# Strategy Before Software > [!question] Lingering Questions > - When AI removes the engineering-hours constraint entirely, what replaces it as the forcing function for prioritization discipline besides self-imposed mission filters? > - Does a "good idea backlog" of AI-generated, already-built PRs actually get reviewed and shipped more slowly than a pre-AI backlog of proposed-but-unbuilt features — did the bottleneck just move, or did it get worse? > - Is "taste and discernment" actually non-automatable, or is it automatable-later — a durable claim or a snapshot of current model capability? (Flagged in the source's own Logical Validity check as a hasty generalization.) > [!tip] Article Ideas > - A worksheet walking a WordPress product team through writing their own one-sentence mission filter and testing three roadmap items against it. #articleseed > - "Own the customer problem, not the product category" — a positioning piece on services/product convergence for WordPress agencies pivoting to SaaS. #articleseed > [!link] Cross Reference > - [[3. Resources/Permanent Notes/Implementation Is Cheap So Taste Becomes the Bottleneck]] — already captures this source's "AI collapses build cost, not judgment cost" claim, and already cites this exact episode as evidence > - [[3. Resources/Permanent Notes/The Bottleneck Moved From Code Generation to Review and Judgment]] — the review-capacity half of the same shift (30 PRs waiting) > - [[3. Resources/MOCs/Product & Startup Instincts MOC]] > - [[3. Resources/MOCs/AI & the Future of Knowledge Work MOC]] > [!quote] Vocabulary > - **TAM/SAM/SOM** — total addressable / serviceable addressable / serviceable obtainable market; a market-sizing funnel credited to Rev One Ventures. > - **Fuse score** — Gravity Kit's support-triage formula (frequency × users impacted × severity ÷ effort), modeled on Intercom's RICE-style scoring. > - **SLC (simple, lovable, complete)** — Brad Griffith's reframing of MVP, arguing "minimum viable" invites shipping something "barely alive." ## Bib capsule WP Product Talk, "Strategy Before Software: What Organizations Should Clarify Before Building," YouTube, August 13, 2026. Guest: Brad Griffith, Buckeye Innovation. https://www.youtube.com/watch?v=KJGd4KW06CM ## Author's thesis (Q1+Q2) Agentic AI coding has made "can we build it?" nearly always answerable yes, which makes strategic clarity more important, not less — the traditional engineering-hours bottleneck no longer forces prioritization discipline. A clear mission filter, an honestly-scoped TAM, and the discipline to name who a product does not serve are what let a team say no to an expanding backlog of technically-easy features; AI changes the economics of building, not the need for that filter. ## Key arguments (≤3) 1. **The bottleneck moved from building to deciding and reviewing.** Cheap AI-assisted implementation produces a "good idea backlog" — one host cites 30 PRs waiting for review — shifting the scarce resource from engineering hours to strategic judgment and review capacity. 2. **A one-sentence mission filter is a working strategy, better than a document.** Gravity Kit tests every feature against "people using our products should sell more tomorrow than they sold yesterday" — simple enough to apply in real time to a backlog AI has made cheap to fill. 3. **TAM discipline and direct validation both matter because market-sizing spreadsheets compound too many assumptions to trust alone.** Brad's ice-cream-franchise example (TAM capped near 100 stores, directly setting pricing) shows how a narrow, honest TAM shapes a product differently than WordPress/WooCommerce-scale markets do. ## Claim atoms ### Strategy separates real problems from solutions seeking one > [!quote] ID:rw_1 > you building a solution that is in seek of a problem that's not good Where/why: ~00:04:20, Brad Griffith's opening framing — sets the episode's central thesis that AI makes building cheap, not strategy unnecessary. ### AI moved the bottleneck from building to code review > [!quote] ID:rw_2 > we got 30 PRs waiting for review right now because every good idea is sitting in front of the team Where/why: ~00:14:50, Ian, describing the team's current backlog — direct evidence for the episode's "review capacity is the new constraint" claim. ### A one-sentence mission filter decides every feature > [!quote] ID:rw_3 > people using our products should sell more tomorrow than they sold yesterday that's it Where/why: ~00:22:39, Ian, describing Gravity Kit's operative mission filter — the strongest reusable operational artifact in the episode. ### Name who you are not serving > [!quote] ID:rw_6 > who are you serving and who are you not serving? Because it's really important. A lot of people just skip over that. Where/why: ~00:53:57, Brad, closing-round advice — positions negative-scope definition as equally important to positive targeting. ## Permanent candidates - The "AI collapses build cost, not judgment cost" claim (Z2/H2/H7) is already captured by [[3. Resources/Permanent Notes/Implementation Is Cheap So Taste Becomes the Bottleneck]], which already cites this exact episode as evidence. No duplicate created. - **New**: the "strategy as a saying-no filter, not a document" claim (Z1/H1/H3) is not yet captured — promoted below as [[3. Resources/Permanent Notes/A Working Strategy Is Measured by What It Lets a Team Refuse]]. ## Processing checklist - [x] Checked for existing Literature Note — none found - [x] Checked for existing Permanent Notes covering this source's claims — found [[3. Resources/Permanent Notes/Implementation Is Cheap So Taste Becomes the Bottleneck]] already covering the taste/judgment claim and already citing this episode; did not duplicate - [x] Connected into [[3. Resources/MOCs/Product & Startup Instincts MOC]] and [[3. Resources/MOCs/AI & the Future of Knowledge Work MOC]] ## Links - [[3. Resources/MOCs/Product & Startup Instincts MOC]] — mission filter, TAM discipline, SLC framing are all product-strategy instincts - [[3. Resources/MOCs/AI & the Future of Knowledge Work MOC]] — the build-cost-collapse/judgment-scarcity theme