The personal finance industry in India has spent a decade asking the wrong question.
The question it has been asking: how do we get users to engage with their financial data?
The question it should be asking: how do we get users to act on their financial situation?
These are not the same thing. And the distance between them is where most fintech products quietly fail the people they were built to serve.
The word is everywhere now. Personalised insights. Personalised recommendations. Personalised dashboards. Every fintech app in India uses it. Almost none of them mean it.
What personalisation means in practice, across most products in this space, is this: we show you your own data, filtered and categorised, in a slightly different order than we show it to everyone else.
That is not personalisation. That is segmentation with better UI.
Real personalisation is contextual. It knows not just what your numbers are, but what your numbers mean given your specific life situation right now.
Consider this: a ₹40,000 savings balance means something completely different for a 24-year-old with no dependents and a stable job than it does for a 28-year-old supporting two parents with an EMI landing in four days. Same number. Completely different financial reality. Completely different recommended action.
No Indian finance app today can tell the difference. They see the ₹40,000. They do not see the situation around it.
That is the personalisation gap.
There are two reasons, and they are not equally hard to fix.
The first is data infrastructure. Until the Account Aggregator framework arrived, building a complete picture of a user’s financial life required screen scraping or manual input. Both break at scale. Both erode user trust. AA changes this fundamentally. For the first time, a consented, real-time, multi-institution view of a user’s finances is technically achievable. Salary account, savings account, credit card, investments: one view, with the user’s explicit permission. The data infrastructure problem is being solved at a regulatory and ecosystem level. This is not a startup problem anymore. It is a solved problem that most startups have not yet built on top of.
The second reason is harder. It is a product philosophy problem.
Most fintech apps were built to display data, not interpret it. The engineering effort went into aggregation, categorisation, and visualisation. Beautiful charts. Clean transaction feeds. Fast sync. The question of what to do with the data once you have it was treated as secondary. Something to layer on later, with AI, once the data layer was stable.
Later is now. And most apps are not ready for it. Because displaying data and interpreting a situation are not engineering problems that exist on the same spectrum. They require a fundamentally different orientation from the beginning. You cannot retrofit a dashboard into a decision engine. The architecture is different. The product philosophy is different. The thing you are optimising for is different.
Blume’s Indus Valley report identified AI and personalisation as the defining themes of the next wave of Indian consumer technology. That framing is correct. The opportunity is real and the timing is right.
But the report, by necessity, describes the opportunity from the outside. From the inside, building a product in this space every day, the picture is more specific.
The personalisation opportunity in Indian personal finance is not about better recommendation engines. It is not about showing users the right mutual fund at the right moment based on their age and risk profile. Those are useful features. They are not the structural opportunity.
The structural opportunity is this: build a product that understands the difference between a user’s financial data and their financial situation, and closes that gap in real time.
That requires three things working together.
Complete financial data through AA. Without it, the picture is always partial and interpretation is always guesswork.
A conversational intelligence layer that can interpret context, not just process queries. A user asking “how much can I spend this weekend” is not asking for their account balance. They are asking for an assessment of their current financial situation relative to their upcoming commitments. Those are different questions that require different answers.
A product architecture oriented around decisions rather than dashboards. Every screen, every insight, every notification should exist to move a user from knowing to doing. If it does not do that, it is noise.
All three are hard to build. None of them is impossible. And the window to build them before incumbents wake up is 2025 to 2027.
Last month, a user we will call Arjun opened Finanjo at 6pm on a Sunday.
His first message to Jo, Finanjo’s AI layer, was three words: “Show my expense breakdown.”
Standard. That is how most users start.
What happened in the next 13 minutes was not standard.
Jo identified two recurring EMIs totalling ₹57,449 a month against a salary of ₹1.67 lakhs. EMI-to-income ratio of 34%, technically healthy, but clearly something Arjun wanted to improve. Then Arjun asked something that reframed the entire session: “Shall we look at my spending pattern to see how much more I could comfortably put towards prepayments?”
He did not ask how much he was spending. He already knew. He asked Jo to look at everything together and find the room between his current life and a better financial outcome.
What Jo found: in April, 97.6% of Arjun’s spending had gone to EMIs. The remaining 2.4% to essentials. Zero discretionary. This is a disciplined person. He does not need to be told to spend less. He needs to know what his discipline makes possible.
