In April, OpenAI announced its acquisition of Hiro, an AI-powered personal financial management (PFM) app that billed itself as your “Personal CFO in your pocket.” Hiro, founded by Ethan Bloch, helped users understand their financial situation, make financial plans, and manage their assets. Shortly after the acquisition OpenAI rolled out their new personal finance experience, bringing what Hiro built to life across its platform with global scale and over a billion users. Restive invested in Hiro with the thesis that AI unlocks personalization of financial advice at scale- a product that wasn’t feasible previously due to the nature of the work and associated cost structure.
Hiro was, in a sense,"unfinished business” at the time Ethan and Rushabh Doshi started it. Ethan Bloch's first company, Digit, had a similar vision and was a great outcome for everyone involved, but Ethan's real dream, the one he and Rushabh had been chasing the whole time, was bigger: a truly intelligent partner that could understand your entire financial life and actively help you navigate it.
Now that the dust has settled on the Hiro acquisition, what does this mean for the category of personal financial management? Are PFMs as a category now “finished business”? If not, where could they go from here?
We don’t believe PFMs are a finished product, but the playing field has dramatically changed.
The final form of the PFM is you ask (insert your favorite chatbot or agent) how you're doing and it tells you. It will eventually be able to take action on your behalf: “cancel that subscription”, “re-book my flight if the price drops”, or “rebalance my portfolio to my desired asset allocation”. We can safely assume that other foundational model players, plus any banking or fintech interface will layer some PFM-light product into their experience.
That product is going to be ubiquitous and good enough for 95% of consumers. Trying to compete with it head-on is a category error.
But the death of the center doesn't kill the edges. It just relocates where the interesting work happens. We continue to believe consumer financial services is one of the largest opportunities in the market. That belief is why we backed Hiro, and it's why we're still looking hard at the categories around what OpenAI just absorbed.
A few areas worth watching:
The permission layer underneath.
ChatGPT as a personal CFO only works if it, and whatever agents get built on top of it, can find your accounts and act on them. Plaid solved data aggregation for the app era by letting a developer connect to a user's bank and pull transactions. The agentic version of that problem expands the surface area of the product needed. An agent has to access the right accounts, authorize transactions inside set guardrails, manage consent, and leave an audit trail. Someone is going to build the Stripe of agentic financial permissions. Crossmint and Manif are examples of companies building some of this infrastructure.
Execution and orchestration on top of the recommendation.
When ChatGPT tells you to refinance, pay down your highest-APR card, or move idle cash into a money market fund, an agent has to go and do those things, often across several institutions at once. Each of those actions is regulated. Moving money, opening accounts, and placing trades all require a licensed entity willing to take on the liability of acting on a recommendation. The execution and orchestration layers are what turns "an AI told me to do X" into "X happened across my accounts in 90 seconds and the receipts are in my inbox." For most of their history, personal finance apps handed users data and left the strategy to them. With the Hiro acquisition, OpenAI is changing that dynamic, with the model now supplying the strategy. Agent builders are stepping in to provide execution, and Natural is one of the companies building the infrastructure for agentic finance.
Verticalization
The more regulated the action, the less willing a horizontal super app is to take it. Tax filing, fiduciary financial planning, estate planning, equity comp optimization for tech employees, business cash management, and expat tax are complicated products that touch several regulatory bodies at once. ChatGPT will likely not own these flows simply because a company with over a billion users is not built to advise in categories where it has to absorb liability, and that reluctance leaves an opening for startups. A focused company can go deep on one regulated problem, own it end to end, and organize the entire business around carrying that risk.
Outside of regulated verticals, there are opportunities for consumer finance companies to build durable businesses by serving specialized groups of consumers. ChatGPT is generally optimized for mass-affluent, English-speaking, mid-to-high-income, salaried consumers with more traditional financial profiles. Companies will exist to serve gig workers, immigrants sending remittances, families managing multi-generational finances, small business owners with mixed personal/business flows, high and ultra-high-net-worth individuals, and people in cash economies, just to name a few. Many people in those groups will likely have a super mediocre experience with the super app, and we’d expect specialized agents with specialized data integrations to win here.
Superior User Experience
People like pretty things and they always will. Building a beautiful, context rich app with well-designed dashboards, forecasts, trends, and more, like Copilot, is a differentiator that will keep consumers coming back. The winners here are and will continue to be those who can package the insights that language models unlock into beautiful representations and visualizations designed elegantly.
The pattern here is familiar to those who have watched as a horizontal platform absorbs a category. When Google ate search, auxiliary winners were vertical search (Zillow, Yelp, Indeed) and the infrastructure beneath (Cloudflare, Fastly). When the iPhone ate mobile, the winners were native mobile-first vertical apps (Uber, Instagram, DoorDash) and the platforms beneath (Twilio, Stripe). Horizontal eats the middle. Verticals and infrastructure survive and thrive.
PFMs as a standalone consumer category are probably over. That doesn't mean nothing gets built in consumer finance, rather that the bar is higher and the angle has to be sharper. You need to be either deeper than ChatGPT will ever go on a specific problem, underneath it as the rails it runs on, or pointed at a user it doesn't serve.
Hiro was the last great attempt at the PFM as a destination. It sold to the destination where everyone is already. Ethan and Rushabh got to take the dream they've been chasing since Digit and put it in front of a billion users overnight. That's the right ending for that story, and we couldn't be prouder to have been part of it.
The next decade of consumer fintech happens at the edges. We're going to be there for it. If you’re building in this category, get in touch!
© 2026 Restive®, Inc.
