I’ve spent most of my career building fintech products for distribution apps, platforms, and now WhatsApp‑based AI flows so I know deeply about DSA Behaviour in Indian Fintech. Some worked, some failed, and one in particular the OneAndro app at Andromeda, completely changed how I think about behaviour, technology, and legacy systems in Indian fintech.
This is a detailed case study of that journey: why OneAndro struggled despite a large partner base, what I learnt about agents, RMs, bankers, and culture, and how those lessons helped me design WhatsApp‑native flows at Duvanta that actually got used every day.
Before OneAndro: The “App‑First” Mindset
Before joining Andromeda, I was already deep into building fintech apps. At Afinoz, I worked on an all‑in‑one app for loans, mutual funds, and insurance, and then on Afinoz Lite for instant personal loans from multiple lenders. The default thinking was clear: if you want scale and control, you build an app; if you want more adoption, you add features to that app.
So when I joined Andromeda, it was natural for me to think “partner app” as the solution. The idea was straightforward: give employees and DSAs a single app to sell and manage all financial products.
That idea became OneAndro.
OneAndro: Vision vs Reality
What OneAndro was supposed to be
OneAndro was designed to be the single operating system for Andromeda’s employees and partners:
- Source and manage leads for all financial products
- Upload client documents and data in a structured way
- Track file status, approvals, and disbursals
- Reduce dependency on phone calls, manual follow‑ups, and scattered communication
In discussions and pre‑launch conversations, employees and partners said they would use it. On slides, it looked like the right move – one app, full visibility, better control.
The real numbers on the ground
Once OneAndro was live (launched in May 2024), the numbers told a different story.
-
Around 42,000 partners were registered with Andromeda.
-
Roughly 17,000 had downloaded the OneAndro app and had at least some work happening in a month.
-
On a normal day, only about 300–500 partners were actually active inside OneAndro.
-
When top management pushed hard, daily activity could spike to around 1,700 users for 2-3 days, then drop back to 300-500 again.
We also tried to enforce that a specific NBFC’s business must run only through OneAndro. In that phase, Andromeda saw roughly a 40% drop in business for that NBFC from our side.
On paper, 17K users and spikes to 1.7K daily actives look decent. In reality, behaviour told me that OneAndro was not being chosen voluntarily.
The Hidden Triangle: Partner, RM, and Banker
To understand why OneAndro struggled, I had to look closely at the existing workflow inside Andromeda.
The real operating model for many years was a triangle:
Partner/employee → WhatsApp → Andromeda RM → Banker and backend → File allocation under Andromeda.
In practice:
- Partners and employees would collect documents and basic information from customers.
- Instead of using any platform, they would forward everything over WhatsApp to an Andromeda relationship manager or sometimes directly to a banker.
- The RM or banker would handle almost everything: file login, eligibility checks, documentation validation, back‑and‑forth with the lender, and getting approvals.
- Once done, the banker would allocate the file under Andromeda’s name because they also had their own sales targets, and they often provided extra support or small facilities to Andromeda partners to keep the relationship strong.
Over time, this created a very clear pattern:
- Partners did not need to know detailed eligibility rules for each bank/NBFC.
- They did not need to learn systems beyond WhatsApp.
- A lot of intellectual and operational work shifted to bankers and RMs, while partners became more dependent on this setup.
In that context, OneAndro wasn’t just a new app. It was asking partners and employees to take back work they had long outsourced to RMs and bankers.
“Lazy” or “System‑Shaped”? What I Actually Learnt
In the early phase, my internal narrative was simple: partners and employees are lazy; they don’t want to do tasks themselves; they prefer just forwarding files on WhatsApp. But over time, and after speaking to competitors, my understanding changed.
I realised:
- People were not inherently lazy or anti‑tech.
- They were behaving exactly as the system had trained them for decades – in a model where offloading work to RMs and bankers was normal, and still produced business and payouts.
So the problem at Andromeda was not just “people don’t want to change.” It was that the entire ecosystem, incentives, processes, support had been built around a non‑digital, WhatsApp + RM + banker‑driven model since 1990.
OneAndro was trying to overwrite 30+ years of behaviour in one shot.
Why the Same Partners Use Basic’s App But Not OneAndro
After leaving Andromeda (April 2025), I spoke to founders of competitors like Basic Home Loan and Ambak, who are friends. Those conversations were eye‑opening.
Basic’s founder, Atul, told me directly:
- The same partners who were not adopting OneAndro at Andromeda were working with Basic and successfully using Basic’s app.
- From day one, Basic’s clear policy was: partners must work through the Basic platform (app/web). There was no parallel workflow where they could offload everything to bankers over WhatsApp.
- Once people joined Basic, they adapted to that system without major issues, because the operating rules were digital‑first from the start.
Ambak had a similar stance:
- They were clear that without digital, they would not run the business.
- They provided support to partners at their own cost, but the entire process had to be driven through their app or web portal. There was no informal “bypass” path.
The contrast:
- Andromeda: Founded in 1990, legacy processes, deep WhatsApp + RM + banker habits, many parallel “offline” paths.
- Basic (founded around 2020): Digital‑first policy, platform usage non‑negotiable, behaviour shaped from day one.
- Ambak: Digital‑only mindset; strong support, but always via their app or portal.
This is where the key insight became clear to me:
It’s not the people, it’s the system.
The same employee or partner can resist an app in one company and adopt it smoothly in another, depending purely on the culture, rules, and incentives of that organisation.
The Bigger Picture: India Is WhatsApp‑First, Not App‑First
Zooming out from my personal experience, the market trend supports the same story.
