The Agentic AI Market Will Hit $41.8B by 2030 (and your CX budget better be ready)
PLUS: Turn Your Failed AI Projects Into Learning Gold
Start every workday smarter. Spot AI opportunities faster. Become the go-to person on your team for what’s next.
Exclusive Launch Bonuses Offer Ends at Midnight Tonight
Now Available: The Psychology of CX 101: Master the Hidden Drivers of Customer Behavior
Get 7 Exclusive Launch Bonuses
📅 September 22, 2025 | ⏱️ 3-min read
🎯 The big picture
Here's the thing: the game has completely changed. We're not talking about chatbots anymore—we're talking about AI that actually does stuff. And that shift is creating a massive new market that's forcing every CX leader to ask: Am I building for AI that just chats, or AI that solves real problems?
📊 Today's lineup
• The agentic AI market is exploding 175% annually (and you need to know where that money's going)
• An AI startup just raised $150M to build the "plumbing" we all desperately need
• Hertz's AI car scanner is teaching us exactly what not to do (ouch)
• A financial firm is showing how to actually use Google's AI for 1.5 million customers
• GrubMarket's new AI agent proves why industry expertise beats generic solutions every time
1️⃣ The agentic AI market is about to explode (and it's bigger than you think)
What's happening:
• New Omdia research just dropped some wild numbers: the agentic AI market is going from $1.5 billion in 2025 to $41.8 billion by 2030.
• We're talking about a 175% compound annual growth rate—that's nearly double what we saw with regular generative AI at this stage.
• And here's what caught my eye: the top two use cases are automated code development ($8.2B) and virtual customer assistants ($7.7B).
What's so great about this: Look, this isn't just about better chatbots anymore. We're talking about AI that can handle the messy, multi-step stuff—processing a complex refund, rebooking your entire vacation when flights get canceled, or walking someone through a complicated onboarding process. All without ever touching a human agent.
What's next: Time for some real talk with your team. If your 2026 roadmap still just says "improve chatbot," you're already falling behind. Start thinking about what problems your customers face that require multiple steps to solve.
Go deeper: Omdia
2️⃣ This startup just raised $150M to solve AI's biggest infrastructure headache
What's happening:
• Baseten just closed a massive $150 million round to build the high-speed infrastructure that makes complex AI actually work at scale.
• Think of them as the "payment rails for AI"—they want to make sure your customer-facing AI doesn't crash when Black Friday hits.
• Their sweet spot? Fast-growing companies whose AI usage is already breaking traditional enterprise systems.
What's so interesting: Here's the thing nobody talks about: as you move from AI that just answers questions to AI that actually does things, the backend gets incredibly complex. This funding round is basically a giant flashing sign that says "the AI model isn't the hard part anymore—running it reliably for millions of customers is."
What's next: Next time you're in a meeting with your CIO, ask them about your "inference strategy." If they give you a blank stare, you've got a problem that could sink your future CX plans.
Go deeper: PYMNTS
3️⃣ Hertz's AI car scanner is a perfect example of how not to do automated CX
What's happening:
• Hertz thought they were being clever: AI cameras that automatically scan for car damage and bill customers. No humans involved.
• Then customers started getting hit with automated charges—like one consultant who got slapped with a $350 bill—and couldn't easily dispute it through the app.
• The AI is actually pretty accurate (it only finds damage on less than 3% of cars), but the customer experience? Total disaster.
What's so great about this: This is exactly what happens when you optimize for operational efficiency but completely forget about customer trust. Hertz nailed the automation part but created a customer experience nightmare. It's like they forgot that being right isn't enough if your customers feel screwed over.
What's next: Here's your new rule: If your AI can automatically charge someone money, there better be a one-click way to get a human involved. Go audit your automated systems right now—I guarantee you'll find this vulnerability somewhere.
Go deeper: Customer Experience Dive
4️⃣ This financial firm is showing everyone how to actually use Google's AI
What's happening:
• Insignia Financial (a major Australian wealth manager) is using Google Cloud's Vertex AI to completely transform how they serve their 1.5 million members.
• Their goal? Provide the kind of hyper-personalized financial advice that used to be reserved for millionaires.
• Plus, their AI agents are helping human advisors by giving them instant access to exactly the information they need.
What's so interesting: Finally, someone gets it! Instead of just using AI to cut costs, Insignia is using it to scale up the high-touch, advisory experience that customers actually want. They're not replacing the human touch—they're making it available to way more people.
What's next: The bar for "personalization" just got a lot higher. If your idea of personalization is still putting someone's first name in an email, you're about to get lapped by companies doing stuff like this.
Go deeper: Google Cloud
5️⃣ GrubMarket's AI agent proves that industry expertise beats generic solutions every time
What's happening:
• Food supply chain platform GrubMarket just launched a Reporting AI Agent that actually understands their industry's weird jargon and complex data relationships.
• It can handle questions like, "Which restaurant clients ordered two weeks ago but not last week?" and automatically email the results to the right sales manager.
• The secret sauce? It's built on a Food Supply Chain Industry Knowledge Graph that gives it deep contextual understanding.
What's so great about this: This is exactly why the future belongs to specialized AI, not generic solutions. Anyone can build a chatbot, but GrubMarket built an AI that thinks like someone who's worked in food supply chains for years. That's the difference between a novelty and a tool you can't live without.
What's next: Start thinking about your company's unique "knowledge graph." What do you know about your industry, your customers, your processes that your competitors don't? That's where your AI advantage lives.
Go deeper: PR Newswire
⚡ Quick hits
• Workato just launched a massive training program to skill up 3,000+ students in AI → the talent war for AI-ready CX teams is going global, folks.
• GSA and Meta are bringing open-source AI to every U.S. federal agency → when even the government thinks AI adoption is urgent, you know something big is happening.
• Fidelity International just launched "Freya," an AI assistant for 500,000 investors → yet another financial giant betting that AI is the key to keeping customers happy.
💡 CX Prompt Tip of the Day
Turn Your Failed AI Projects Into Learning Gold
Act as a post-mortem analyst reviewing our recent AI implementation that didn't quite hit its goals.
Context: [Tell me about the AI project, what you were trying to achieve, and where it fell short. For example: "We launched an AI to handle returns, but somehow escalations to human agents went up 30%."]
Generate:
1. Three potential root causes for why this didn't work out.
2. Two early warning signs we probably missed along the way.
3. One specific change to our AI governance process that could prevent this next time.
4. A one-sentence lesson for leadership that they'll actually remember.
Keep it focused on actionable process improvements, not pointing fingers.
Quick win: Try this on your last AI project that didn't quite live up to the hype. Here's the thing: your failures often teach you way more than your successes, but only if you take the time to really dig into what went wrong.
🤔 CX reflection
Question of the day: If AI agents can now handle complex, multi-step customer problems from start to finish, what's the single most important skill you should start training your human agents on right now?
See you tomorrow!
Mark
Follow me on Linkedin