Add blog post: AI-first program management

dictionary.txt

@@ -836,9 +836,17 @@ workspace
 # AI / Agentic workflows
 agentic
 andon
+async
 autocomplete
 boilerplate
+boundaryless
 Copilot
+differentiator
 handoff
+LLM
+LLMs
 PRs
-scaffolding
+scaffolding
+standups
+trackable
+workstreams

src/content/posts/2026-03-19-ai-first-program-management.md

@@ -0,0 +1,118 @@
+---
+title: "AI-first program management: amplifying judgment, not replacing it"
+description: "AI-augmented program management is the natural evolution of async-first and engineering-inspired workflows — amplifying human judgment, not replacing it."
+---
+
+When I first wrote about [the nine things a program manager does](/2021/03/26/nine-things-a-technical-program-manager-does/), the hardest part of the job was the Monday morning scramble: six programs, four teams, three time zones, a dozen threads that moved over the weekend, two risks that materialized, and someone who quietly changed the scope of a critical deliverable in a comment buried three levels deep. All of it needed synthesizing into a coherent picture before the 10 AM leadership sync.
+
+That scramble hasn't gone away. But increasingly, I'm not doing it alone.
+
+## From async to AI-augmented
+
+The evolution makes sense in hindsight. I've argued that teams should [default to async communication](/2022/03/17/why-async/) — written, durable, discoverable artifacts over synchronous [meetings](/2023/04/20/meetings-are-a-point-of-escalation/). That managers should [use the same tools engineers use](/2023/01/10/manage-like-an-engineer/) — issues, pull requests, project boards — to plan and track their own work. And most recently, that [AI agents are extending the same patterns](/2026/03/18/agentic-workflows/) of transparency and code review that made open source successful.
+
+These ideas aren't disconnected. They're an evolution.
+
+If the shift from synchronous to async was about decoupling communication from presence, and managing like an engineer was about applying developer workflows to leadership, then AI-first program management is the next logical step: using AI to amplify the judgment, pattern recognition, and relationship work that makes program managers effective.
+
+The through-line is the same principle I've been writing about for years: [make work visible](/2022/02/16/leaders-show-their-work/), make it durable, and reduce the friction between having an idea and acting on it. AI doesn't change that philosophy. It accelerates it.
+
+## AI across the PM toolkit
+
+Here's how each core PM responsibility shifts when AI enters the picture — and where it doesn't.
+
+### Communication, coordination, and facilitation
+
+Program managers are professional context-switchers. You spend your day translating between engineering teams, product managers, designers, and executives — each with different mental models and vocabulary.[^1] It's an O(n²) communication problem, and it only gets worse as the organization grows.
+
+AI doesn't eliminate that complexity, but it compresses the information-shuttling work. An LLM can summarize a 200-message thread into a paragraph. It can draft a status update that translates "we're blocked on a schema migration that requires a backward-compatible rollout strategy" into language an executive actually cares about. It can take meeting notes and extract the three things that matter from the twenty that were discussed.
+
+I've [written about how we communicate at GitHub](/2023/10/04/how-to-communicate-like-a-github-engineer/) — optimizing for clarity, discoverability, and low-context readers. AI handles the mechanical work of drafting and reformatting for different audiences. The editorial choices — *what* to communicate, what to emphasize, when to pick up the phone instead — stay with you.
+
+### Capture and track work
+
+One of a PM's most underappreciated responsibilities is ensuring that work doesn't fall through the cracks. Every conversation, decision, and commitment needs to land somewhere trackable — usually an issue or a project board.
+
+AI can auto-generate issues from meeting transcripts. It can scan Slack threads and surface commitments that never made it into a tracker. It can flag stale issues that haven't been updated in weeks, or detect when a project board is drifting from reality. Think of it as a continuous reconciliation process — a linter for your program's state, catching drift between what people *said* they'd do and what the artifacts actually reflect.
+
+This matters because the biggest risk in program management isn't that something goes wrong. It's that something goes wrong *and nobody notices* until it's too late.
+
+### Risk identification and mitigation
+
+Speaking of risk — PMs are supposed to see around corners. In practice, that means reading a lot of threads, attending a lot of standups, and developing an intuition for when something *feels* off.
