September 28, 2026
Outmarket, an insuretech startup that uses AI to automate the paperwork grind for insurance agencies and brokers, just closed a $34.5 million round only months after its last one, according to TechCrunch. The dollar figure isn't the headline. The speed is. Investors don't usually double back that fast unless a company is proving out real usage numbers, not just a good pitch deck. What this tells a business reader is simple: the AI money that's actually sticking around isn't chasing novelty, it's chasing tedium. Insurance paperwork is exactly the kind of unglamorous, high-volume, error-prone process that eats hours from skilled staff every week. That's the same logic behind why so many operations teams are automating helpdesk tickets, HR onboarding, and finance reconciliation instead of waiting for some flashier use case. If you run a business with a paperwork bottleneck, Outmarket's back-to-back rounds are a signal that the market believes narrow, boring automation is where the near-term ROI actually lives -- not in general-purpose assistants.
AI agent startup Instinct raised a $1 billion Series C at a $10 billion valuation, TechCrunch reported, with the company declaring it's 'just getting started.' That's an enormous number for a product still described as viral rather than proven at enterprise scale, and it's worth sitting with why investors are comfortable writing that check. My honest read: this is a momentum bet, not strictly a fundamentals bet. Viral growth curves are seductive, but they don't always translate into the kind of durable, governed deployment that a business actually needs from an agent handling customer data or making decisions on its behalf. We've written before about how quickly agent hype can outrun agent reliability -- the Salesforce Agentforce bug is a recent reminder that bolt-on agents built for speed can leak in ways that cost far more than the round that funded them. None of that means Instinct is destined to repeat those mistakes. But a $10 billion valuation sets an expectation of enterprise-grade trust that viral growth alone doesn't buy. Companies evaluating any agent tool, Instinct included, should ask what governance and audit trail comes with the product, not just what it can do in a demo.
Google is testing purchases from Walmart-owned Flipkart directly through Gemini and AI Mode in India, TechCrunch reported, with the limited pilot covering select products and users ahead of a broader October rollout. This is a small test on paper, but it fits a pattern we've flagged before: the real fight in consumer AI right now isn't which chatbot is smartest, it's who owns the moment of checkout. For merchants, this matters because it changes who the customer relationship belongs to. If Gemini becomes a genuine shopping front door, the business selling the product may end up several steps removed from the person actually buying it -- with less control over pricing display, upsells, and the post-sale relationship. That's a serious strategic question for any retailer building around marketing or CRM systems designed to own the full customer journey. I'd rather see this rolled out slowly and transparently than rushed, precisely because checkout is where trust and money both change hands at once -- get the disclosure and dispute-handling wrong here and it's a much bigger mess than a bad recommendation.
Which of these three would you bet on holding up a year from now: the boring automation, the viral agent, or the AI-run checkout?
Sources