While Rome Burns, We Orchestrate Agents
This is a rant, don’t read it.
Hype and FOMO are eating our organisations alive. Every feature team needs to become AI-native to have the right to exist.
What the hell does AI-native even mean?
Definitions I have seen circle around orchestrations of swarms of autonomous agents and magic workflows. None has been about product focus.
Meanwhile management is still impressed by one- or two-person teams showing greenfield AI PoCs or peripheral tooling automation. What actual product has been shipped? Which enterprise application got better? How many customers have been won?
In all seriousness, we allow, even encourage, teams to tell their Product Owners: “Sorry, we can’t deliver features right now. We’re busy figuring out how to become AI-native.”
Are you kidding me?
There was a time when improving your tools, skills, and productivity was simply part of being a professional engineer. You didn’t ask permission. You didn’t stop delivering. You didn’t make your personal learning curve the company’s problem.
You figured it out and got better at your damn job.
Now we’re celebrating technology playgrounds and forgetting our actual customer commitments. Developers build pet projects, management applauds flashy demos, and everyone congratulates themselves on the great AI transformation.
Zero features. Zero customer value. But congratulations, AI-native!
We are in brutal global competition, and nobody gives a damn about our agent orchestration, our million-token workflows, or the twenty AI agents that magically turn a Jira ticket into a pull request nobody understands. Customers care about products that solve problems. Performance. Reliability.
Give talented, motivated teams clear outcomes. Let them choose their tools, experiment and automate as much as they like. Then hold them accountable for the customer value they deliver.
AI can be a competitive advantage. It is still a means to an end. Stop rewarding experiments as if they were outcomes.
And for God’s sake, GET BACK TO SHIPPING!