JPMorganChase's $20bn tech spend yields outdated app

PayPal is cutting 4,760 jobs. Not because the business is failing. Because AI can do the work instead. Their CEO said it plainly: "We will accelerate our AI adoption across our operations." Coinbase cut 700 people the same week. Same reason. This isn't a layoff wave. It's a restructuring wave. And it's just getting started. Here's what most people aren't saying: The finance professionals who survive this aren't the ones who know finance best. They're the ones who know how to work with AI. The analyst who uses AI to process 10,000 data points in an hour. The CFO who uses AI agents to flag risk before it becomes a crisis. The compliance officer who uses AI to audit every call — not just a sample. Those people become 10x more valuable. The rest become a line item to cut. The question isn't "will AI take my finance job?" The question is: "Am I the person who knows how to use it?"

Coinbase cut 14% of its employees. PayPal is cutting 20% over two to three years. BILL is cutting up to 30%. All three framed it as AI transformation. Alex Johnson's Fintech Takes send from earlier this week breaks down why the "AI restructuring" narrative is more complicated than the headlines suggest. Five things he argues are simultaneously true: 1️⃣ Software companies have been overstaffed for years. The 2022-2024 layoff waves were the first correction. These cuts are the second. AI is the new rationale, but the overcapacity predates it. 2️⃣ AI is collapsing the cost of building competing software, which makes incumbents' moats a lot less defensible than they used to be. 3️⃣ White-collar productivity gains are real, but who captures them first matters. Right now, it's the companies. 4️⃣ Nobody actually knows what an AI-native org looks like. The companies making these cuts are guessing. 5️⃣ Investors and policymakers are equally lost. PayPal and BILL made nearly identical announcements in the same week. One stock went up 8%. The other dropped 10%. The bottom line: if you restructure early and guess right, it's strategy. If you guess wrong, it's panic. The tricky part is that no one can reliably tell the difference yet. Read the full analysis at the link in comments.

This morning I stopped by my local JPMorganChase branch, looking to cross an item off my To Do list, a withdrawal over the ATM limit. 15 minutes in line. No big deal. That’s part of the tradeoff when you need human help. Then an employee steps out and says: “All Chase systems are down. We can’t process any transactions.” And there goes the ENTIRE value proposition of the physical branch. 💨 This isn’t really about Chase. These things happen. Systems fail. Technology is imperfect. Standing in line only to learn nothing can be done, isn’t a technology failure, it’s a communication failure. In healthcare, we talk a lot about experience: client experience, referral partner experience, and caregiver experience. What a reminder this was that experience is often shaped before the interaction even starts. * Could there have been signage at the door? * A push notification to customers? * An alert on the location specific page on the website and app? * Someone proactively greeting customers walking in? Those are small things. But they change everything. Technology outages are inevitable. Poor experiences around them are not. The organizations that win aren’t the ones that avoid failure, they’re the ones that design for it. And zooming out, this problem is only getting bigger. We’re consolidating onto fewer platforms, fewer vendors, and more shared infrastructure. It drives efficiency… until it doesn’t. When something breaks now, it doesn’t impact a system, it impacts everything. We’ve all seen it recently with CrowdStrike, AWS, and Cloudflare. These aren’t edge cases anymore, they’re the new operating reality. Fewer systems. Bigger blast radius. The differentiator isn’t uptime only, it is how you operate when you DON’T have it. I am curious about how others are thinking about this. How are you designing your front-line experience for the moment your systems go down? #chase #customerexperience #outages

I've been covering impact of AI on financial institutions for a while In last week's Finance Upturned newsletter, I pointed to Coinbase and Cloudflare's recent ~20% layoffs and predicted it will soon come to banking It's here sooner than I expected with Standard Chartered now announcing job cuts due to AI I can only hope these large companies are doing their best to support employees laid off

PayPal cut 4,760 jobs on Tuesday. Coinbase cut 700 the same morning. Both CEOs are named AI. The cuts followed Anthropic shipping ten finance-specific AI assistants by 48 hours. That sequence is the read. PayPal's new CEO, Enrique Lores, who started on March 1, announced cuts of roughly 20% of the 23,800-person workforce as part of a two- to three-year reorganization targeting $1.5 billion in run-rate savings. Lores explicitly framed the cuts as part of an AI integration plan, not a stand-alone austerity move. Coinbase CEO Brian Armstrong cut approximately 700 employees — 14% — describing the move in his memo as rebuilding the company "as an intelligence with humans around the edge." Coinbase expects $50-60 million in restructuring charges in Q2. The structural detail is what the cuts are paying for. Two days earlier, Anthropic shipped 10 finished AI assistants for finance work, pitchbook generation, KYC, earnings review, and month-end close, running as plugins inside Microsoft 365, with the Moody's 600-million-company database wired in directly. Agents shipped on Sunday. Layoffs landed on Tuesday. The labor reduction at the buyer and the agent shipment from the lab landed inside the same 72-hour window. That is not a coincidence. That is the buyer and the seller pricing the same trade. The expensive part of a financial workflow used to be the analysts. It is now the API call. The honest read for any finance-services leader is that the productivity payoff is being banked before the productivity actually arrives. If your role intersects pitchbook prep, KYC, fraud screening, month-end close, or earnings analysis — what is the one workflow you would scope as agent-first this quarter, and what would the headcount answer have to look like to make it real? Full read on Substack #ArtificialIntelligence #FinancialServices #EnterpriseAI #AILayoffs #FutureOfWork #Anthropic #PayPal #Coinbase

