Resources
AI Playbooks
166 sequenced workflows for revenue teams. Every step names an owner, a tool, and a definition of done. Start at the level your team actually operates at, then move one rung. Tools are catalogued in the AI Tool Library.
The L1 to L6 scale shows how deeply AI is built into the work. L1 Starter adds one tool without changing the workflow. L2 Assisted lets AI draft while people approve. L3 Integrated connects AI to systems such as CRM and Slack. L4 Orchestrated assigns an owner, a sequence, and a measure. L5 Autonomous lets an agent run the workflow while a person audits the output. L6 Rebuilt changes the process itself. Use the level to choose the next playbook your team can run now, not the most advanced one on the page.
Read the full L1 to L6 framework, with the tech, upside, downside, and gate at each rung →
166 playbooks
Sales / L1 Starter
Always-on AI meeting notes (L1)
Every customer-facing meeting gets transcribed, summarized, and pushed into a shared Slack channel automatically. Zero workflow change for reps, they just keep taking the meeting. Granola, Fathom, Fireflies, or Otter, pick one and standardize.
4 steps
Sales / L1 Starter
ChatGPT for first-draft cold emails (L1)
The simplest possible AI lift: reps paste a LinkedIn profile + one pain point into ChatGPT and get a 3-variant cold email draft. No integrations, no automation, just a shared prompt and a 10-minute Loom showing how to use it.
4 steps
Every team / L1 Starter
GitHub Copilot for everyday autocomplete (L1)
Every engineer gets GitHub Copilot in their IDE. No process, no review board, no shared prompts, just turn it on. The cheapest way to start measuring AI productivity in engineering.
3 steps
Every team / L1 Starter
Perplexity for daily research (L1)
Replace Google for any work question where you need cited sources fast. Every IC gets a Perplexity Pro account; managers expect to see Perplexity links shared in Slack instead of Google SERPs.
3 steps
Sales / L1 Starter
Personal AI assistant for repetitive sales admin (L1)
Give each rep a personal AI assistant (Lindy or Relevance) wired to their Gmail + Calendar. It drafts follow-up emails after meetings, books internal reviews, and chases unanswered threads. Saves 30,60 min/rep/day with no CRM changes required.
4 steps
Sales / L1 Starter
Zapier + ChatGPT inbound lead summarizer (L1)
When a demo request comes in, Zapier sends the company website + form responses to ChatGPT, gets back a 5-line account brief (what they do, likely use case, 2 discovery questions), and posts it to Slack before the AE picks up the phone. Your first real AI-in-the-loop workflow.
4 steps
Sales / L2 Assisted
AI call notes via Otter (L2)
A sanctioned recorder summarizes calls and posts to Slack. It is not coaching, it is transcription with lipstick.
5 steps
Whole company / L2 Assisted
AI literacy bootcamp for GTM (L2)
4-week curriculum: prompting, tools, ethics, security. Measurable certification. Without this, every other initiative leaks.
5 steps
Every team / L2 Assisted
Bardeen + Make.com personal workflow agents (L2)
Equip every IC with a personal automation stack: Bardeen for browser-side capture (scrape, save to Notion, route emails), Make.com or n8n for cross-tool flows (HubSpot \→ Slack \→ Sheets). Saves 30\,60 min/IC/day on the long tail of admin nobody owns.
4 steps
Whole company / L2 Assisted
Centralized AI prompt library (L2)
A shared, versioned, reviewed prompt library. Boring infra; necessary.
5 steps
Whole company / L2 Assisted
Claude for Work shared projects (L2)
A whole team standardizes on Claude for Work (Projects + shared knowledge). PM, design, and eng all chat to the same context, product specs, design docs, codebase snippets, instead of copy-pasting into 5 different chats.
4 steps
Customer success / L2 Assisted
CSM call summaries via Gainsight + AI (L2)
Sanctioned summarizer posts to account record. Saves CSMs time, does not change retention.
5 steps
Every team / L2 Assisted
Cursor + Claude for whole-team coding (L2)
Standardize the team on Cursor or Windsurf with Claude Sonnet as the default model. Shared .cursorrules / .windsurfrules file in every repo so the AI follows your team conventions, not generic JavaScript advice.
4 steps
Whole company / L2 Assisted
Notion AI + Claude for company knowledge ops (L2)
Stop people from re-asking the same 100 questions in Slack. Centralize docs in Notion, enable Notion AI Q&A, and run a weekly \"What got asked in Slack that should be a doc?\" ritual driven by Claude scanning #ask-* channels.
4 steps
Sales / L2 Assisted
Sanctioned email + meeting assistant (L2)
One approved AI tool for the whole revenue team (e.g. Lavender, Regie, Copilot). Centralized billing, basic usage tracking, prompt library owned by enablement.
5 steps
Sales / L2 Assisted
Sequencer AI variants (Outreach/Salesloft) L2
Native AI generates step variants. Lifts open rates a bit; does not fix bad ICP targeting.
5 steps
Revenue operations / L2 Assisted
Voice-of-Customer → Product & Enablement Loop (L2)
Two-prompt agent reads every call transcript + Salesforce context daily and outputs a weekly dashboard with key themes, use-case library, product-team actions, enablement actions, and a champion-quote sizzle reel. As shown by Nate Follen, Head of GTM Systems & Ops at Perplexity, on the GTM AI Podcast.
4 steps
Revenue operations / L2 Assisted
Zapier Agents for sales ops chores (L2)
Agents handle small recurring chores: 'every Monday email me deals with no activity'. Saves an hour, frees nothing strategic.
5 steps
Revenue operations / L3 Integrated
Agentic Nightly CRM Hygiene + Ownership Sync (L3)
Nightly batch job that cleans account ownership, hierarchies, and contact-domain mismatches in Salesforce. Replaces brittle after-save flows that overlap and break at scale. From Nate Follen at Perplexity on the GTM AI Podcast.
