AI for revops
RevOps sits at the intersection of sales, marketing, and CS — translating between functions and turning messy systems into legible processes. AI helps with the parts that scale poorly: cleaning data, documenting processes, surfacing pipeline patterns, building stakeholder briefs. The judgment about what to build, what to automate, and what to leave alone stays human.
What AI handles well
CRM data hygiene at scale
The problem. Your CRM is a dumpster. Inconsistent industry tags, missing email domains, duplicate accounts, wildly different "company size" values. Cleaning manually = months.
What AI does. Use AI to normalize fields in bulk. Have it generate the rules first (so you can review them), then apply. Tag confidence scores so a human reviews ambiguous cases.
Pipeline analysis without manual SQL
The problem. Sales leadership wants to know why deals are stalling. Your CRM dashboards show what — not why. Building cohort views every time is expensive.
What AI does. Use AI to interpret pipeline data with context. Surface stage-by-stage drop-off, identify the deals that look like outliers, and suggest the 2-3 hypotheses worth investigating before pulling more reports.
Lead scoring logic that's actually defensible
The problem. Marketing scores leads on engagement signals. Sales says they're garbage. You're stuck in the middle defending a scoring model nobody trusts.
What AI does. Use AI to build a metric tree from "qualified opportunity" backwards. Identify which signals actually correlate with deal close, not just engagement. Output: a defensible scoring model with reasoning.
Process documentation new hires read
The problem. Your sales process docs are a wall of text in Notion. New AEs don't read them. They learn by watching peers — which means tribal knowledge, inconsistent execution.
What AI does. Use AI to convert your tribal knowledge (paste call transcripts, top-rep messages, deal notes) into role-specific runbooks. Format as actionable checklists, not philosophy.
Compensation plan modeling
The problem. Leadership wants to redesign comp. You need to model 5 scenarios across 30 reps. Each tweak requires 2 hours of spreadsheet wrangling.
What AI does. Have AI structure the model — quotas, accelerators, edge cases, payouts under different scenarios. You provide the assumptions, AI surfaces the implications and where the model breaks at edge cases.
Cross-functional briefs (sales+marketing+CS → exec)
The problem. Monthly business review. You need to pull data from 3 systems, synthesize across functions, and present it to the exec team. The briefs always read like data dumps.
What AI does. AI structures the brief: what changed, why, what we're proposing, the asks. Built around the decision the exec team has to make, not the data they'll skim.
Forecast sanity checks
The problem. Sales submits a $2.5M forecast. Your gut says it's soft. You don't have time to deep-dive every rep's pipeline.
What AI does. Have AI flag anomalies — deals stuck in stage too long, deals that grew suspiciously near month-end, reps with unusual close-rate spikes. AI highlights what to ask about, you do the deep-dive.
Tool/vendor evaluation
The problem. Sales wants 4 new tools this quarter. Each vendor pitches well. You need to make defensible recommendations to the CRO.
What AI does. Use AI to build a structured evaluation: feature parity, integration cost, switching cost, what we're currently doing without it. Output: a clear "yes/no/maybe later" with reasoning.
Your AI stack
Start with the foundation. Add specialized tools as the work calls for them.
Foundation LLM
Specialized add-ons
Prompts ready to use
Get started in 35 minutes
Pick a foundation LLM and load your business context
10 minSet up a Claude Project with: your sales motion, your stages, your ICP, your top objections, your team structure. Now any prompt about your pipeline starts pre-loaded.
Run the data-gotchas prompt on your most-used CRM report
15 minPick a report your team relies on. Have AI list selection effects, definition shifts, and survivorship bias risks. You'll find at least one thing worth fixing.
Convert one tribal-knowledge process into a checklist
10 minPick a process new hires struggle with (deal handoff, stage progression rules, etc.). Use the tutorial-to-checklist prompt. Ship the result to your team.
Common mistakes
Trusting AI-generated forecasts without sanity-checking. AI is great at extrapolation. It's bad at knowing when the trend is about to break. Use AI to surface anomalies, not to predict the future.
Letting AI define lead scoring without validating against actual outcomes. The model that scores high on engagement might score low on close-rate. Validate, don't assume.
Generating dashboards AI thinks reps need. Talk to actual reps. AI doesn't know what your team's real friction is.
Pasting customer pipeline data into public AI. Use enterprise tier or anonymize identifiers before pasting.
Process docs that read like AI wrote them — generic, polished, soulless. Edit the parts that need to sound like a real human at your company. Otherwise no one adopts them.