PayPal is repositioning itself as a technology first business once again, placing artificial intelligence at the centre of a major restructuring effort... Read more; https://lnkd.in/dHrWrHfY #AfricaBusinessInsight #AfricaBusinessNews

PayPal will cut 4,760 jobs. Their new CEO said it clearly: "We are aggressively adopting AI in our development processes." Target: $1.5 billion in savings. They're cutting while profitable, not because the business is struggling. Q1 revenue: $8.4 billion. Up 7%. Strong earnings + AI restructuring = fewer people, same (or better) output. And PayPal isn't the first. Block did it. Coinbase did it. Cognizant did it. #AI #Automation #FutureOfWork #Ubiqflow

Coinbase had 8 degraded-performance incidents in the past month including a 6.5-hour outage on May 7–8 that put markets into cancel-only mode. The last one gets the headlines because AWS overheated a data center in Virginia. Easy story, external villain. One outage is bad luck. But eight in thirty days is a pattern — and patterns are architectural, not accidental. What eight incidents in a month actually means: somewhere in the system, the resilience work nobody wanted to prioritize is overdue. The runbook that never got written. The cross-zone dependency nobody mapped. The failover path that was “tested in staging” two years ago and never since. That work is invisible right up until the month it isn’t. Three days before the big one, the CEO announced 14% layoffs and pitched an AI-native future where one person ships what teams used to. Cause and effect? No — these architectural choices predate the layoffs by years. But it’s a signal about what the org values going forward. And the failure mode on display is exactly the kind that gets worse, not better, when speed beats judgment. I wrote earlier this week about the prototype gap (https://lnkd.in/edqVH5Te) — the difference between AI-assisted shipping and production survivability. AI is genuinely good at knowledge: syntax, patterns, working prototypes in two days. AI cannot give you the instinct to ask “what happens when an aws zone goes down and our control plane lives there?” That instinct comes from someone who got paged at 2am, three years ago, for the same class of bug. Maybe it’s bad luck. And honestly, it could happen to us someday — I won’t pretend we’re immune. That gap — between companies that treat “MVP” and “production-ready” as interchangeable — is going to keep growing. And as AI makes it easier than ever to ship fast, the discipline to ask “but what breaks at 2x load, with one AZ down, on a Friday?” becomes a competitive moat, not a footnote. My job, and the job of our engineering team, is to make sure we don’t get “unlucky” when building systems.

PayPal just said out loud what most legacy tech companies won't admit: they stopped being a tech company. That's not a rebrand. That's a confession. The plan now: → AI-driven automation → Modernizing a bloated tech stack → $1.5B in savings (read: job cuts) Here's the uncomfortable truth. When a company says "we're becoming a tech company again," it usually means two things happened: → They got fat and slow → A competitor scared them awake AI is the story they're telling Wall Street. Restructuring is the story they're telling themselves. What actually matters: can they move fast enough before someone eats their lunch? PayPal has the brand. The users. The data. But incumbents with all three still lose all the time. Turnaround or last act? #PayPal #AI #Fintech #TechStrategy #FutureOfWork

A handful of extra data points. Over $700K in additional annual recovery. That's the actual delta from a recent enrichment project with one of our customers at justt. We encouraged the merchant to provide additional data points most merchants already have sitting in their systems: identity signals, fulfillment details, account history, transaction context, and more. A few months later, the numbers told the story: Overall win rate up 25% Fraud recovery on Stripe: more than doubled Fraud recovery on Chase: nearly tripled Same disputes. Same merchant. A few more data points landing in the right argument. The system continuously adapts. As new data points become available, it dynamically generates and tailors new arguments and evidence combinations for each case to maximize win rates at scale. The lesson isn’t simply “use more data.” It’s that the right data, applied intelligently, turns into real revenue. #Chargebacks #RevenueRecovery #Fintech