4 steps
Whole company / L3 Integrated
AI council + intake process (L3)
Cross-functional council reviews AI tool requests weekly. Stops shadow IT, sets standards, owns budget.
5 steps
Marketing / L3 Integrated
AI customer-interview synthesis (L3)
Run discovery / win-loss / churn interviews, auto-transcribe with Granola or Fireflies, and use Claude (Projects + 200K context) to synthesize themes across 20+ interviews into a single PMM-ready insight doc. Replaces 2 weeks of manual qualitative coding with 2 hours.
4 steps
Sales / L3 Integrated
AI meeting intelligence rolled out properly (L2→L3)
Most teams buy Gong/Fireflies and stop. To move from L2 to L3, the transcripts feed back into CRM fields, coaching plans, and forecast calls.
5 steps
Marketing / L3 Integrated
AI SEO clusters with Clearscope/Frase (L3)
Pillar + cluster strategy generated and graded by AI. Tied to GSC data. This actually moves traffic.
5 steps
Sales / L3 Integrated
AI-assisted RFP responses (L3)
Loopio / Responsive + LLM drafts from your answer library. Cuts RFP time 60%+ when library is clean.
5 steps
Sales / L3 Integrated
AI-augmented outbound workflow (L3)
AI is embedded in the actual outbound workflow inside CRM/engagement, research, segmentation, sequencing, follow-up, not a side tool. This is the threshold between "we use AI" and "AI is part of how we sell."
6 steps
Revenue operations / L3 Integrated
AI-driven price-quote acceleration with CPQ (L3)
LLM drafts quote from opp + email + product config. Human approves. Cuts quote time 70%.
5 steps
Sales / L3 Integrated
AI-First BDR, Internal Signals, Not LinkedIn Scrapes (L3)
Sendoso cut from 15 BDRs to 1, then rebuilt to 4-5 with AI. Pipeline went from <15% to >30% of total. The unlock: marrying Snowflake product usage + Salesforce closed-lost/champion history + UserGems, not the LinkedIn-scrape email everyone else sends. From Austin (Sendoso) on the GTM AI Podcast.
5 steps
Marketing / L3 Integrated
AI-generated event follow-up (L3)
Trade show scans → enrich → personalize → sequence within 24h. The window matters.
5 steps
Customer success / L3 Integrated
AI-generated QBR decks (L3)
Pull product usage + support + CSM notes → LLM generates a draft QBR deck. CSM edits, does not author.
5 steps
Every team / L3 Integrated
Claude Code for agentic refactors and migrations (L3)
Use Anthropic\'s Claude Code CLI as a teammate that can read your whole repo, propose multi-file refactors, run tests, and open PRs. Goes beyond autocomplete: you give it a goal, it iterates on the codebase like a junior engineer with a coffee IV.
4 steps
Sales / L3 Integrated
Clay → Smartlead → Lindy outbound stack (L3)
A modern outbound stack with zero SDRs: Clay sources + enriches accounts on intent triggers, Smartlead/Instantly handles deliverability + sending, and a Lindy agent handles every positive reply (book meeting, route to AE, log to CRM). Cost ~$2k/mo, output of 2\,3 SDRs.
4 steps
Sales / L3 Integrated
Clay + GPT inbound enrichment (L3)
Every inbound demo request is enriched with funding stage, hiring signals, tech stack, and routed with a one-page brief to the AE. This is where most teams should aim first.
5 steps
Sales / L3 Integrated
Cold call AI roleplay simulator (L2-3)
Reps practice cold calls with an AI persona before live dials. New-hire ramp drops 30%.
6 steps
Revenue operations / L3 Integrated
CRM dedupe + enrichment with LLMs (L3)
Clay + LLM merges duplicate accounts, fills missing firmographics, normalizes job titles. Pays for itself in 30 days.
5 steps
Sales / L3 Integrated
Earnings-call digest for enterprise AEs (L3)
LLM ingests target accounts' earnings calls + 10-Ks; AE gets a 1-pager before exec meetings.
5 steps
Sales / L3 Integrated
Gong call → CRM auto-fill (L3)
Gong AI extracts MEDDPICC fields and writes them to the opportunity. Stop asking reps to update Salesforce.
5 steps
Sales / L3 Integrated
HubSpot MCP for AE chat (L3)
AEs ask 'show me my deals slipping past close date with no activity in 14 days' in chat. HubSpot MCP returns the answer. Self-serve ops.
5 steps
Customer success / L3 Integrated
Intercom Fin AI for support deflection (L3)
Deploy Intercom Fin (or Ada/Zendesk AI) trained on your docs + past ticket history. Fin answers ~50\,70% of support tickets without a human, routes the rest with context attached. Each deflection pays for itself in 2 tickets.
4 steps
Marketing / L3 Integrated
Lifecycle email AI personalization (L3)
Customer.io / Iterable + LLM choose subject line, hero copy, and CTA per segment. Real lift if you have the data.
5 steps
Sales / L3 Integrated
Lindy.ai meeting prep agent (L3)
Agent reads calendar 1h before each meeting, pulls company news + CRM history + last 3 emails, drops a brief in Slack.
5 steps
Sales / L3 Integrated
MindStudio agent for partner intros (L3)
Agent monitors LinkedIn changes in your network, suggests warm intro paths to target accounts.
6 steps
Sales / L3 Integrated
Momentum.io call-to-CRM autofill (L3)
Momentum.io listens to every sales call (Zoom/Meet/Gong), extracts MEDDPICC fields, next steps, risks, and competitor mentions, and writes them back into Salesforce/HubSpot deal records, automatically. Reps stop CRM-updating; managers get reliable pipeline data.
4 steps
Enablement / L3 Integrated
Notion-Native Account Plan Agent, "Yes Chef" (L3)
5 chained prompts in Notion AI generate a full account plan, stakeholder map, tiered contacts, history, recommended actions, in 90 seconds. Killed the 6-week change-management drag of rolling out account planning. By Justin Dries, Director of Sales Enablement at Legora, on the GTM AI Podcast.
