comes to use to assess the creditworthiness of individuals and offer flexible and affordable financial products tailored to their needs," says Rohit Rathi, co-founder & CEO Karmalife, the credit-solutions provider for India's gig and blue-collar workers.

"In a competitive world where customer retention is a challenge, deploying technology helps us efficiently assess the risk profile of a customer, which is a breakthrough for underwriting. Moreover, AI-powered chatbots help solve customer queries and make the claim process quick and seamless. AI has indeed made the insurance experience for customers fast, flexible, and reasonably progressive," says Saurabh Tiwari, Chief Technology Officer, Policybazaar.com, an online marketplace for insurance products.

Not just the consumer mapping and deliverables portfolio, businesses leverage tech to scale in analytics, product development innovations and reach out to new markets.

"Earned Wage Access product, one of our recent products allows workers to access a portion of their salary before payday. It is powered by AI algorithms that assesses users' financial behavior and needs," explains Rathi.

Angel One's pipeline is fuelled by the recent breakthroughs in interactive and generative AI. "With chatGPT, transition from UI/UX to conversational interfaces is underway. Another example of new ideas is how our API platform is supported. In place of multiple Software Development Kits (SDKs) integration guides, with a bit of SEO, we can direct the LLM to generate code to use smart API," says Raiturkar.

In addition to AI and ML (machine learning) that have revolutionised the landscape, hopes are also on closed-loop operations through blockchain to bring much secure, efficient, and inclusive operations. "Fairer premiums across processing, fraud detection and better data transparency will be possible through the chain," says Tiwari of Policy Bazaar.

Automation has indeed unlocked the hidden possibilities and brought in time, cost, and operational efficiencies. But enterprises still choose to rely on the human touch to take the experience further and avoid over-reliance. Smart engineering talent is the driving force behind these enterprise tech systems.

Tiwari reasons out thus. "Customer experience is of paramount importance, so human vigilance is indispensable to navigate through tech."

Fintechs are finding success due to small nimble tech-focused teams and tech with customer focus as compared to incumbent players, says Raiturkar. "Building such deep tech systems is a moat only some teams can broach with commodity integrations. Nevertheless, to control cost in building and maintaining AI," he adds. "We have partnerships with select startups who are SMEs in niche areas, work in the PODs (Product-Oriented Development) model when going live with the projects."

Though angel investing and crowd funding has democratised access to business capital for these new-age fintechs, more regulatory support, as Tiwari points out, "…from strategic regulation formulation to tactical matters like technical feasibility, t-shirt sizing etc," especially in this time when international banks have collapsed and traditional banking is in distress, will help in expanding innovation. As for sustainability, Rathi prefers to stand by transparency in processes and upholding stringent data privacy standards. "A thorough understanding of the regulatory landscape and responsible use with strong commitment to user benefit and protection will be the ideal route to create lasting impact, win customer support and grow."

ML in fintech

From customer profile to improving accuracy, fintech takes to AI