Trent Gillespie
February 19, 2026
I often start my keynote talks with Bon Jovi. Not Livin' on a Prayer — although that may be apropos for the moment. The Pitbull collaboration: Now or Never.
I use it because I believe we are at an inflection point where the decisions leaders make now will dictate how well their companies succeed over the coming years. How their careers will be impacted. And what goes unsaid: what their future financial security looks like.
I've been saying that for three years. I work hard to help leaders understand what is happening and prepare their organizations for it — so they and their employees have jobs in the disruption that is coming.
I've been told my presentations might be terrifying. That they might be too advanced. At the same time, audiences rate me a consistent 4.8 out of 5 stars, with comments like "Finally, a useful AI keynote."
So I must be hitting something right.
But here's the thing: there is only so much I can share in a 50-minute keynote or a weekly newsletter. Often, that is limited by what leaders can safely view from their current observation point. I don't tell them the rest.
This week, I want to share the wider view. Because I think we just crossed a line — and I can't get it out of my head.
We invented the machine of innovation. And that changes everything.
The good news: it's still early. Most businesses haven't grasped what's happening yet, which means the leaders who start preparing now have a real window to get ahead. That window is closing fast. But it's open.
When ChatGPT launched three years ago, I wrote a whitepaper sharing a little-known concept called the invention of the machine of innovation — the IMI. The core idea: the primary job of most businesses is to innovate. Continually improve products, create value for customers, generate growth. But innovation is also what's hardest. The people who know how to do it are expensive. Technology investment is capital-intensive. It fails about 90% of the time.
The IMI thesis is this: when machines can innovate, when AI can do that work for pennies on the dollar in hours instead of years, it doesn't just change efficiency. It changes how value is created. How value is destroyed. What competition looks like. What business itself is.
Until now, it was theoretical. It isn't anymore.
Look at just the last few weeks:
Clawdbot went viral: a personal AI agent built by one developer that does things on your behalf, managing your calendar, sending your messages, running your errands across every app you use.
Anthropic also released a new version of Claude Code and was immediately one-upped by OpenAI's Codex: AI tools that write, debug, and optimize entire software projects automatically, handling work that would take human teams months. An Anthropic developer shared publicly that Claude wrote almost the entire Claude Cowork codebase itself.
Anthropic released Claude Sonnet 4.6 and Opus 4.6, its most capable models yet, plus launched Agent Teams: multiple Claude instances coordinating in parallel, each owning its own piece of a project, like a real team. OpenAI is rumored to have its own major model release imminent. The race is accelerating weekly, not quarterly.
Anthropic launched Claude Cowork on both Mac and Windows: AI that lives on your desktop and works alongside you locally, inside your actual files, apps, and workflows. Not a browser tab you switch to. A colleague that's already there.
OpenAI launched Frontier: an enterprise platform for managing AI agents the way you manage human employees, with onboarding, permissions, institutional knowledge, and performance feedback. One early customer reduced a six-week manufacturing task to a single day. Another freed up 90% of their sales team's time. OpenAI's own framing: these are AI coworkers, not tools.
"Gas Town" launched as an open-source “software factory” modeled after a human village—running dozens of AI software developer agents in parallel, building entire software products from a single prompt. One person can now run what used to require a development team. It’s capabilities scared Anthropic.
I spoke with the CEO of an AI startup this week. He told me the hardest thing for developers right now is getting over the fact that they no longer look at code — at all. They describe what they want. The machine builds it. Their job is to supervise the output and think of test cases to ensure it works right.
And the people at the top of this industry aren't being quiet about where that leads. Mustafa Suleyman, CEO of Microsoft AI, told the Financial Times last week that most white-collar tasks, accounting, legal, marketing, project management, "anyone sitting down at a computer," will be fully automated within 12 to 18 months.
These aren't predictions from outside observers. These are the people building the tools.
This is what the IMI actually looks like. Not AI helping humans work faster. AI innovating on itself. Software that writes better software. Agents that spawn and improve other agents. The feedback loop is now closed — and it only accelerates from here.
Most leaders have spent three years asking "how do we use AI to be more efficient?" That's always been the wrong question. Now it's just crystal clear.
Markets are imperfect, but they're the best real-time prediction machine we have. And over the last two weeks, they made a prediction about your industry that you should take seriously.
It started with software stocks. The S&P North American software index dropped roughly 15% in January — its worst monthly decline since 2008. Salesforce is down nearly 26% year to date. ServiceNow has lost 28%. Intuit is down more than 34%. Nearly $1 trillion in market value has been erased from the software sector in six weeks.
Then it spread. Sector by sector, day by day:
Wealth management: Charles Schwab fell 7.4%, Raymond James dropped 8.3%, and LPL Financial lost 8.8% in a single session after a startup launched an AI tax planning tool. Bloomberg Intelligence called it "fee compression concerns and potential market-share shifts." One analyst said it best: "I think at the end of the day people just want to trust their money with somebody, a person." That's a clue — we'll come back to it.
Trucking and logistics: A former karaoke machine company — with a $4.8 million market cap — announced an AI freight platform claiming it could scale shipment volumes 300–400% without adding headcount. CH Robinson fell 15%. RXO dropped 20%. Landstar lost 16%. The entire Russell 3000 Trucking Index dropped 6.6% in a single day.
