Agent-Ready with CXO.dev
By Claire Vo + Zach Davis
Categories: technology
What AI transformation actually looks like, from inside the companies and teams doing it. Every Friday, Claire Vo (host of How I AI) and her CXO.dev co-founder Zach Davis share what they’re seeing as they help technical leaders build agent-ready teams: how the code and the org are changing, which tools and releases matter, and what’s coming next on the frontier. Each episode runs under 10 minutes and ends with one thing to try this week. Written by humans, voiced by AI.
26 episodes indexed on PodSearch.
Recent episodes
- Verification is the New Review: How to Write Code That Actually Works
AI-generated code can't be trusted, say the skeptics, and they're not totally wrong. We make the case that there's no such thing as code you can't trust, only code that hasn't been verified yet, and walk through the layers of verification it takes to ship AI code with confidence.Read the note…
- Tokenmaxx or Token Cap? Controlling the cost of AI
Everyone wants to control AI spend. Hard caps make the budget predictable, but they also cap your return and punish your best builders. We make the case for visibility, education and soft caps instead, plus what Apple, Accenture and Cloudflare are doing.Read the note…
- Feel Behind on AI? Start With One Small Experiment
Everyone feels behind, from the X timeline to the big-company CEO still working on a strategy. We make the case that the way out isn't panic or caution, it's a small experiment with how you operate, plus what we're reading on the most AI-pilled company in the world, Mark Cuban on AI, and where the…
- Still using PRs? What to do with code review
Code review was built for humans reading every line, and agents just broke that assumption. We make the case for rebuilding review around evidence of correctness and less human toil, plus Will Larson's small-change, Thorsten Ball on pull requests, and the paper Zach co-wrote on agentic code…
- Software Factories: What They Are and Why They're Not Scary
Fifteen years ago, calling your engineering team a software factory was an insult; now everyone wants one. We explain what a software factory actually is and why it's not that scary, plus software factories for skeptics, Warp's CEO on rolling them out, and a reference architecture for a secure…
- What Makes a Company AI-Native? A Working Definition
Someone asked us to define AI-native on the spot, so we did, then checked our answer. We walk through a working definition of an AI-native company and the part most companies skip, plus Atlassian's AI SDLC playbook, Ethan Mollick on the dot and the swarm, and how agent-ready your codebase is.Read…
- AI Isn't About Speed: Set Your Phasers to Ambition
Most people think AI is about going faster, but speed is the least interesting thing it buys you. We make the case for setting your phasers to ambition, plus the CEO of Linear on getting better along with our tools, a writing tool built in a weekend, and agents that monitor production deploys.Read…
- Cheaper Models, Bigger Bets: Your Ambitious Projects Are Now in Play
Frontier intelligence just got a lot cheaper, and a new kind of model, Jev, makes decisions instead of text. We make the case that your most ambitious projects are now in play, and that the bottleneck is people, not technology, plus Opus 5.5 versus GPT-6 Sol and a level-headed Jev explainer.Read…
- Rolled Out AI Tools? Why Adoption Stalls Without Process Change
Most companies handed their teams AI tools and called it done. We make the case that real change in means real change out, and AI adoption without process reengineering stalls, plus Warp's COO on running non-engineering teams like engineering teams and inside OpenAI's agentic software factory.Read…
- Put AI in the Group Chat: Why Multiplayer AI Beats Solo
Working with AI has quietly become a one-player game, even though teams working with AI beat individuals working with AI. We make the case for putting AI in the group chat with cloud agents and Slack, plus the multiplayer AI manifesto and why Shopify is going back to native.Read the note…
- Measuring AI Impact: You Can't Improve What You Can't See
Uber's agent requests grew more than nine times this year while its AI spend went flat. We make the case that you can't improve what you can't see, and that includes your software factory, plus self-improving cloud factories, running a factory at Uber scale, and why Ramp built its own coding…
- Stalled on AI? What Holds Companies Back Early
After four months inside a handful of companies, we keep seeing the same things hold them back: procurement, codebases that aren't ready for agents, and executives who aren't walking the walk. We share what stalls AI transformation early, plus Anthropic's AI-native development playbook and a…