5 steps
Sales / L3 Integrated
Outbound from intent signals (L3)
Bombora/6sense intent → Clay → AI personalized 3-touch sequence. Replace spray-and-pray.
6 steps
Marketing / L3 Integrated
Paid search AI bidding + creative (L3)
Google Performance Max + AI-generated headlines/descriptions, refreshed weekly by an LLM agent reading your top landers.
5 steps
Sales / L3 Integrated
Perplexity Comet for agentic account research (L3)
Use Perplexity\'s Comet agentic browser to do real account research: it navigates the prospect\'s site, 10-K, recent press, and competitor pages, then assembles a structured one-pager for the AE before each first call.
4 steps
Every team / L3 Integrated
Portable Chat Memory Across Models (L3)
Open-source browser extension (Chat Archive) exports Claude/ChatGPT/Gemini/Grok threads to JSON or Markdown, fully local, zero outbound inference. Switch providers mid-thread without losing context. Built by John Williams (FXOps) and demoed on the GTM AI Podcast.
4 steps
Sales / L3 Integrated
RAG-powered competitive intel (L3)
Crawl competitor sites, G2, Reddit, earnings calls → embed → reps query in Slack. Beats stale battle cards.
5 steps
Revenue operations / L3 Integrated
RevOps data cleanup with LLMs (L3)
Use LLMs to deduplicate accounts, normalize titles/industries, and reconcile CRM ↔ billing. Unsexy, highest-ROI play in most orgs.
6 steps
Sales / L3 Integrated
Sales coaching with AI scorecards (L3)
Gong / Chorus AI scores every call against your rubric. Managers coach to the gap, not to vibes.
5 steps
Customer success / L3 Integrated
Support deflection with AI search (L3)
Intercom Fin / Zendesk AI answers tier-1 tickets. Real cost savings if your knowledge base is decent.
5 steps
Marketing / L3 Integrated
Voice-of-customer aggregator (L3)
LLM continuously aggregates calls, tickets, NPS verbatims, reviews into a weekly digest by theme.
5 steps
Revenue operations / L3 Integrated
Weekly RevOps Leadership Deck, On Cron (L3)
Replaces the manual hour-long deck prep. A scheduled agent pings reps to refresh in-month pipeline, re-reads the thread, pulls from Slack/Linear/Snowflake, and ships the deck to leadership. Demoed by Nate Follen at Perplexity on the GTM AI Podcast.
4 steps
Marketing / L3 Integrated
Win/loss interview synthesis (L3)
Record win/loss calls, LLM extracts themes monthly, posts to PMM Slack. Replaces the consultant.
5 steps
Marketing / L4 Orchestrated
Account-based MQA scoring (L4)
First-party + 3P signals feed a model that scores accounts (not leads) by ICP fit + intent + engagement. Sales gets a ranked list, not a queue.
5 steps
Whole company / L4 Orchestrated
Agent observability + cost monitoring (L4)
Every AI call logged with cost, latency, prompt, output. Without this, you cannot scale agents responsibly.
5 steps
Revenue operations / L4 Orchestrated
AI agent for inbound triage and routing (L4)
Agent reads inbound (email, web, chat), classifies intent, enriches, and assigns to the right rep with a brief. No more lead round-robin lottery.
6 steps
Revenue operations / L4 Orchestrated
AI-augmented deal desk (L4)
Deal desk uses agent to summarize the deal, flag risks, suggest terms. Reduces approval cycle from days to hours.
5 steps
Marketing / L4 Orchestrated
AI-driven ICP refinement (L4)
Re-derive your ICP every 6 months from closed-won data using LLM clustering. Stop guessing in a whiteboard session.
6 steps
Revenue operations / L4 Orchestrated
AI-driven territory design (L4)
Annual ritual; AI optimizes territories on opportunity density + travel + rep skill. Replaces the spreadsheet horror.
5 steps
Sales / L4 Orchestrated
AI-Generated Single-File HTML Sales Proposals (L4)
Replace PowerPoint with a single-file HTML proposal generated from Salesforce + call history. AI auto-selects testimonials, goals, and pricing context; rep can hand-edit or "let AI decide." Cloudflare auth keeps each proposal private. From Egan at Sendoso on the GTM AI Podcast.
5 steps
Revenue operations / L4 Orchestrated
Automated Lead Qualification and Personalization Engine
This is a lead processing factory, not a simple connector. It uses a cascaded enrichment waterfall in Clay to build a complete lead profile, then uses OpenAI to qualify, score against an ICP, and generate personalized sequence-starters. It's designed to route only qualified, data-rich leads to sales, with specific paths for top-tier, mid-tier, and unqualified targets.
8 steps
Revenue operations / L4 Orchestrated
Contract Scraper → CRM + Upsell Signal Engine (L4)
Self-hosted n8n workflow extracts line items from every closed-won PDF (OCR + image + text into Opus), runs a deterministic JS auditor, escalates only low-confidence cases to a human, then writes structured plan data to Salesforce. ~45 min saved per renewal. Built by Egan at Sendoso, shown on the GTM AI Podcast.
5 steps
Customer success / L4 Orchestrated
Customer health scoring with AI signals (L4)
Combine product usage, support sentiment, and exec-engagement signals into a single churn-risk score with explicit experiments.
5 steps
Revenue operations / L4 Orchestrated
Data-anchored lead scoring (L4)
Lead scoring tied to first-party product/usage + third-party intent data, with explicit experiments measuring win-rate lift. The model is owned, not bought as a black box.
5 steps
Revenue operations / L4 Orchestrated
Forecast augmentation with Gong + ML (L4)
Combine rep-submitted forecast with Gong call-signal model. Manager sees both; uses the gap as coaching trigger.
