Actionable insights — Accenture (ACN): the narrative-overshoot screen

1. The narrative-overshoot screen — buy quality the market has condemned by category

1. Find a high-quality franchise sold off hard because it got swept into a category narrative (here: "AI kills anything touching enterprise") rather than judged on its own data.
2. Steelman the bear case honestly — concede the parts that are real (commodity work will be automated; the federal drag is ~1%).
3. Then ask whether the narrative has been "extrapolated well beyond what the actual data support."
4. Buy where the gap between the story and the data is widest.

• Whole sectors repriced on a single thesis; the best names inside them being sold for their "ZIP code," not their fundamentals.

2. Use bookings / book-to-bill as the truth serum against a disruption story

1. For a services/contract business, go to forward indicators — new bookings, book-to-bill, big-client count — not trailing revenue.
2. A book-to-bill above 1.0 means future revenue is growing: $1.20 of new work won per $1 recognized is the opposite of a business being automated away.
3. "Companies whose work is being automated away do not produce record booking numbers" — let the order book overrule the narrative.

• Book-to-bill direction and record-client counts as the leading tell — they move before reported revenue.

3. The "sell the AI-arms-dealer" inversion — own the integrator the disruption needs

1. When a technology shift is feared as a threat to a company, ask whether that same shift actually creates demand for what the company sells.
2. For AI specifically: every dollar of hyperscaler AI infrastructure requires a dollar of enterprise services to integrate, govern, secure and operationalize it — and AI misfires make expert human supervision more valuable, not less.
3. Find the company sitting on the demand side of the disruption (the integrator/governance layer), priced as if it's on the casualty side.

• "Picks-and-shovels" or integration/governance providers to a feared technology — they often benefit from what's supposed to kill them.

4. Anchor on FCF yield, not P/E, for a quality compounder at a trough

1. Compute the free-cash-flow yield (FCF ÷ market cap) and compare it to the company's own history and to risk-free rates.
2. A 10%+ FCF yield on a 26%-ROE, dividend-growing business is the "most striking single metric" — it implies either permanent impairment or a mispricing.
3. Cross-check the forward P/E against its own decade range: ~15× vs a 25–35× norm means the 50% drop is pure multiple compression, not earnings collapse.
4. Use a record-high dividend yield "not far below 2-yr Treasuries" as the paid-to-wait floor.

• Quality names whose FCF yield blows out to double digits while ROE and the dividend stay intact — the valuation, not the business, has broken.

5. Time the catalyst to reset expectations — buy ahead of a beatable quarter

1. Identify the next earnings date and check whether consensus estimates have already been cut hard (a low bar).
2. Confirm the leading data (bookings, AI revenue) is running ahead of the lowered guidance.
3. A clean beat plus reaffirmed guidance against reset expectations is the trigger for "at least a partial upward re-rating" — position before it, not after.

• A reset-low estimate bar plus a record order book heading into the print — the asymmetry tilts to the upside.

6. Pre-commit to the real bear case — and the metric that would prove it right

1. Write down the most credible bear case explicitly — and notice it's usually not the popular one.
2. For a labor-leveraged services firm, the real risk is margin compression (AI does the same work with fewer people → busier but less profitable), not demand destruction.
3. Name the falsifier in advance: "a sustained pattern of revenue growth without margin expansion would validate the bear case" — monitor that exact metric.
4. Set the holding period to match how the thesis resolves: re-rating "only comes from quarterly data accumulating" → plan a 12–18-month-minimum hold.

• The operating-margin trend each quarter as the make-or-break metric; a multi-quarter horizon for the multiple to mean-revert.