- Things Are Getting Interesting: Agents With Actual Jobs
Slack, Block and Grok all shipped agents with actual jobs this week. We dig into multiplayer agents, agents hired for a job, and the meta-harness, plus Grok Bot, Grok 4.6 and Cursor Origin, Slack Code, and Warp Factories.Read the note…
- Building a Company Brain? Why It Needs a Librarian
Your company brain can't tell which of three roadmaps is the real one. We make the case that every company brain needs a librarian, plus why we should stop calling it a brain, what Spotify learned scaling AI coding agents, and the top reasons AI stalls inside companies.Read the note…
- Glimpses of the Future: Why You Can't Stop Experimenting With AI
A year ago, merging code with no human looking was unthinkable; now Intercom's bot does it with one rule. We share glimpses of the future and why you can't afford to stop learning and experimenting, plus Steve Yegge on what's coming, Amp's case for Orbs, and how Intercom taught AI to review…
- Build or Buy? Choosing Your Tools in the Age of AI
A QA engineer vibe coded a replacement for Jira, and a few months later the team went back to Linear. We dig into build versus buy in the age of AI, plus why you probably don't need your own coding agent and how Anthropic secures its AI-native development lifecycle.Read the note…
- Don't Make Quality a Tradeoff: How AI Can Make Software Better
AI doesn't have to make your software worse; it can make it better. We make the case that quality shouldn't be a tradeoff, plus what makes quality software, Gumroad's support agent in its own words, and how Sierra built its MCP gateway.Read the note: https://www.cxo.dev/blog/qualityChapters0:00…
- Let AI Review Your Code: Humans Were Never That Good, Anyway
Intercom let AI approve some pull requests, and deployments doubled while downtime from bad changes fell. We make the case that human review was never that good, plus how Sierra built one agent for the whole company and the strange case of isRecord.Read the note…
- Automating With Agents? Automate the Workflow, Not the Task
Uber gets most of its pull requests from agents, and the lesson for the rest of the company isn't more agents for more tasks. We make the case that most of us are automating the wrong noun, plus Uber's agentic pods, Figma's 2026 AI report, and the future of product design.Read the note…
- Raise Your Ceiling, Raise Your Floor: Closing the AI Super-User Gap
About thirty percent of one big software company's people are AI super-users, and most leaders can't see the other seventy. We make the case for raising your ceiling and your floor, plus agentic autonomy levels, how Coinbase keeps AI spend flat while usage grows, and Atlassian and Dropbox on AI…
- Benchmarking Against the 2%? What Real AI Adoption Looks Like
The headlines say everyone runs fleets of autonomous agents, but Notion found only two percent do. We make the case that what separates the two percent is mostly unglamorous work you can start on Monday, plus Notion's great renovation report, Block's BuilderBot, and why Warp's CEO says we're…
- Is Code Review Your Bottleneck? Why Confidence Matters More
Ramp reads every change that touches money line by line, and treats the rest like a black box. We make the case that code diffs aren't the bottleneck, confidence is, plus ten thoughts on reviewing and verifying code, agentic code review audited, and a junior engineer's perspective.Read the note…
- Code Is Free: Why Judgement Is Your Real Edge
Coding is no longer the bottleneck; judgement is. We make the case that when code is basically free, your edge is judgement and fast feedback, plus Slack's former head of design on prototyping, modern engineering values, and Shopify's internal hosting tool, Quick.Read the note…
- Past the Basics: What Comes After AI Adoption
Uber burned through its AI coding budget for the year in four months, and the work got better. We dig into what comes after adoption, when the question becomes what the spend actually bought, plus DORA on the era of tokenmaxxing, how Anthropic uses skills, and how OpenAI built its data agent.Read…
- AI Tools Have Grown Up: Is Your Org Keeping Pace?
One company put a single login in front of every AI tool, and agent use jumped to ninety-five percent with the same models. We make the case that the tools have grown up, which moves the hard part onto your org, plus Cursor's developer habits report and a no-hype take on Claude Opus 4.8.Read the…
PodSearch is a free, independent podcast search engine. Read about PodSearch, the privacy policy or contact us.