5 steps
Revenue operations / L4 Orchestrated
Gong + Momentum forecast augmentation (L4)
Layer Gong\'s conversation-intelligence sentiment + Momentum\'s structured MEDDPICC autofill into your forecast model. Replace the rep-call-on-Friday with an AI-generated forecast that reps adjust, not author. Forecast accuracy +20-30 points typical.
4 steps
Revenue operations / L4 Orchestrated
Lead scoring v2 with first-party + LLM features (L4)
Replace the rule-based score with a model that uses product usage, intent, and LLM-derived fit-from-website features.
6 steps
Every team / L4 Orchestrated
Local Inference as Token-Budget & Outage Insurance (L4)
Run a small local model (Gemma, Mistral, Llama via Ollama) as redundancy for routine work, privacy-sensitive analysis, and GPU brownout days. Stay productive when Claude/OpenAI go down. From John Williams (FXOps) on the GTM AI Podcast.
4 steps
Sales / L4 Orchestrated
Momentum.io autonomous deal desk (L4)
Momentum.io upgraded from CRM autofill into a full deal-desk agent. After every call it (1) updates MEDDPICC, (2) flags stuck deals with no champion in 14+ days, (3) drafts the follow-up email referencing actual call quotes, (4) escalates risk to leadership in Slack with proposed mitigation, and (5) auto-builds the QBR deck from the quarter's call data.
6 steps
Sales / L4 Orchestrated
Multi-thread agent on at-risk deals (L4)
Agent reads CRM + Gong, flags single-threaded deals, drafts exec emails to add stakeholders. Human sends.
5 steps
Customer success / L4 Orchestrated
Onboarding milestone agent (L4)
Agent monitors product events; nudges customer and CSM when a milestone slips. Reduces time-to-value.
5 steps
Every team / L4 Orchestrated
OpenAI Codex cloud agents for PR automation (L4)
Wire OpenAI Codex (the 2025 cloud coding agent) to your GitHub. Triage incoming issues, propose fixes as draft PRs, run tests in sandboxed environments, and ping the on-call engineer when ready for review. Your repository now has a 24/7 junior contributor.
4 steps
Revenue operations / L4 Orchestrated
Pricing experimentation with AI (L4)
AI suggests deal-level discounts based on win probability and competitor presence. Stop the across-the-board 10%.
5 steps
Sales / L4 Orchestrated
Relevance AI BDR agent (L3-4)
Multi-step agent: research → personalize → email → handle reply → book meeting. Real, not magic; needs constant tuning.
6 steps
Customer success / L4 Orchestrated
Renewal forecasting with AI (L4)
Replace CSM gut with model trained on usage, support, sentiment. CFOs love this.
5 steps
Customer success / L4 Orchestrated
Salesforce Agentforce service triage (L4)
Agentforce reads cases, classifies, drafts response, escalates. Real deflection if your data model is clean.
5 steps
Sales / L4 Orchestrated
Voice-AI inbound qualification with Vapi/Bland (L4)
When a high-intent inbound lead submits a demo form, a Vapi or Bland.ai voice agent calls them within 60 seconds, qualifies BANT-style, and books the meeting on the AE\'s calendar. Cuts inbound speed-to-lead from hours to seconds; converts 2\,3x.
5 steps
Whole company / L5 Autonomous
AI Acceptable Use Policy + Agent Commerce Guardrails (L5)
Two open specs to bring sanity to org-readiness: an AI AUP baseline that combats shadow AI without over-restricting, and Agent Commerce guardrails that define what terms/$ limits an agent may auto-approve. From John Williams (FXOps) on the GTM AI Podcast.
4 steps
Every team / L5 Autonomous
Autonomous engineering team with Devin + Claude Code (L5)
A handful of senior engineers + a fleet of autonomous agents (Devin for project-level work, Claude Code for repo-level chores, Codex for issue triage). Humans set strategy and architecture, write the hard 20% of code, review everything. Output per human engineer 3\,5x.
4 steps
Revenue operations / L5 Autonomous
Data Cloud-powered NBA in Salesforce (L5)
Data Cloud unifies product, billing, support; Einstein recommends next action in the AE flow. Heavy lift, big payoff.
5 steps
Revenue operations / L5 Autonomous
Next-best-action engine for AEs (L5)
Every morning each AE gets a ranked list of actions by expected $ value. Reps stop guessing what to work on.
5 steps
Revenue operations / L5 Autonomous
Next-best-action engine for AEs (L5)
For every open opp, the system recommends and ranks the next action by expected value. Managers manage to the plays, not to dials.
5 steps
Marketing / L5 Autonomous
Predictive churn → marketing save (L5)
CS health model triggers marketing save plays automatically. Tied to NRR.
5 steps
Revenue operations / L5 Autonomous
Self-Hosted Deep Research over Customer Comms (L5)
Self-hosted vector DB of all customer calls/emails/surveys with a hybrid (semantic + text) re-ranker. An agent harness fans out: project → queue → per-deal analyzer → synthesizer → HTML report. The reason most AI search tools feel wrong: they pretend to read every deal but actually rank-and-snip. From Egan at Sendoso on the GTM AI Podcast.
5 steps
Whole company / L6 Rebuilt
AI-native GTM operating model (L6)
Org, roles, and comp redesigned around AI capabilities. Leaner middle management, AI copilots default, central AI platform team, EV-portfolio of bets.
5 steps
Whole company / L6 Rebuilt
Comp redesign for AI-augmented reps (L6)
Quota up, headcount flat or down, comp tied to expected-value actions completed. Most orgs are not ready. Do not start here.
5 steps
Sales / L5 Autonomous
Autonomous pipeline generation with a human audit lane (L5)
An agent researches accounts, writes the sequence, enrolls the contact, and books the meeting. Humans stop approving each send and start auditing a weekly sample. Works on the stack most teams already run: Clay or ZoomInfo for data, Outreach or Salesloft for delivery, Salesforce or HubSpot for the record.
