Aakash GuptaThe Claude Workflow Nobody at the VP Level Is Showing You
At a glance
WHAT IT’S REALLY ABOUT
A VP-level Claude workflow for strategic decks, not PRDs
- The core leadership skill for effective AI use is rigorous problem decomposition, because LLMs will flatten complexity and generate “slop” if the problem isn’t broken into crisp parts.
- Matt’s two-hour all-hands workflow starts by inventorying raw materials (strategy docs, prior presentations, Zoom transcripts) and then using Claude to pivot that raw material into the company’s strategic themes.
- He builds slides before writing the talk track, then feeds screenshots of the finished slides back into the same Claude session to produce a non-redundant narrative that leverages accumulated context.
- Claude often behaves like an eager junior employee—rushing to deliverables, inventing unsuitable formats, or misreading audience politics—so Matt slows it down with iterative context, explicit constraints, and staged revelation of the true domain.
- Beyond Claude, his weekly AI stack includes Slack-based agents for data querying (Snowflake access), a “product scanner” that flags threads needing PM attention, and “Chiefy,” which audits new docs against a corpus of gold-standard company artifacts to detect inconsistencies and staleness.
IDEAS WORTH REMEMBERING
5 ideasTreat AI use as a decomposition test, not a prompting contest.
Matt argues leaders create value by exploding a messy business problem into distinct parts before asking an LLM to transform or assemble anything. If you don’t, the model simplifies the space and produces outputs that executives quickly dismiss as noise.
Start with an inventory of raw materials before opening Claude.
Instead of “make me an all-hands deck,” he lists available sources—demo-day Zoom recordings, transcripts, strategy themes, prior artifacts—then decides which pieces should be transformed and which should remain human-owned (narrative intent, audience framing).
Use Claude for strategic reshaping: pivot raw updates into the strategy’s categories.
He combines a Zoom transcript (engineering-centric) with the strategy doc (theme-centric) and asks Claude to reorganize presentations by investment theme—his “matrix multiplication” idea—so the content matches the executive narrative shape.
Build slides first, then generate the talk track from slide screenshots.
By finishing the visual “show” layer first, he can ask Claude to write speaker notes that add context rather than repeat bullets. Feeding screenshots back into the same session forces the model to align to what’s actually on screen.
Stop the model from racing ahead by explicitly controlling the next step.
Claude will try to jump to premature deliverables (e.g., drafting a full Word strategy doc) without clarifying questions. Matt counteracts this by drip-feeding context, issuing “don’t recommend next steps” instructions, and preferring long iterative sessions over revising a rushed first draft.
WORDS WORTH SAVING
5 quotesAI for leaders is ultimately a test: how good are you at decomposing problems? AI is very good at solving a problem, but it will simplify the problem space if you don't properly decompose it.
— Matt Wensing
A lot of junior employees, which I would consider Claude one of, very talented but very junior, is very eager to please, and because of that eagerness, it will go too far too fast.
— Matt Wensing
It's better to have that 50, 100, even 200 iteration session with a great deliverable at the end than to say yes to that first ask to generate that deliverable and then try to revise it.
— Matt Wensing
As a leader, you don't wanna microwave your output, right? If that's all you're doing, that's low value. You wanna really slow cook these things more often and produce something that is really, you know, to produce something that's really compelling, impactful, resonates with your audience, you've gotta slow it down.
— Matt Wensing
If you use it to generate alignment, I think executives are the best at filtering out noise and detecting BS and detecting slop.
— Matt Wensing
High quality AI-generated summary created from speaker-labeled transcript.
