3 FINTECH NEWS STORIES
#1: Experian Strengthens its Orchestration Capabilities
What happened?
Experian is acquiring NeuroID:
NeuroID’s behavioral analytics solutions are available now through CrossCore on the Experian Ascend Technology Platform as a key fraud-detection capability. … With NeuroID seamlessly integrated into Experian, clients can use one service provider to proactively monitor and analyze a user’s real-time digital behavior (for example, how they navigate a form and enter information).
The emergence of generative AI-driven fraud has motivated companies across industries from financial services to healthcare and e-commerce to seek new types of fraud-detection technology, such as behavioral analytics, within their originations and account management fraud strategies. Insights from behavioral analytics help mitigate fraud in real time and prevent identity theft, account takeover, bot attacks, next-generation bot attacks and fraud rings — empowering businesses to provide a seamless customer experience.
So what?
I’m a bit biased on this one, as I have known the NeuroID folks for years now (shoutout Whitefish!) That said, I think this is a good idea for both sides.
If you are a fraud detection start-up, your destiny, 99 times out of 100, is to be acquired by one of the public-market data broker behemoths like LexisNexis or Experian. The deal terms weren’t disclosed, but I think Experian is a logical home for NeuroID (LexisNexis bought BehavioSec a few years ago).
And on the Experian side, this is a smart acquisition.
Experian has invested a lot over the years in wrapping its data assets (most of which are commodities) with sophisticated decisioning technology. The goal is to provide the system that orchestrates the data that banks and other companies acquire and executes decisions based on that data. Providing this technology makes Experian challenging to rip out and replace (and it gives them a leg up in selling their commoditized data assets).
NeuroID provides a behavioral analytics platform that can passively assess the digital body language of companies’ customers and prospects, allowing them to dynamically determine the best path forward for each individual (streamlined account opening, step-up verification or authentication, etc.) in a way that maximizes fraud prevention while minimizing unnecessary friction.
It’s a powerful orchestration tool for a company that is all in on orchestration.
#2: How Autonomous Should AI Agents Really Be?
What happened?
Skyfire Systems, a payments network for autonomous AI agents, raised a seed round:
Skyfire created a payment network specifically for AI agents to make autonomous transactions. Now, obviously, AI agents are hard to control today, so the idea of one tied to your bank account is terrifying. However, Skyfire uses a number of safeguards to prevent AI agents from overspending, making the whole thing a little less scary.
Skyfire assigns each AI agent a digital wallet with a unique identifier, where businesses can deposit a set amount of funds they want the agent to spend, so they don’t get unlimited access to a bank account. Skyfire also allows customers to set limits on how much an AI agent can spend in one transaction and over time. If an AI agent tries to overspend, it will ping a human to review it. Skyfire also offers a dashboard to view exactly how much, and where, their agent is spending.
On Wednesday, Skyfire officially launched its payment network and announced $8.5 million in seed funding raised from Neuberger Berman, Inception Capital, Arrington Capital, and other investors.
So what?
I don’t get this one.
Skyfire’s argument is that the only way to have a truly autonomous agent acting on your behalf is to enable that agent to buy stuff. Here’s Skyfire’s Co-founder and Chief Product Officer:
AI agents can’t do anything if they can’t make payments; it’s just a glorified search … Either we figure out a way where agents are actually able to do things, or they don’t do anything, and therefore, they’re not agents.
But here’s my thing — a much better search engine actually sounds just fine!
Build me an AI agent that can search for better solutions, negotiate prices, and package up everything for me to review and approve. I don’t mind being the one who pushes the button and authorizes the transaction.
In fact, given the onerous-sounding safeguards that Skyfire has built into its payments network (individual wallets for different expense categories, spend limits, a dashboard to review spending, etc.), I think it might be less work to just keep a human in the transaction loop.
This isn’t high-frequency algorithmic trading. There is no performance requirement here that only AI can meet.
Two other things about this make me nervous.
First, Skyfire isn’t building AI agents. It just wants to be the payments network. Here’s TechCrunch:
Notably, Skyfire doesn’t build the AI agents … Even though Skyfire has added safeguards, the founders say that aligning AI agents to act responsibly is ultimately up to the companies behind them.
I get why Skyfire just wants to do the payments part (much easier and less expensive!), but I’d feel better if they were taking a more vertically integrated approach.