4 steps
Sales / L6 Rebuilt
Rebuild account coverage around agents, not headcount (L6)
The process itself changes. Instead of adding reps to cover more accounts, you split coverage into an agent tier and a human tier, then rewrite quota, routing, and comp to match. This is an operating model change, not a tool rollout.
4 steps
Marketing / L1 Starter
AI first drafts inside the marketing tools you already pay for (L1)
No new vendor. Use the AI already bundled in HubSpot, Canva, and Google Workspace to produce first drafts of social posts, email subject lines, and ad variants. Humans still edit and publish everything.
4 steps
Marketing / L2 Assisted
Brief to draft content assembly with human approval (L2)
AI drafts, humans approve. The brief, the source material, and the draft live in one place so the writer edits rather than starts from blank. Notion or Google Docs plus Claude or ChatGPT is enough.
4 steps
Marketing / L6 Rebuilt
Rebuild demand generation around answer engines (L6)
The channel changed, so the process changes. Instead of writing pages that rank, you publish sourced answers that get cited by AI search, and you measure citations and assisted pipeline rather than blue-link position alone.
4 steps
Revenue operations / L1 Starter
AI record summaries inside Salesforce or HubSpot (L1)
Turn on the summary feature your CRM already ships. Every account and opportunity gets a plain-language recap so managers stop reading raw activity logs. No integration work, no new contract.
4 steps
Revenue operations / L6 Rebuilt
Rebuild the revenue data model so agents can act on it (L6)
Agents fail on messy data long before they fail on reasoning. This rebuilds definitions, permissions, and access paths so an agent reads and writes the same governed truth a human does. Warehouse plus CRM plus a governed access layer.
4 steps
Customer success / L1 Starter
AI ticket and conversation summaries in Zendesk or Intercom (L1)
Turn on the native summary feature so every ticket and chat carries a plain recap and a suggested next step. Agents keep full control. This is the lowest-risk entry point for a post-sale team.
4 steps
Customer success / L5 Autonomous
Autonomous renewal risk agent with human audit (L5)
An agent watches product usage, support volume, sentiment, and stakeholder change, then opens a risk case and drafts the save plan on its own. The CSM audits and decides. Gainsight, Salesforce, or HubSpot plus your product telemetry.
4 steps
Customer success / L6 Rebuilt
Rebuild post-sale coverage around agents and pooled CS (L6)
Named-CSM coverage for every account stops being the default. Low-touch accounts move to an agent-plus-pool model, high-touch accounts get deeper human coverage, and retention targets are set per tier rather than blended.
4 steps
Enablement / L1 Starter
AI call highlight reels for coaching (L1)
Use the highlight and topic tracking your call recorder already produces. Managers coach from three-minute clips instead of hour-long recordings. Gong, Chorus, Fathom, or Fireflies all do this out of the box.
4 steps
Enablement / L2 Assisted
Onboarding ramp assistant built on your own content (L2)
New hires ask questions in Slack and get answers drawn only from your approved enablement library. AI drafts the answer, enablement approves the source set. Notion AI, Guru, or Glean over the content you already have.
4 steps
Enablement / L4 Orchestrated
Certification with AI roleplay scoring (L4)
Reps certify on discovery, pricing, and competitive objections against an AI buyer that scores against your rubric. Orchestrated means it is scheduled, owned, scored, and tied to a gate before reps touch live accounts.
4 steps
Enablement / L5 Autonomous
Autonomous coaching nudges from call data (L5)
The system detects a coachable pattern across a rep's calls, drafts the nudge, and delivers it in Slack without waiting for a manager. Managers audit the nudges rather than write them.
4 steps
Enablement / L6 Rebuilt
Rebuild enablement as a living knowledge system (L6)
Static decks and quarterly trainings are replaced by a maintained source of truth that feeds reps, agents, the website, and support from the same content. Enablement stops producing artifacts and starts owning a system.
4 steps
Whole company / L1 Starter
One sanctioned AI assistant for every employee (L1)
Give everyone a licensed enterprise assistant so work stops happening in personal accounts. ChatGPT Enterprise, Microsoft Copilot, or Google Gemini for Workspace. This is the cheapest way to reduce shadow AI risk.
4 steps
Every team / L6 Rebuilt
Rebuild internal operations around agent workflows (L6)
Recurring internal processes stop being human checklists with AI help and become agent-run workflows with human exception handling. Approvals, reporting, and request intake are the first three to move.
4 steps
Every team / L2 Assisted
The 30-day first workflow (L2)
Most teams buy a platform before proving a single workflow, then spend nine months explaining the invoice. This produces a defensible before-and-after number on one workflow inside 30 days for the price of ten seats of ChatGPT Team, Claude Team, Gemini Business or Microsoft 365 Copilot. Strongest at 50 to 500 employees where one person owns the whole thing and there is no procurement committee.
7 steps
Enablement / L2 Assisted
Turn dormant seats into one shipped workflow (L2)
Companies buy assistant seats broadly and get real usage from a handful of people while the prompts that work already exist in five power users' heads. This packages them so the other 90% get the same output without learning to prompt, and reclaims the seats nobody logs into. Strongest at 100 to 2,000 employees with a company-wide ChatGPT Team, Claude Team, Gemini Business or Microsoft 365 Copilot rollout that has gone quiet.
8 steps
Revenue operations / L3 Integrated
Borrow the engineering harness for revenue (L3)
When engineering has AI tooling and revenue does not, the blocker is rarely appetite. Engineering already cleared vendor review with tools like GitHub Copilot, Cursor or Claude Code, wrote the data policy, and learned the procurement path, so copying their paperwork removes roughly two months from your timeline and costs one meeting. Strongest at 200 to 5,000 employees, software and technology-forward services, where engineering adopted AI tooling 6 to 18 months before the revenue org.