Jo built him a specific prepayment goal for May. Built from his actual surplus, his actual salary average, his actual fixed commitments. Arjun set it up before he closed the app.
The whole session: 13 minutes.
This is not a product demo. This is what the personalisation gap looks like when it starts to close. Arjun did not get a pie chart and a generic tip about reducing food delivery spending. He got a plan built for his situation, his numbers, his goal. That is the difference between knowing a user’s balance and knowing their situation.
Across 5,000+ conversations with Jo, the pattern is consistent.
60% of user queries fall outside every intent category we built into our classification system. Not balance checks, not budget queries, not spending breakdowns. Something else. Questions too specific, too contextual, too human to fit a predefined taxonomy.
“If my salary stopped next month, how long would I last?”
“Should I put ₹5,000 in SIP or pay off my credit card this month?”
“My in-hand is ₹19K. I give ₹15K to my parents. I want to do a yoga certification. How do I save for it?”
These users are not confused about finance. They understand their situation. What they are missing is someone to think through it with them, someone who has all their data and can help them decide.
That 60% is not a classification failure. It is the clearest signal we have that the financial lives of young Indians are more contextual, more specific, and more complex than any dashboard was built to handle.
AA makes the data available.
AI makes the interpretation possible.
The product philosophy is the only variable left. And product philosophy is the one thing that cannot be copied quickly. It has to be built from the beginning, from a clear answer to the question: what is this product actually for?
If the answer is “to show users their financial data,” the competition is every team with a good design sprint and an AA integration. That category gets commoditised fast.
If the answer is “to close the gap between a user’s financial situation and their next best action,” the competition looks completely different. Because that requires knowing the user, earning enough trust that they tell Jo their wife’s name and their parents’ expectations and their salary-day anxiety, and building a product architecture that converts all of that context into a clear and specific next step.
The apps that get to that second answer first, and build deeply enough that the first 600 users become the foundation of a proprietary understanding of how young India makes financial decisions, those are the apps that matter in 2030.
The personalisation gap is real. The infrastructure to close it is here. The window is open.
The question is who builds in it.
The personal finance industry in India has spent a decade asking the wrong question.
The question it has been asking: how do we get users to engage with their financial data?
The question it should be asking: how do we get users to act on their financial situation?
These are not the same thing. And the distance between them is where most fintech products quietly fail the people they were built to serve.
The word is everywhere now. Personalised insights. Personalised recommendations. Personalised dashboards. Every fintech app in India uses it. Almost none of them mean it.
What personalisation means in practice, across most products in this space, is this: we show you your own data, filtered and categorised, in a slightly different order than we show it to everyone else.
That is not personalisation. That is segmentation with better UI.
Real personalisation is contextual. It knows not just what your numbers are, but what your numbers mean given your specific life situation right now.
Consider this: a ₹40,000 savings balance means something completely different for a 24-year-old with no dependents and a stable job than it does for a 28-year-old supporting two parents with an EMI landing in four days. Same number. Completely different financial reality. Completely different recommended action.
No Indian finance app today can tell the difference. They see the ₹40,000. They do not see the situation around it.
That is the personalisation gap.
There are two reasons, and they are not equally hard to fix.
The first is data infrastructure. Until the Account Aggregator framework arrived, building a complete picture of a user’s financial life required screen scraping or manual input. Both break at scale. Both erode user trust. AA changes this fundamentally. For the first time, a consented, real-time, multi-institution view of a user’s finances is technically achievable. Salary account, savings account, credit card, investments: one view, with the user’s explicit permission. The data infrastructure problem is being solved at a regulatory and ecosystem level. This is not a startup problem anymore. It is a solved problem that most startups have not yet built on top of.
The second reason is harder. It is a product philosophy problem.
Most fintech apps were built to display data, not interpret it. The engineering effort went into aggregation, categorisation, and visualisation. Beautiful charts. Clean transaction feeds. Fast sync. The question of what to do with the data once you have it was treated as secondary. Something to layer on later, with AI, once the data layer was stable.
Later is now. And most apps are not ready for it. Because displaying data and interpreting a situation are not engineering problems that exist on the same spectrum. They require a fundamentally different orientation from the beginning. You cannot retrofit a dashboard into a decision engine. The architecture is different. The product philosophy is different. The thing you are optimising for is different.