I’d also feel better if they were using TradFi payments infrastructure, but alas:
Skyfire is solely focused on creating the payments network these agents can transact on, and did it using blockchain technology. The founders were early executives at the cryptocurrency startup Ripple, helping to build a cross-border payments network that processed more than $50 billion during their time there.
Businesses can deposit and withdraw U.S. dollars from Skyfire, but under the hood, the platform is converting those dollars into a digital stablecoin. Skyfire uses USDC, a digital stablecoin pegged to the American dollar’s value, and holds it in a wallet tied to that agent.
Skyfire collects 2% to 3% of every transaction to generate revenue
Again, this decision seems more about creating a good business for Skyfire (building on Circle leaves room for a 2% to 3% transaction fee) than it does about maximizing customer value (businesses might prefer, for a number of reasons, to use their corporate cards).
#3: Pipe’s Pivot
What happened?
Pipe officially launched its Capital-as-a-Service offering:
Pipe’s embedded Capital-as-a-Service solution for vertical ISVs and payment companies … allows SMBs to access Pipe’s capital offers inside of the software and payment systems they’re already using, and gives partners a new way to monetize payments while better serving their end merchants. Pipe’s new offering supports its mission, funding businesses beyond recurring revenue to include predictable, charge-based revenue across a wide range of vertical markets, stages, and sizes.
Initial partners in Pipe’s embedded finance program include Priority, a leading payments and banking solutions provider that helps businesses take and make payments and monetize payment networks, Infinicept, a leading provider of embedded payments, and Boulevard, creator of the client experience platform purpose-built for appointment-based, self-care businesses.
So what?
I wrote about the shift from direct small business lending to embedded small business lending in a recent article, so if you want to go deep on this space, that’s an excellent place to start.
I’m guessing the details of Pipe’s offering are similar to what is offered by others in this market (Stripe, Adyen, OatFi, Kanmon, etc.):
- White-labeled lending products offered by the SaaS platforms directly serving small businesses.
- Loans that are underwritten based on a real-time view into the small business’s cash flows (facilitated through the platforms’ payment processing relationship with the business) and repaid directly out of the business’s future incoming payments.
- The participating platforms split the revenue with Pipe, but Pipe takes on 100% of the credit losses (and exercises a lot of control over underwriting and servicing).
I like this pivot for Pipe (especially given that I was never a big fan of its prior model). One of the areas that the company seems to be leaning into, in terms of its competitive differentiation, is user journey design, according to an article from Peter Renton:
The key is design and focusing on the user journey. Each Pipe client will have a very different user journey, which is why Pipe’s head of design is involved very early on in the process. You want to surface this capital offering at the point of need or at the place where the user visits regularly. Every step in the user journey needs to be considered to ensure success.
This thoughtfulness regarding the user journey shouldn’t be surprising, given that Pipe’s new CEO (Luke Voiles) comes from Square and has had ample opportunity to learn what successful embedded banking and lending look and feel like.
The question will be how successful Pipe can be in scaling up its new Capital-as-a-Service business, given the increasing competition that we are seeing in the embedded B2B lending infrastructure space.
We shall see.
2 FINTECH CONTENT RECOMMENDATIONS
#1: Hot Takes: FDIC’s Proposed Broker Deposit Change (by Jason Henrichs, Breaking Banks) 🎧
This is the podcast episode you’ve been waiting for. Jason Henrichs, Jason Mikula, Kiah Haslett, Alexandra Steinberg Barrage, and I talking about brokered deposits for an hour.
Thanks to Mr. Henrichs for pulling us together (and keeping things spicy!)
#2: “Death of a Salesforce”: Why AI Will Transform the Next Generation of Sales Tech (by a16z) 📚
I am enjoying a16z’s meditations on how generative AI will change various incumbent businesses and legacy business processes.
This one on the impact of AI on B2B sales and CRM systems is especially interesting.
1 QUESTION TO PONDER
What products for small businesses has Stripe killed or deprecated in recent years?
I feel like Stripe has pivoted from a strategy of “build products for small businesses” to “build products for enterprise businesses, including platforms (like Shopify) that serve small businesses.”
One casualty of this shift has been Stripe’s embedded bank account product for small businesses (Opal, I think it was called?), but I am guessing there have been others, and I just haven’t noticed.
Does anyone have a list?