+
+AI augments that intuition with pattern recognition at scale. It can analyze velocity trends across repositories, flag pull requests that have been open too long, identify dependencies that no human mapped, and surface cross-project risks that are invisible when you're looking at one team's board in isolation. I've called [transparent collaboration the andon of knowledge work](/2023/08/30/transparency-collaboration-is-the-andon-of-knowledge-production/) — pull the cord when something's off. AI is an andon that monitors every cord at once, catching a staffing conflict between two programs before either PM realizes they're competing for the same engineer's time.
+
+AI won't replace the experienced PM's gut feeling that "this one's going to slip." But it surfaces the signals earlier, giving you more time to act. The best risk management has always been about buying time.
+
+### Reporting up and across
+
+Let's be honest: nobody became a program manager because they love writing weekly status reports. And yet, clear upward reporting is one of the highest-leverage things a PM does. It's how leadership knows where to pay attention and where to stay out of the way.
+
+AI can draft these reports from project activity — commits, merged PRs, issue closures, comment threads. Instead of spending Friday afternoon manually assembling a narrative from memory and half-updated boards, you start with an AI-generated draft that reflects what actually happened, then layer on your interpretation and recommendations.
+
+The key word there is *start*. An AI-generated status report is a first draft, not a finished product. The PM's job is to add the "so what" — the strategic interpretation that turns a list of activities into insight. AI gives you the *what*. You provide the *why it matters* and the *what to do about it*.
+
+### Relationship management
+
+Program management is fundamentally a relationship business. When I first moved into the TPM role, I remember watching senior PMs set up "just wanted to say hello" meetings and dismissing it as socializing on company time — until I realized that's exactly the point.[^2] You make regular deposits to the social capital bank long before you need a withdrawal. AI doesn't replace that. You can't automate trust.
+
+But AI can help you *maintain* relationships at scale. It can remind you that you haven't checked in with a particular stakeholder in two weeks. It can surface context before a meeting — "last time you spoke with this team, they raised concerns about the timeline for the API migration." It can help you draft a thoughtful response to a tricky message when you're running low on cognitive bandwidth at 4 PM on a Thursday.
+
+The human work — building rapport, reading a room, knowing when someone's frustrated versus when they're genuinely blocked — stays firmly in the PM's domain. AI just helps you show up more prepared and more responsive than you could be on your own.
+
+### Consensus and conflict resolution
+
+Driving consensus across teams with competing priorities is among the hardest things a PM does. It requires understanding not just each team's *position*, but their *interests* — the underlying needs that drive those positions.
+
+AI can help by synthesizing different viewpoints, drafting proposals that incorporate multiple perspectives, and suggesting compromises based on stated constraints. Need an RFC that reflects six teams' input? AI can produce a first draft that no single person could assemble manually — but you still need to facilitate the conversation that turns a draft into a decision.
+
+Where AI falls short is in reading the political dynamics. Understanding that Team A's objection is really about being burned in a previous launch, or that a particular VP's silence means something different than a junior engineer's silence — that's pattern recognition of a deeply human kind. No model is going to learn the org chart's shadow topology from a prompt.
+
+### Boundaryless engagement
+
+One of the program manager's superpowers is the ability to engage across boundaries — teams, organizations, time zones — without being constrained by reporting lines. The cost of this superpower is context-switching. You might go from a security review to a design critique to a vendor negotiation in the span of an hour.
+
+AI dramatically lowers the cost of those context switches. It can brief you on a project you haven't touched in a week. It can summarize what happened in a channel while you were focused elsewhere. It can help you draft a message calibrated for an audience you don't interact with daily.
+
+If [async communication](/2022/03/17/why-async/) decoupled work from time zones, AI decouples engagement from the cognitive overhead of maintaining context across a dozen concurrent workstreams. It's a working memory upgrade for your entire portfolio.
+
+### The long tail of program work
+
+Every PM knows the catch-all responsibility: do whatever needs to be done to keep the program moving. Sometimes that means writing a missing spec. Sometimes it means scheduling a meeting nobody wants to own. Sometimes it means reformatting a spreadsheet at 9 PM because a VP asked for a different view of the data.