- WhatsApp is India’s dominant communication layer, with hundreds of millions of users and the country as its largest market.
- In fintech and BFSI, WhatsApp Business API has become a primary engagement surface for many players used for KYC, EMI reminders, transaction alerts, and end‑to‑end service journeys.
- Industry data shows significantly higher open and response rates on WhatsApp compared to email/SMS, and many fintechs report lower drop‑offs when they move journeys into chat.
So while some newer players like Basic and Ambak intentionally built app‑first systems, a large part of the Indian ecosystem still runs on WhatsApp‑driven workflows.
At Andromeda, partners were fully embedded in that WhatsApp reality. OneAndro was trying to pull them away from it; that’s why it felt heavy.
The Pivot: Moving From App to WhatsApp at Duvanta
After Andromeda, I joined Duvanta as a founding member for AI and product. This time, I decided not to fight WhatsApp.
The thesis was simple:
- If partners live on WhatsApp, build the product on WhatsApp.
- Instead of asking them to learn a new app, let them continue chatting – but turn those chats into structured, automated workflows.
At Duvanta, we built WhatsApp‑based AI tools that allowed partners to:
- Create loan files inside a WhatsApp conversation
- Run automatic loan eligibility checks
- Identify required documents based on the case
- Perform credit score checks and basic analysis directly in chat
Behind the scenes, we wired the logic, integrations, and AI; on the front end, partners were still just sending messages and documents in the same app they use all day.
Duvanta’s Numbers: From 1-2 Files per Month to 1-2 Files per Day
The contrast in output between OneAndro and Duvanta was very clear.
At Andromeda with OneAndro:
- Even with 42K registered partners and 17K app downloads, we were often seeing only 1-2 files per month coming through the app from some partners.
- Most of the real activity was still happening via WhatsApp and manual workflows.
At Duvanta, with WhatsApp‑native flows:
- We onboarded around 80 partners in beta.
- Out of those, roughly 55-60 actively used our WhatsApp flows.
- Instead of 1-2 files per month, we started seeing 1-2 files per day flowing through the WhatsApp journeys.
Same type of users (DSAs/partners), same category (loans), same market but:
- OneAndro asked them to change tools and behaviour.
- Duvanta met them inside their existing behaviour and tools.
The output difference from monthly trickles to daily flows confirmed that this was not just a small tweak; it was a fundamentally better alignment with reality.
Case Snapshot Table: Andromeda vs Basic vs Ambak vs Duvanta
This table makes one thing obvious: the same human being behaves differently depending on what the system normalises and enforces.
Key Learnings From OneAndro vs Duvanta
1. Behaviour beats features
OneAndro had features. But it fought against established behaviour: “send & forget on WhatsApp, let RM/banker handle the rest.” Duvanta respected that behaviour and embedded structure and intelligence inside WhatsApp itself.
Now I start with behaviour and build technology around it, not the other way around.
2. Adoption is about voluntary repeat use, not downloads
OneAndro:
- 42K registered partners
- 17K app downloads
- 300-500 genuine daily active users most days
- Spikes to ~1,700 under pressure for 2-3 days
- 1-2 files per month from some partners through the app
Duvanta:
- 80+ onboarded partners
- 55-60 active on WhatsApp flows
- 1-2 files per day instead of per month
Today, I don’t celebrate installs or forced DAU spikes. I look for behaviour that continues even when nobody is pushing.
3. Forcing tools can reduce, not increase, business
When we tried to enforce OneAndro for a specific NBFC, business for that NBFC through Andromeda dropped by about 40%. Partners simply routed that NBFC’s cases to other companies who allowed WhatsApp‑based workflows.
In a relationship‑driven ecosystem, technology is always optional. If your platform adds friction, people will quietly route around it.
4. It’s not the people, it’s the system
Basic’s founder told me clearly: the same employees and partners coming from Andromeda adopt technology at Basic without issues, because the rules and expectations are set that way from day one.
So the pattern is:
- At Andromeda, the system rewarded offloading work to RMs and bankers.
- At Basic and Ambak, the system rewarded using the platform.
- At Duvanta, the system embraced the channel partners already lived in (WhatsApp) and layered intelligence on top.
This shifted my lens from “people are lazy” to “systems shape behaviour.”
5. WhatsApp is now a primary surface in Indian BFSI
Across the industry, WhatsApp has become a core interface for banks and fintechs – from Union Bank’s WhatsApp banking to many fintechs running onboarding, KYC, support, and reminders through WhatsApp Business API.
At Duvanta, designing the product as WhatsApp‑first aligned with this broader shift: conversational journeys, higher engagement, and lower friction for both partners and customers.
What This Means For Fintech Builders and Leaders
If you’re building for DSAs, brokers, or similar networks in India, here are my straight conclusions from these years:
- Do not assume an app will automatically fix behaviour; design for where users already live (often WhatsApp), then decide what should be app vs chat.
- Watch what people do, not just what they say in internal meetings. Pre‑launch “yes” means nothing if the actual habit is still WhatsApp + RM.
- Be very intentional about your operating model: will you allow parallel offline/WhatsApp flows, or is your platform truly the system of record?
- If you inherit a legacy culture, start with low‑friction conversational experiences and gradually move structure into them instead of forcing a hard jump to heavy apps.
For me, OneAndro was a difficult but important lesson. It showed me how powerful legacy behaviour and incentive structures can be. Duvanta is where I started applying those learnings and the shift from 1-2 files per month to 1-2 files per day told me very clearly which side of the behaviour vs app debate reality is on