Elena Verna Says AI Confidence Theater Is Eroding Trust in Every Brand. She's Right.

Elena Verna, the growth leader at Lovable and one of the most credible growth voices in tech, published a plea last week that a lot of people have been muttering privately for a year. Her ask was simple: stop performing AI mastery you don't have. And it lands harder coming from someone who spends every day inside one of the fastest growing AI companies on the planet.

Her test for the performers is blunt. When someone tells her AI changed their life, she asks them to prove it. What she usually gets back is Slack summaries, email drafts, some scheduled scans. Useful, sure. Life changing, no. The gap between the claim and the demo is the whole story, and she gave that gap a name that is going to stick: AI Confidence Theater.

We read her piece as a growth essay. But strip it down and what she's actually describing is a brand crisis in slow motion. Three of them, stacked on top of each other.

Personal brands built on borrowed proof

Verna makes a sharp observation about how flexing has evolved. The 2020 version was hustle: 5am alarms, cold plunges, inbox zero. The 2026 version is the opposite pose. Nobody works anymore. Agents do everything. The wisest thing you can do is burn tokens and think big thoughts.

Same performance, different costume. And it's a personal branding strategy with a short shelf life, because the proof never arrives. Verna's point is that the token counts and agent screenshots keep showing up while the business outcomes don't. In a market where anyone can generate expert-sounding language about MCP and RAG and vector databases on demand, fluency stopped being a credential. She notes that hiring has already absorbed this lesson: verbal interviews are close to worthless now, and work trials are becoming the only reliable filter.

That's what happens when a signal gets cheap. Everyone can afford it, so it stops signaling. The creators who built audiences on "this INSANE workflow will replace your job" are holding a depreciating asset. The ones who show real receipts, including the failures and the babysitting their agents still require, are building the only kind of personal brand that survives an audit.

Product brands cashing trust they haven't earned

Verna is generous to marketing teams. Their job has always been to sell the best version of the product, and she doesn't fault them for doing it. Her problem is structural: with AI, the distance between the demo and the Tuesday afternoon reality is bigger than it has ever been for any category of software. Traditional SaaS was deterministic. Click the button, get the outcome. AI products are brilliant sometimes and confidently wrong other times, and the landing pages admit none of this.

We've covered this dynamic before, and it keeps ending the same way. Every overpromise gets discovered. Every discovery taxes not just the brand that made the promise but the next AI product the buyer evaluates. Verna frames this as trust-based growth, and she's right that it cuts both ways. Hype is a loan against your brand. The category is collectively over-leveraged.

The honest positioning play is sitting right there, mostly unclaimed. A brand willing to say "this saves your team 20 minutes a day and here's exactly where it still fails" would be so unusual in this market that the candor itself becomes the differentiator. Almost nobody will do it, because almost nobody's board will let them.

Employer brands rotting from the top

The part of Verna's piece that deserves the most attention from leadership is her chain of pressure. VCs were promised miracles, so executives demand miracles, so employees are asked to produce miracles with tools that underperform, so employees learn that performing AI confidence is how you protect your job. The theater isn't an individual failing. It's an incentive structure, and it was installed from the top floor.

Think about what that does inside a company. Your best people know the internal AI wins are being inflated. They watch inflated claims get rewarded. The lesson they absorb is that honesty about what the tools actually do is a career risk. That is a culture problem wearing an innovation costume, and it corrodes the employer brand long before it shows up in a Glassdoor review.

Her prescription for companies is unglamorous and correct. Executives should manage investor expectations down to reality. Managers should anchor AI work to outcomes and give teams actual room to build. Employees should treat learning AI as the job itself, on the clock, not a weekend hobby. None of this makes a good keynote slide. All of it compounds.

The part nobody posts about

The line from her piece that will outlast the discourse is her observation that the first prompt is the fun part and the next thousand are entirely where the value gets created. AI systems drift. Models change, prompts decay, integrations break silently. The real work is the monitoring and the tuning and the maintenance, and that work is invisible on LinkedIn because it doesn't screenshot well.

Which is exactly why it's the credibility opportunity. The market is drowning in first-prompt content. There is almost no thousandth-prompt content. Whoever fills that gap, as a creator, a product, or an employer, is going to own the trust that everyone else is busy spending.

Verna's closing argument is the one we'd tattoo on every AI marketing brief: the technology is already the most impressive thing built in our lifetimes. Pretending it's also magic is not ambition. It's a liability, and the bill is coming due.