Cushion AI’s backstory is a little on the personal side. It didn’t want to make the flashiest product in fintech. Instead it wanted to build a product that was painfully relatable. Something almost everyone suffered with and needed help: a fight back against bank fees. And for millions of people this sounded like a win. For a long time, it was actually a win.
But at the end of 2024, after 8 years, with $21.6 million in funding, 200,000 paying customers, and over $3 million in annual revenue, Cushion AI shutdown. The founder and CEO of Cushion AI made the news official through a LinkedIn post, it said:
The outcome wasn’t what we hoped for, but we moved the industry forward
Cushion had the product and the funding, but like many, it struggled with startup scale. Let’s find out what went wrong and see what we can learn.
What Cushion Set Out to Do
As I mentioned earlier, the idea behind Cushion was a little personal. Paul Kesserwani while helping his parents manage their bank accounts overseas noticed how hidden fees piled up fast. Which then made him look at his own account and to his surprise he saw more than $400 in random fees, which he hadn’t even noticed in the first place.
That’s how Cushion was born.
The app was super helpful and it connected to your bank account. After which it scanned your transaction history, and automatically identified fees like overdrafts, ATM charges, and card replacement fees. But that was not all, there was more. Cushion even negotiated refunds with banks on your behalf, and cut a commission for it only if it was successful in getting your money back.
How It Grew
In the year 2016, Cushion was launched and it quietly built is userbase. In just three years it had over tens of thousands customers and had recovered over $1 million in bank fees! Then in 2 years, in 2022 it had raised $12 million Series A, pushing its total funding to $21.6 million. It’s safe to say that investors loved the model and the concept it served. It helped people save money and it aligned incentives. Furthermore, the company operated without pushing credit cards or ads like many other fintech apps, which obviously people loved.
By 2023, Cushion had:
- 1 million+ users
- Over 200,000 paying customers
- $3 million ARR (Annual Recurring Revenue)
- $300 million in processed BNPL (Buy Now Pay Later) loans
So, What Went Wrong?
Despite having all that momentum, Cushion failed to survive. Kesserwani said it simply: “We didn’t reach the scale needed to sustain the business.”
Here’s my input on what likely went wrong:
1. Scale vs. Stickiness
Despite millions of people downloading the app, not all stayed or paid. Realistically fee negotiation is not something people need help with daily. No doubt that Cushion had a cool feature, but it did lack on daily utility. Which thereby made customer retention hard.
2. Platform Challenges
Initially Cushion used Plaid, which is a popular net banking app. But eventually it got kicked off as it was holding user credentials, something that Plaid didn’t allow. So what did the company do next? The CEO tried building his own infrastructure, which actually worked, but took a lot of time and recourses. ‘
3. Regulatory Hurdles and Bank Behavior
Banks were not a fan of automated negotiators and did not welcome them. The reason behind it was quite simple, if millions of users started auto-requesting refunds, banks could easily shut down their co-operations. Or instead they could change their systems to block third-party apps.
4. Burn Rate vs. Growth Rate
Even though Cushion had paying users. The cost it had to bare for infrastructure, hiring engineers and handling customer support was likely more.
What We Can Learn from Cushion
Cushion leaves back valuable lessons that not just startup founders can learn, but for anyone building something meaningful:
1. Infrastructure Can Be a Moat or a Trap
When Cushion lost the partnership with Plaid, it had stand back up with it’s own help. It gave it control and power but also added complexity. One must know when to build vs. partner.
2. Growth Isn’t Enough, You Need Frequency
Solving once-in-a while problem once makes the retention hard. No one keeps visiting the apps once their work is done. Even if you help your customers in a huge way, if they don’t come back often, it makes monetizing hard.
3. Aligning Incentives Helps, But Isn’t a Business Model on Its Own
Cushion AI made money only when users saved money. No doubt that it is noble and that helped them build trust. But what it also meant was unpredictable revenue. Along with great ethics one needs great economics.
Cushion did not fail because it lacked vision. It failed even after having good ethics, and being well funded, this is the truth of startups, they face brutal realities in the fintech world, just like Dunzo did in its own shutdown story.