7 steps
Revenue operations / L4 Orchestrated
Put an answer layer on the warehouse you already paid for (L4)
A modern warehouse with no natural-language layer means every revenue question queues behind an analyst, and pointing a model at raw tables returns a confident number nobody can reconcile. Freezing a small set of metric definitions in a semantic layer like dbt Semantic Layer, Cube or LookML and letting the assistant answer only from those turns the warehouse into something a CRO can query on a Sunday night. Strongest at 500+ employees with a working data platform and a real analytics team; skip it if metric definitions are still argued about in meetings.
8 steps
Sales / L3 Integrated
Make conversation data do work (L3)
Call recordings are the only genuinely un-copyable data a revenue team owns, and in most companies they sit in a library and rot. Extracting a fixed handful of qualification fields from tools like Gong, Chorus or Avoma with the supporting quote attached gets the forecast populated before the pipeline review starts, letting the meeting be about decisions instead of reading status aloud. Strongest at 200 to 5,000 employees, B2B with a real multi-call sales cycle and a qualification framework already in use.
7 steps
Marketing / L3 Integrated
Signal layer before more sequences (L3)
World-class sending machinery in Outreach or Salesloft pointed at an unscored list produces volume, not meetings, because AI makes bad targeting faster rather than better. This converts the ICP into a machine-checkable filter, adds exactly two signals, and caps the queue at 25 accounts per rep per week, the cap being the entire mechanism. Strongest at 100 to 2,000 employees running outbound with a defined ICP, best where the addressable market is under roughly 20,000 accounts.
7 steps
Customer success / L5 Autonomous
Support deflection with an audit lane (L5)
Real deflection is available on a small number of high-volume, docs-answerable intents, and it depends entirely on the content underneath since stale documentation ships a confident liar to your customers. The audit lane, a human reading a daily sample from tools like Zendesk, Freshdesk or Intercom Fin and fixing content the same day, is what separates autonomous from reckless. Strongest at 500+ employees or any size with high repetitive ticket volume, requiring a maintained help center and a support function that can staff a daily review.
7 steps
Marketing / L3 Integrated
Lifecycle content engine with a brand gate (L3)
Teams with strong delivery infrastructure usually carry a content backlog measured in quarters, and production is exactly what AI is good at. This puts a voice file in front of generation on Braze, Klaviyo, Iterable, Marketo or HubSpot, ships everything as a test against the incumbent, and caps monthly output at what review can absorb. Fits 100 to 2,000 employees with meaningful lifecycle volume and a small content team, and anywhere with a claims-approval requirement should treat the voice file and review gate as non-optional.
7 steps
Revenue operations / L4 Orchestrated
The agent control plane (L4)
Most teams building agents have no logging, no cost ceiling and no way to tell whether output quality moved. A retry loop in a broken agent can spend a quarter's budget over a weekend, and model providers update silently so a workflow degrades without a line of code changing. This routes every AI call through one gateway such as LiteLLM, Portkey, Cloudflare AI Gateway, Azure AI Foundry, Vertex AI or Bedrock, and is effectively mandatory once more than two AI workflows touch customers or revenue data.
7 steps
Whole company / L3 Integrated
Get shadow AI into the light (L3)
People are already pasting customer data into personal accounts on personal cards, and an ungoverned L1 is more dangerous than no AI at all because exposure is live and visibility is zero. Banning tools drives usage underground, so the fix is amnesty first, then making the sanctioned path on Okta, Entra ID or Google Workspace SSO faster than the shadow one. Fits any company above roughly 50 employees, and is urgent in healthcare, financial services, insurance, legal, defense and education.
7 steps
Enablement / L4 Orchestrated
Point retrieval at the revenue corpus (L4)
Reps re-ask the same handful of questions every week, pricing exceptions, competitor comparisons, security questionnaire answers, and enablement answers them by hand each time. Corpus scope and a freshness gate decide the outcome, both governance rather than engineering, and unowned content is how the system starts lying with confidence. Fits 500+ employees with a sales team large enough that the same question gets asked by different people, and is a weak fit under roughly 20 reps.
7 steps
Whole company / L6 Rebuilt
Rebuild coverage around agents, not headcount (L6)
Every other playbook makes an existing process faster; this one changes the process itself, replacing a coverage model designed years ago for humans only. The tell that you have actually reached L6 rather than dressed up L4 is that the comp plan changed. Fits 2,000+ employees with several AI workflows already stable in production and a CRO with the political capital to redraw comp, and it requires playbooks 4, 5 and 9 already live and stable.
7 steps
Revenue operations / L4 Orchestrated
Score your ICP on stack density (L4)
What a company already runs predicts what it will buy next better than industry or funding stage does, and a prospect running several sales-engagement tools and a modern data platform needs a different first conversation than one running a CRM and nothing else. The step everyone skips is redrawing territories to follow the score, which is exactly why scoring projects die in the spreadsheet they were born in. Fits any size selling B2B software into a technology-buying market, strongest at 50 to 1,000 employees with an addressable market of roughly 5,000 accounts or more.
7 steps
Customer success / L1 Starter
AI first-draft renewal briefs (L1)
Before every renewal call the CSM pastes the account's last three tickets, last QBR notes, and usage summary into the sanctioned assistant and gets a one-page brief: what they bought, what they use, what is broken, what to ask. No integration, no automation, one shared prompt.
4 steps
Customer success / L1 Starter
Tone and clarity rewrite for customer replies (L1)
Support and CS paste their own draft reply into the assistant and ask for a clearer, calmer version before sending. The human still writes the answer. AI only fixes structure, tone, and length. This is the lowest risk way to get a support org comfortable with AI.
4 steps
Customer success / L2 Assisted
Assisted onboarding plan generator (L2)
A shared assistant workspace turns the signed order form and kickoff notes into a 30, 60, 90 day onboarding plan with owners and milestones. The CSM edits rather than starts from a blank template, and every plan comes out in the same shape.