+
+AI is exceptional at this long tail of tasks. Drafting documents, cleaning up data, formatting reports, writing follow-up emails, creating meeting agendas from prior action items — all the mundane-but-necessary work that eats a surprising percentage of a PM's week. Offloading these tasks means more time for the work that actually requires your judgment, empathy, and institutional knowledge.
+
+## What doesn't change
+
+It would be easy to read all of this and conclude that AI is about to make program managers obsolete. The opposite is true.
+
+The responsibilities haven't changed. Communication, risk management, relationship building, consensus-driving — these are still the job. What's shifted is the ratio of *information processing* to *judgment* in a PM's day. AI handles more of the former so you can focus more on the latter.
+
+The things AI can't do are precisely the things that make great PMs great: reading a room, building trust over months, knowing when to push and when to back off, navigating organizational politics, and making the hard call when the data is ambiguous. These are human skills, and they become *more* valuable as AI handles the routine work, not less. When everyone has access to the same AI tools, the differentiator is the human wielding them.
+
+## What changes for PMs
+
+That said, AI-first program management does require new muscles — or at least, new applications of existing ones. Four capabilities stand out:
+
+**Prompt craft matters.** The quality of what AI produces depends entirely on the quality of what you ask for. PMs who can write clear, specific prompts — essentially, good requirements — will get dramatically better results than those who can't. It turns out that years of writing crisp issue descriptions and well-defined acceptance criteria is excellent training for working with LLMs. Good requirements have always been a PM superpower. Now they're a superpower twice over.
+
+**Knowing when *not* to use AI is as important as knowing when to use it.** A sensitive personnel conversation, a politically charged escalation, a message to a stakeholder who's having a rough week — these require a human touch that AI can't fake. Developing judgment about when AI helps versus when it gets in the way is a skill in itself, and it's one that's hard to teach in the abstract.
+
+**Verification becomes a core competency.** AI will confidently summarize a thread and miss the most important nuance. It'll draft a status report that's 90% accurate and 10% dangerously misleading. PMs need to be skilled editors and fact-checkers, not just consumers of AI output. The role shifts from *author* to *editor-in-chief* — you set direction, review drafts, and ensure quality before anything ships.
+
+**AI fluency is table stakes.** Just as PMs were expected to be comfortable with project management tools, version control, and whatever collaboration platform their teams use, they'll be expected to work fluently with AI assistants. Not as a novelty, but as a core part of the daily workflow — the way we already think about Slack or email or GitHub Issues.
+
+## The PM as orchestra conductor
+
+The mental model I keep coming back to is the program manager as orchestra conductor. A conductor doesn't play every instrument — they don't play *any* instrument during the performance. But they're essential: they set the tempo, bring in the right sections at the right time, interpret the score, and turn individual performances into a coherent whole.
+
+AI agents are new instruments in the orchestra. They play fast, they play consistently, and they can handle parts that used to require a human musician. But someone still needs to read the score, understand the audience, and make the hundred small decisions that turn a technically correct performance into something that actually moves people. That's the PM's job — and it always will be.
+
+The shift from synchronous to async made program management more intentional. Managing like an engineer made it more transparent. AI makes it more leveraged. Each evolution builds on the last, and each one makes the human judgment at the center of the work more important, not less.
+
+If you're a PM wondering where to start, pick the task that consumes the most time but requires the least judgment — status report assembly, meeting note cleanup, stakeholder update drafts — and hand it to an AI. You'll free up hours for the work that actually drew you to the role: solving hard problems with smart people across organizational boundaries.
+
+[^1]: I once counted the number of distinct audiences I communicated with in a single week as a TPM. The answer was genuinely depressing. Every additional team or stakeholder doesn't just add one more communication channel — it adds one per *existing* channel. This is why PMs' calendars look the way they do.
+
+[^2]: As much as I might like them to be, [human-to-human requests are unlike server-to-server requests](/2021/03/26/nine-things-a-technical-program-manager-does/). A properly authenticated request from a never-before-seen client is less likely to be fulfilled, or fulfilled in a timely manner, even if it's facially valid. Invest in the relationship before you need the favor.