--- title: The AI-Native Team source: https://steadman.ai/newsletters/david/ai-native-team.html published: 2026-04-25 summary: Interactive tool for modelling the Director, AI Builder, Auditor ratio. Diagnoses three gaps in how organisations staff AI work and provides a calculator for the team shape one AI Builder actually needs. --- # The AI-native team Output has multiplied. Checking hasn't. Your best people are stuck doing the wrong jobs. ## Three gaps ### 1. Directors are drowning. AI has made output several times faster than a year ago. Reports, briefs, analyses, models: all arrive quicker. Directors are now reviewing more drafts than they can judge properly. The bottleneck isn't production. It's the queue of work waiting for a senior person to look at it. **Implication:** Take audit responsibility off them. Make sure work is checked before it reaches their desk. And you need more Directors per AI-using team than you did before, not fewer. (See The Director, below.) ### 2. The AI frontier is a full-time job. More than ever, staying at the frontier requires focus. The people doing the best AI work spend all day on it. They rebuild workflows every quarter and know which model fits which task. We should expect everyone to use AI. We can't expect most people to keep up with the frontier. Senior and client-facing staff are worth more on judgement and relationships. The frontier belongs to a specialist. **Implication:** Create a dedicated AI Builder role. Stop asking Directors to stay current on the tools and stop spreading frontier fluency thinly across the whole team. (See The AI Builder, below.) ### 3. The checking is broken. Four failure modes, all common. Senior staff burning expensive hours on verification that sits below their pay grade. AI Builders pulled off the frontier to re-check their own output. Staff who look diligent but skim rather than check, passing errors through. Or worst: nobody checks at all. The work ships because the AI Builder is confident and the Director is busy. Assurance has been treated as a side-task. It's a craft. **Implication:** Make auditing a specialised role, not a side-task bolted onto someone's existing job. The Auditor doesn't stop the work. They let it ship. (See The Auditor, below.) ## Two scenarios ### Everyone uses AI. Nobody specialises. AI inside every existing role. No accountability for any of the three jobs. Everyone directing, building and auditing their own work with varying skill. Nobody breaks through because nobody has time to get genuinely good. ### Director, AI Builder, Auditor Three roles. All use AI. One specialises. The difference is who specialises in what. ## The three roles ### The Director Frames the problem. Decides what good looks like. Edits the work to make it great. Owns the outcome. Uses AI every day and gets real value from it. **Who becomes this?** Your existing senior leader. No change in seniority, new mandate. ### The AI Builder Lives in AI all day, every day. The one seat where staying at the frontier is the job. Knows which model to use for which task, runs sessions in parallel, ships at pace. Practises it as a craft because the tools change every quarter. Wants to be minimally distracted by anything else. **Who becomes this?** Defined by appetite, not rank. Could be a mid-career specialist or a sharp junior. ### The Auditor A different breed. The Director asks 'is this the best version of this?' The Auditor asks 'is any of this wrong?' Same draft, different question, different person. Uses AI every day and uses it well. Also not an expert in the tools. The skill is judgement, accountability, and a name on the file. A machine can fail. Only a person can be accountable for the failure. *A senior generalist reads an AI Builder's draft memo. Traces three citations to source. Two check out. One is a confident fabrication. Reruns the key regression: direction right, magnitude wrong. Stress-tests the argument in a second model. Signs off, or doesn't. Their name goes on the file.* **Who becomes this?** An experienced generalist with a nose for nonsense. Often someone who has done the work themselves. The handoff matters. The AI Builder ships fast and rough. The Director makes it great. The Auditor confirms it's right. If any one of the three tries to do another's job, the model breaks. ## Build your team (interactive calculator) Start with one AI Builder. They're the engine. The question is how many Directors can they serve, and how many Auditors does their output need. The underlying logic: when AI agents make execution free, verification becomes the expensive thing. Ajey Gore named this for software engineering. Martin Fowler amplified it. The same economics apply to knowledge work. ### Inputs 1. **Decisions each Director handles per week** (default: 5). Client briefs, project scoping, strategic calls. The moments where someone frames what good looks like and signs off at the end. 2. **Things the AI Builder ships per day** (default: 5). Reports, analyses, dashboards, research briefs, models, websites. Substantial pieces of work, not emails. 3. **Days to properly audit each output** (default: 1 day). Tracing citations to source, rerunning key numbers independently, stress-testing the argument in a second model. Not skimming. Checking. ### Calculation - One AI Builder ships N things per day (N x 5 per week). - Each Director generates roughly one piece of AI Builder work per decision. - Directors one AI Builder can serve = floor(AI Builder weekly output / decisions per Director per week). - Auditors needed = ceil(AI Builder daily output x days to audit each output). ### Default result (5 decisions/week, 5 things/day, 1 day to audit) - 5 Directors - 1 AI Builder - 5 Auditors - 11 people total **Most organisations have zero Auditors.** ### Benchmark: most organisations today - 5 Directors - 0 AI Builders - 0 Auditors The gap between the benchmark and the calculated ratio is the diagnostic. ## Attribution From [David's Saturday AI Thoughts](https://steadman.ai/newsletters/david/), Edition 10: Rise of the auditors. Published 25th April 2026.