4 steps
Customer success / L2 Assisted
Escalation timeline builder (L2)
When an account goes red, someone spends half a day reconstructing what happened. This play has the assistant assemble a dated timeline from tickets, call notes, and email threads so the escalation meeting starts with facts instead of memory.
4 steps
Customer success / L5 Autonomous
Autonomous onboarding nudge agent (L5)
An agent watches onboarding milestone data and acts without a CSM in the loop for the routine cases: stalled setup, unused seats, missed training. Humans handle exceptions and anything touching commercial terms. Every action is logged and reversible.
4 steps
Customer success / L6 Rebuilt
Agent led support tiering rebuild (L6)
Stop staffing tier one as a human queue. Rebuild the model so agents own resolution for the known corpus, humans own judgment and relationships, and the tiering map is redrawn around what the agent actually resolves instead of ticket volume.
4 steps
Customer success / L6 Rebuilt
Expansion signal desk rebuilt around agents (L6)
Expansion stops being a quarterly CSM guess. A standing signal desk combines product usage, support themes, and contract terms into ranked expansion candidates with the evidence attached, and the coverage model is rebuilt around working that queue.
4 steps
Enablement / L1 Starter
Prompt of the week for reps (L1)
One prompt, one Loom, one channel post, every Monday. The cheapest possible enablement motion for AI adoption, and the one that builds the habit that every later level depends on.
4 steps
Enablement / L1 Starter
Day one AI setup for new hires (L1)
Every new revenue hire leaves week one with the sanctioned assistant provisioned, the notetaker joined to their calendar, the prompt library bookmarked, and the acceptable use rules signed. Adoption is an onboarding problem before it is a change management problem.
4 steps
Enablement / L2 Assisted
Objection library built from real calls (L2)
Stop writing objection handling from imagination. Pull the actual objections out of last quarter's recorded calls, cluster them, and publish the top ten with the responses that correlated with progression.
4 steps
Enablement / L3 Integrated
AI assisted call scorecards wired to coaching (L3)
Every recorded call gets scored against your methodology automatically, scores flow into the coaching system, and managers coach the two lowest scoring behaviours per rep. The model scores, the manager decides.
4 steps
Enablement / L3 Integrated
Retrieval answer desk for reps (L3)
Reps ask product, security, and pricing questions in Slack and get an answer with a source link from your approved corpus, not from model memory. Unanswered questions become the content backlog.
4 steps
Enablement / L4 Orchestrated
Orchestrated competitive response (L4)
When a competitor ships a launch or a price change, an orchestrated workflow drafts the updated battlecard, flags every open deal where that competitor is present, and notifies the owning reps with the changed lines only.
4 steps
Enablement / L5 Autonomous
Autonomous ramp plan adjustment (L5)
New hire ramp plans adjust themselves. The agent reads certification results, call scores, and pipeline creation, then reorders the next two weeks of ramp work. Managers approve exceptions and own the final readiness call.
4 steps
Enablement / L5 Autonomous
Live skills graph for the revenue org (L5)
Instead of an annual skills survey, maintain a live graph of demonstrated skill per rep built from call scores, certifications, deal outcomes, and content usage. Enablement plans against evidence rather than opinion.
4 steps
Enablement / L6 Rebuilt
Content supply chain rebuilt for retrieval (L6)
Rebuild enablement content as a maintained corpus with owners, expiry dates, and machine readable structure, because agents and reps both read it now. Decks become an output of the corpus, not the corpus itself.
4 steps
Enablement / L6 Rebuilt
Frontline manager operating system (L6)
Rebuild the frontline manager week around AI produced evidence. Deal reviews, coaching, and forecast prep all run from the same assembled evidence pack, and manager time shifts from assembly to judgment.
4 steps
Every team / L1 Starter
Summarize before you forward (L1)
Nobody reads the 40 page vendor PDF, the board pack, or the long thread. The rule is simple: anything long you forward gets a five bullet summary and a stated ask on top. One prompt, every team, no tooling.
4 steps
Every team / L3 Integrated
One answer layer over internal systems (L3)
Employees stop guessing which system holds the answer. A retrieval layer indexes the docs, tickets, and wikis people are allowed to see and answers with citations and permissions respected.
4 steps
Every team / L4 Orchestrated
Orchestrated cross team handoffs (L4)
The expensive failures live between teams: sales to onboarding, support to product, marketing to sales. An orchestration layer moves the work with its context attached and enforces the service level on every handoff.
4 steps
Every team / L5 Autonomous
Autonomous back office agents with audit (L5)
Routine internal work runs without a human in the loop: vendor renewal reminders, access reviews, expense policy checks, document filing. Every action is logged, bounded, and reversible, and exceptions route to a named human.
4 steps
Every team / L5 Autonomous
Agent spend control across the company (L5)
Once agents run continuously, consumption becomes a real line item. This play puts every AI workload on a named owner, a budget, and an alert, so cost is managed before finance discovers it in a quarterly true up.
4 steps
Every team / L6 Rebuilt
Company operating policy rebuilt for agents (L6)
Policy stops treating AI as a tool question and starts treating agents as actors. Access, approval, retention, and accountability rules are rewritten so an agent taking an action has a named human accountable for it.
4 steps
Every team / L6 Rebuilt
Planning cycle rebuilt around live evidence (L6)
Annual and quarterly planning stops running on hand built decks. Plans are assembled from live system evidence, assumptions are written as testable statements, and the plan is re-scored monthly against what actually happened.
4 steps
Marketing / L1 Starter
AI outlines from an approved brief (L1)
Writers stop starting from blank. Paste the approved brief, the ICP definition, and two reference pieces, and get an outline with a claim list. Writing stays human. Structure gets faster.
4 steps
Marketing / L1 Starter
Weekly campaign recap in one page (L1)
Every Monday the channel owners paste their platform exports into a shared prompt and get one page: what moved, what dropped, what changed, what to do this week. It replaces four dashboards nobody reads.