Real estate services: CBRE dropped 12%, Jones Lang LaSalle fell 12.5%, Cushman & Wakefield lost nearly 14% over two days. The analyst note from Keefe, Bruyette & Woods said it plainly: "Investors are rotating out of high-fee, labor-intensive business models viewed as potentially vulnerable to AI-driven disruption."
Insurance brokers. Legal services. Each one hit in turn.
A Jefferies strategist described it as "shoot first, ask questions later." A Goldman Sachs analyst compared the outlook for legacy software to newspapers after the internet arrived — a sector that declined an average of 95% between 2002 and 2009 and never fully recovered.
Are these markets overreacting? Probably, in the short term. But when institutional money moves this fast and this broadly, it isn't noise. It's a signal.
The signal is this: any industry built on high fees, labor-intensive delivery, or expertise as a moat just had its business model put on notice.
I have a CEO who built her own accounting system and wants to replace QuickBooks. A CFO who programmed a DocuSign replacement — and now gives it to clients. At Stellis AI, we're rebuilding CRM and sales processes that have never worked well. I was just told we should rewrite HubSpot. Plenty of startups are already doing it.
The machine of innovation doesn't just threaten legacy industries. It's already replacing them, one workflow at a time.
Everyone I talk to is focused on what AI can do for their business today — specifically, how to make their employees use it safely.
Barely anyone is asking the more important question: how to use AI to make their employees' and business's futures safe instead.
There's a big difference. One is about compliance and caution. The other is about survival and opportunity. And most organizations are so focused on the first that they're missing the second entirely.
Here's why I tell every leader I work with that daily AI habits matter more than any single tool or initiative: the disruption isn't going to hit your company as a single event. It's going to arrive through your competitors getting faster, your customers expecting more, and new entrants doing in weeks what used to take your team months.
The only defense is an organization that knows how to think and work with AI, not just use it occasionally for a productivity bump.
That means your employees need to be building these skills now, every day, through experimentation and habit. Because when the pressure arrives, you can't train your way out of it in a sprint. You need a team that's already there.
Who else are you going to lean on? These are the people who know your business, your customers, your processes. Getting them ready isn't an HR initiative. It's your most important strategic investment right now.
This is also where my Future Customer™ framework comes in. I developed it from Amazon's "Working Backwards" methodology, the discipline of starting with your future customer's needs and building back toward what you need to create today. The difference now is that "future" has a much shorter timeline than it used to.
Five questions — sit with them this week:
#1. What do your customers say they value most today?
Not what you think they value. What they actually tell you.
#2. Which of those values will still matter when know-how is free and instant?
What will your customers still demand when they can get an expert answer in seconds?
#3. Which of those values can you no longer charge extra for — because it becomes table stakes?
Once smart answers are everywhere, what stops being a premium service?
#4. Where will your customers still want a real person they can trust?
The wealth management analyst who called the selloff "completely overblown" made a telling point: "People just want to trust their money with somebody, a person." She might be right. Where is that true in your business? And where will AI actually make that human trust stronger — not weaker? Define it.
#5. What can you offer now that wasn't possible before?
This is where the opportunity lives. The first four questions show what's at risk. This one is how you build what comes next.
Next week I’ll be working with leaders in an industry expected to lose at least 400,000 jobs due to AI automation, to help plan their AI strategy. We’ll have one goal: create more jobs with AI than are lost. We’ll use the Future Customer™ questions to figure it out. You can do this on your own with my Future Customer™ GPT.
Earlier I said I don't always share the full picture in a keynote or newsletter. I'm changing that. I will be creating a curated group of leaders who are bold and want to be at the forefront of not just using AI, but building companies and careers for what comes next.
This won’t be a learn to use AI course: it’s research, application, and sharing. Training content, live events, office hours, tools, access to experts. Interested? Let me know with this one-click link.
Want to learn more about the Stellis Leadership Lab? |
The companies that come out ahead won't necessarily be the ones who own the AI infrastructure. They'll be the ones who learn to think and innovate with AI faster than their competitors.
That's a skill. And like every skill, it compounds with practice.
1. Run the Future Customer Framework™ before your competitors do.
Set aside two hours this month. Take your leadership team through those five questions. You don't need perfect answers — you need the conversation. The businesses that will struggle are the ones that wait for certainty before asking the hard questions.
2. Identify your trust anchor.
The wealth management selloff came with a telling counterpoint: people still want to trust their money with a person. Where is your trust anchor? What do your customers value that requires a human relationship, human judgment, or human accountability? That's your moat — protect and build it.
3. Start an AI Sprint this month.
Not because AI is going to replace your team tomorrow. Because the learning gap between companies that experiment consistently and companies that watch from the sidelines is already widening — and it accelerates from here. The sprint is simple: what did you use AI for this month, what changed because of it, and what will you test next month? Start there.
For three years, AI has been a productivity tool. Something to experiment with. Something to figure out later.
That just changed. The machine of innovation is here — and it doesn't wait for leaders who are still deciding whether to take it seriously.
You don’t have to have all the answers. But you do need to choose: will you lead?
It’s your Now or Never moment.
Trent Gillespie is CEO of Stellis AI and a keynote speaker helping business leaders understand and operationalize AI in their companies. He spent almost nine years leading global innovation efforts at Amazon before leaving to help other companies build the capabilities they need to compete. Book Trent to speak to your group or book a call to discuss using AI within your business.
Tell us what you thought of today's email. |
Did someone forward this newsletter to you? If you're not already signed up, you can subscribe to AI SPRINT™ for free here.