Using AI to draft a quarterly board narrative arc

The premise

Boards retain decisions, not slides. AI can help you draft a narrative arc that ties what you committed to last quarter, what actually happened, and what you need next.

What AI does well here

• Outline a three-act arc from raw notes and metrics.
• Suggest concrete asks tied to specific risks.
• Reword vague claims into measurable statements.

What AI cannot do

• Judge which risks are politically safe to raise.
• Know unwritten history between board members.
• Decide what to omit for confidentiality.

Using AI to pre-register a pricing experiment

The premise

Pricing tests get rationalized after the fact. AI can draft a pre-reg that locks your hypothesis, sample size, and stop rule before data starts arriving.

What AI does well here

• Draft a one-page pre-reg from a loose hypothesis.
• Suggest realistic minimum detectable effects.
• Generate a checklist for the readout.

What AI cannot do

• Judge whether your traffic is actually independent.
• Know about confounding launches scheduled the same week.
• Decide acceptable downside risk for your business.

Using AI to teardown competitor positioning pages

The premise

Competitor teardowns take hours and produce inconsistent notes. AI can extract a structured teardown from raw page text using the same template every time.

What AI does well here

• Identify the headline promise and target audience.
• List the proof elements and their type.
• Spot weak claims and missing categories.

What AI cannot do

• Visit the live page and watch animations.
• Verify customer logos are real, paying customers.
• Know what the competitor is testing under flags.

Using AI to mine patterns in lost deals

The premise

Free-text loss reasons are noisy. AI can cluster them into a small number of themes and surface the ones that actually move forecast.

What AI does well here

• Cluster freeform notes into 5 to 8 themes.
• Suggest a label and one-sentence definition per theme.
• Flag deals where the note disagrees with the structured field.

What AI cannot do

• Replace a real loss interview with the buyer.
• Know which AE always blames pricing regardless of cause.
• Distinguish polite excuses from real reasons.

Using AI to narrate segment profitability data

The premise

Segment P&Ls drown people in numbers. AI can produce a one-page narrative that names the two or three segment shifts worth a meeting.

What AI does well here

• Identify the largest period-over-period swings.
• Explain swings in plain language with the underlying ratio.
• Suggest what to investigate next.

What AI cannot do

• Know which segment is strategic regardless of margin.
• Distinguish a real trend from a one-time accounting reclass.
• Predict next quarter without external context.

Using AI to size a strategic bet in a memo

The premise

Strategic bets often skip the math. AI can draft a memo that forces explicit ramp, cost, and downside assumptions before approval.

What AI does well here

• Structure a bet memo with explicit assumptions.
• Suggest sensible default ramp curves to challenge.
• Surface the assumption that most affects the answer.

What AI cannot do

• Know your true cost of capital.
• Judge whether the team can execute the bet.
• Account for political resistance from other teams.

Using AI to draft a customer segmentation tree

The premise

Segmentation often starts as a wishlist of dimensions. AI can propose a tree that prioritizes splits by how much they actually change behavior.

What AI does well here

• Suggest a 3-level segmentation tree from customer descriptions.
• Name each branch with a short, distinguishing label.
• Flag splits that look unactionable.

What AI cannot do

• Run statistical clustering on the underlying data.
• Know which split aligns with sales territories.
• Replace a structured customer research project.

Using AI to draft a renewal risk talk track

The premise

At-risk renewal calls often go badly because the CSM avoids naming the risk. AI can draft a talk track that names it explicitly and offers two real paths.

What AI does well here

• Draft an opener that names the risk without blame.
• Generate two concrete paths forward.
• Suggest questions that surface the customer's actual concern.

What AI cannot do

• Know the political dynamics inside the customer org.
• Replace the CSM's relationship and judgment.
• Predict whether the champion will defend the spend.

Using AI to draft a market entry go or no-go memo

The premise

Market entry decisions often drift because nobody writes a clear memo. AI can draft a go or no-go that forces a recommendation backed by named risks.

What AI does well here

• Structure a memo with thesis, evidence, risks, and recommendation.
• Suggest a minimum viable presence specific to the market.
• List entry-specific compliance items to investigate.

What AI cannot do

• Know current local regulatory enforcement posture.
• Judge cultural fit with on-the-ground teams.
• Account for FX and tax structure choices.

Using AI to sanity-check an OKR cascade

The premise

OKR drafts often have unmeasurable KRs and team objectives that do not roll up. AI is good at flagging both before publishing.

What AI does well here

• Flag KRs that lack metric, baseline, or target.
• Identify team OKRs that do not ladder to a company KR.
• Suggest tighter wording for vague objectives.

What AI cannot do

• Judge whether targets are appropriately ambitious.
• Know about secret OKRs at other teams.
• Predict whether the team will actually pursue the OKRs.