4 steps
Marketing / L2 Assisted
Assisted repurposing pipeline (L2)
One source asset becomes the social, email, and enablement versions through a shared assistant workspace loaded with the voice file. Humans approve every output. The gain is coverage, not volume for its own sake.
4 steps
Marketing / L2 Assisted
Assisted inbound reply triage (L2)
Form fills, chat transcripts, and inbox replies get classified and summarized before a human touches them, so real buying questions do not sit behind vendor pitches and job applications.
4 steps
Marketing / L4 Orchestrated
Orchestrated launch runbook (L4)
Launches run from an orchestrated runbook: assets generated from one source of truth, tasks created with owners and dates, readiness checked automatically, and a single dashboard showing what is blocking go live.
4 steps
Marketing / L5 Autonomous
Autonomous experiment engine (L5)
Creative and landing page variants are generated, launched, measured, and retired without a human approving each cycle, inside strict brand and claim guardrails. Humans set the guardrails and read the results.
4 steps
Marketing / L5 Autonomous
Autonomous lifecycle orchestration (L5)
Lifecycle messaging decides its own next best message per contact within approved content, frequency caps, and suppression rules. Humans own the content library, the caps, and the audit.
4 steps
Marketing / L6 Rebuilt
Marketing rebuilt around buyer questions (L6)
Retire the campaign calendar as the organizing unit. Marketing is organized around the real questions buyers ask, each with an owner, an answer surface, and a measurement, because humans and answer engines both retrieve answers now.
4 steps
Marketing / L6 Rebuilt
Evidence first content operation (L6)
Every published claim carries a source, a date, and an owner, enforced at publish time. Content becomes a maintained evidence base that agents and buyers can cite rather than a stream of undated opinion.
4 steps
Whole company / L1 Starter
One page acceptable use rules (L1)
Before any rollout, publish one page everyone can actually read: which tools are sanctioned, what data may be pasted, what must never be pasted, and who to ask. Most shadow AI is a documentation failure, not a discipline failure.
4 steps
Whole company / L1 Starter
Open intake list for AI ideas (L1)
Collect every AI idea in one visible list with the problem, the owner, and the current manual cost. No committee, no scoring model, just visibility so duplicate pilots stop happening in four teams at once.
4 steps
Whole company / L3 Integrated
AI risk register with real review (L3)
Every AI workload that touches customer data, money, or public output gets a register entry with its failure modes, its blast radius, and its rollback. Review is scheduled, not incident driven.
4 steps
Whole company / L4 Orchestrated
Standard AI vendor evaluation (L4)
One evaluation standard for every AI vendor: data handling, meter and pricing mechanics, retrieval quality on your own content, exit path, and roadmap dependency. Orchestrated so procurement, security, and the business team run it in parallel rather than in sequence.
4 steps
Whole company / L4 Orchestrated
Orchestrated AI change cadence (L4)
Model updates, prompt changes, and workflow changes ship on a cadence with release notes, owners, and a rollback, the same as software. Silent changes are the leading cause of broken trust in AI workflows.
4 steps
Whole company / L5 Autonomous
AI incident response runbook (L5)
When an agent sends the wrong thing to a customer or leaks internal content, the clock is minutes. This is the runbook: detect, contain, notify, correct, review, with named owners and pre written customer language.
4 steps
Whole company / L5 Autonomous
Workforce planning with agent capacity (L5)
Capacity planning includes agent throughput as a modelled input rather than a hiring freeze rumour. Leadership sees the tradeoff between headcount, agent capacity, and service level with the assumptions written down.
4 steps
Revenue operations / L1 Starter
AI assisted pipeline hygiene checklist (L1)
Before every forecast call, RevOps runs a shared prompt over the exported pipeline to list the records failing basic hygiene: no next step, stale close date, missing contact roles, value that changed without a note.
4 steps
Revenue operations / L1 Starter
Turn ops requests into written requirements (L1)
RevOps requests arrive as one line Slack messages. The assistant turns each into a structured requirement with the problem, affected process, requested change, and impacted systems, so triage happens on facts.
4 steps
Revenue operations / L2 Assisted
Assisted reporting question desk (L2)
Leaders ask their reporting questions in one channel. RevOps answers with the number, the definition used, and the source, assisted by a saved prompt. It kills the analyst interruption cycle and the metric definition drift.
4 steps
Revenue operations / L6 Rebuilt
Revenue operating model rebuilt for agents (L6)
Process, data model, and roles are redesigned on the assumption that agents perform the routine work. Stages, fields, and approvals are rebuilt around machine readable state rather than around what humans could be bothered to update.
4 steps
Revenue operations / L6 Rebuilt
Continuous forecast rebuilt on evidence (L6)
The weekly forecast ritual is replaced with a continuously updated forecast that combines rep commit, model output, and conversation evidence, with variance attributed to a cause every month.
4 steps
Sales / L5 Autonomous
Autonomous post meeting follow up (L5)
After every customer call, an agent sends the recap and next step email, updates the CRM, and books the agreed next meeting, without waiting for the rep. Reps review a queue rather than authoring each message.
4 steps
Sales / L5 Autonomous
Autonomous sourcing inside a bounded list (L5)
An agent works a defined account list end to end: research, first touch, follow ups, and meeting booking, inside strict messaging, frequency, and suppression rules. Reps take over at the reply.
4 steps
Sales / L6 Rebuilt
Territory and quota rebuilt for agent leverage (L6)
If agents carry sourcing and admin, the old coverage math is wrong. Territories, quotas, and comp are rebuilt around the work humans still do, with the assumptions published so reps can see the logic.
4 steps
Sales / L6 Rebuilt
Evidence standard for every deal (L6)
Deal reviews stop running on rep narrative. Every forecasted deal carries machine verifiable evidence: named economic buyer, documented compelling event, multithread count, and a mutual plan with dates.
4 steps