Blume’s Indus Valley report identified AI and personalisation as the defining themes of the next wave of Indian consumer technology. That framing is correct. The opportunity is real and the timing is right.
But the report, by necessity, describes the opportunity from the outside. From the inside, building a product in this space every day, the picture is more specific.
The personalisation opportunity in Indian personal finance is not about better recommendation engines. It is not about showing users the right mutual fund at the right moment based on their age and risk profile. Those are useful features. They are not the structural opportunity.
The structural opportunity is this: build a product that understands the difference between a user’s financial data and their financial situation, and closes that gap in real time.
That requires three things working together.
Complete financial data through AA. Without it, the picture is always partial and interpretation is always guesswork.
A conversational intelligence layer that can interpret context, not just process queries. A user asking “how much can I spend this weekend” is not asking for their account balance. They are asking for an assessment of their current financial situation relative to their upcoming commitments. Those are different questions that require different answers.
A product architecture oriented around decisions rather than dashboards. Every screen, every insight, every notification should exist to move a user from knowing to doing. If it does not do that, it is noise.
All three are hard to build. None of them is impossible. And the window to build them before incumbents wake up is 2025 to 2027.
Last month, a user we will call Arjun opened Finanjo at 6pm on a Sunday.
His first message to Jo, Finanjo’s AI layer, was three words: “Show my expense breakdown.”
Standard. That is how most users start.
What happened in the next 13 minutes was not standard.
Jo identified two recurring EMIs totalling ₹57,449 a month against a salary of ₹1.67 lakhs. EMI-to-income ratio of 34%, technically healthy, but clearly something Arjun wanted to improve. Then Arjun asked something that reframed the entire session: “Shall we look at my spending pattern to see how much more I could comfortably put towards prepayments?”
He did not ask how much he was spending. He already knew. He asked Jo to look at everything together and find the room between his current life and a better financial outcome.
What Jo found: in April, 97.6% of Arjun’s spending had gone to EMIs. The remaining 2.4% to essentials. Zero discretionary. This is a disciplined person. He does not need to be told to spend less. He needs to know what his discipline makes possible.
Jo built him a specific prepayment goal for May. Built from his actual surplus, his actual salary average, his actual fixed commitments. Arjun set it up before he closed the app.
The whole session: 13 minutes.
This is not a product demo. This is what the personalisation gap looks like when it starts to close. Arjun did not get a pie chart and a generic tip about reducing food delivery spending. He got a plan built for his situation, his numbers, his goal. That is the difference between knowing a user’s balance and knowing their situation.
Across 5,000+ conversations with Jo, the pattern is consistent.
60% of user queries fall outside every intent category we built into our classification system. Not balance checks, not budget queries, not spending breakdowns. Something else. Questions too specific, too contextual, too human to fit a predefined taxonomy.
“If my salary stopped next month, how long would I last?”
“Should I put ₹5,000 in SIP or pay off my credit card this month?”
“My in-hand is ₹19K. I give ₹15K to my parents. I want to do a yoga certification. How do I save for it?”
These users are not confused about finance. They understand their situation. What they are missing is someone to think through it with them, someone who has all their data and can help them decide.
That 60% is not a classification failure. It is the clearest signal we have that the financial lives of young Indians are more contextual, more specific, and more complex than any dashboard was built to handle.
AA makes the data available.
AI makes the interpretation possible.
The product philosophy is the only variable left. And product philosophy is the one thing that cannot be copied quickly. It has to be built from the beginning, from a clear answer to the question: what is this product actually for?
If the answer is “to show users their financial data,” the competition is every team with a good design sprint and an AA integration. That category gets commoditised fast.
If the answer is “to close the gap between a user’s financial situation and their next best action,” the competition looks completely different. Because that requires knowing the user, earning enough trust that they tell Jo their wife’s name and their parents’ expectations and their salary-day anxiety, and building a product architecture that converts all of that context into a clear and specific next step.
The apps that get to that second answer first, and build deeply enough that the first 600 users become the foundation of a proprietary understanding of how young India makes financial decisions, those are the apps that matter in 2030.
The personalisation gap is real. The infrastructure to close it is here. The window is open.
The question is who builds in it.
