AI isn’t a side project—it’s the core of modern work. Most organizations already have licenses and pilots, yet value stalls because AI is treated like a plug-and-play tool instead of a workflow shift with disciplined change management behind it. Durable results show up when adoption, learning loops, and behavior change are built into how people actually work—and when everything ladders to a single growth goal.

I’ve proven the pattern in a tiny creative project and in enterprise programs. For the music experiment below, I used an AI micro-stack—LLM lyrics → AI composition → Midjourney/Adobe visuals → Canva edits → automated mastering/distribution—and cut time and cost by ~80-90% versus a traditional approach. The point isn’t music; it’s that the same stack logic is industry agnostic. Whether you’re in entertainment, marketing, training, operations, support, or R&D, the recipe is the same: draft fast, iterate with feedback, finish consistently, and wire what you learn back into the system with change management to make it stick.

Step 0 (the step most teams skip): Pick One Growth Lever

Before touching any technology, ask: What business outcome are we trying to achieve? Winners anchor AI to one concrete growth lever; dabblers start with prompts, tools, and licenses—and wonder why nothing changes.

Choose one:

• Land & Expand — grow within existing accounts
• Segment Expansion — grow into new ICPs or industries
• Upsell/Cross-Sell — grow wallet share per customer
• Pricing/Packaging — grow value per deal
• Channel/Partnerships — grow through partners
• Geo Expansion — grow across regions
• Operational Excellence — grow by making the engine faster and cleaner

Define a single KPI for the lever (e.g., expansion ARR, revenue per rep, cycle time, win-rate in target ICPs) and measure relentlessly. Change management starts here with clear sponsorship and success criteria.

Phase 1: Map Capabilities (not tools) to your lever

Example: Operational Excellence

• Rep productivity → AI-drafted outreach, summaries, and follow-ups; retrieval over your knowledge for instant context
• Conversion optimization → stage-specific assist (objection handling, value proof), deal-quality checks
• Pipeline coverage → automated hygiene and insights to reduce rework and surprises

This is capabilities-first, tools-second. Your stack should fit the work. Pair every capability with an adoption plan: communications, role-based enablement, and manager coaching.

Phase 2: Build the War Room (where change management lives)

Create a Center of Excellence owned by the growth-lever leader, not IT. Make adoption unavoidable.

• Growth Lever Owner (P&L accountability)
• Ops Lead for that function (process design and instrumentation)
• Top Performer(s) (peer coaching; sets “what good looks like”)
• IT/Data (enablement, security, governance)
• Knowledge/Content Lead (keeps the knowledge supply chain healthy)
• Change Network (front-line managers/power users driving behavior change)

Meet weekly. Review lever metrics, unblock adoption, update process/content, and publish a change log so people see what’s different and why.

Phase 3: Run Lever-Focused Pilots (30-day sprints)

Target one measurable lift that your lever cares about.

Sprint example (Operational Excellence):

• Week 1: Diagnose the biggest productivity leak (e.g., discovery notes, follow-up quality, deal hygiene)
• Week 2: Ship the smallest AI-assisted fix (retrieval over knowledge + macros/templates + tasking)
• Week 3: Measure the gain (cycle time, conversion at stage X, error rates)
• Week 4: Document, templatize, and scale to similar workflows

Failed pilot? Good—you learned quickly. Successful pilot? Scale to the entire lever before starting another. Change management tracks adoption, reinforces behaviors, and removes blockers as you scale.

Phase 4: Measure Outcomes (not vanity metrics)

Track lever-specific KPIs, not tool usage:

• Operational Excellence: revenue per rep, cycle compression, forecast accuracy
• Land & Expand: expansion ARR, multi-product penetration, renewal uplift
• Upsell/Cross-Sell: attach rates, expansion win rate, average products per account
• Channel/Partnerships: sourced pipeline, partner win rates, time-to-first-deal
• Marketing/Creative: pipeline contribution, speed-to-first-asset, performance in target ICPs
• Training/Enablement: time-to-proficiency, completion + effectiveness, defect/rework reduction
• R&D/Product: concept-to-prototype time, experiment throughput, validated learning per sprint

If it doesn’t move the lever, it’s noise. Your change plan should reward outcomes, not ‘AI usage’.

How the playbook applies across industries and functions

• Entertainment & Media: Rapid concepting (scripts, visuals, audio), iterate with audience feedback, standardize finishing moves (mix/master, QC). Change management aligns creative leads and producers on “what good looks like” and version control.
• Marketing: AI-assisted briefs, creative variants, and performance insights tied to one ICP. Change network ensures brand guardrails and review cadence.
• Training & L&D: Script with LLMs, render short updates with AI video, attach to in-flow moments (LMS, help hubs). Change plan equips managers to coach to the new materials.
• Operations: Copilots for procedures and checklists; micro-training at points of work; instrumentation to cut rework/defects. Change management handles SOP updates and certification.
• Support & CX: Retrieval-augmented chat over a curated knowledge base + agent assist; auto-summaries to reduce wrap time; Evolve Loop to improve content with every interaction. Managers coach to behaviors, not just handle time.
• R&D & Product: Assisted research, ideation, and simulation; value comes from clean data and closed-loop learning. Change management formalizes how experiments are proposed, reviewed, and productized.

Across all of the above, change management is the backbone—visible sponsorship, a trained change network, role-based enablement, incentives tied to outcomes, and adoption treated as a deliverable with owners and dates.

30-Day Quick Start

• Week 1: Pick one growth lever + one KPI; name an executive sponsor
• Week 2: Map required capabilities; compose a minimal stack to the work; plan enablement
• Week 3: Stand up the War Room; publish cadence, definitions of done, and the change plan
• Week 4: Launch the first pilot, instrument it, document what worked (and what didn’t), and communicate results

Then repeat—same lever, next adjacent workflow—until the lever moves at scale.

What “Good” Looks Like

• AI sits inside the flow of work (not in a separate tab)
• Knowledge is the substrate: capture → structure → publish → find → apply → learn
• Leaders manage behavior change, not just licenses
• Dashboards show cause, not just count, and tie directly to the growth lever
• Teams compose their workflows: draft fast, finish consistently, wire learning back in with a living change log

If you try one thing this week, take a recurring deliverable (brief, onboarding snippet, campaign draft, customer update, or top support article), draft it with AI, publish it, route the same source into search/chat/LMS, and measure the delta. Repeat next week. The stack matters, but the system—the growth lever, workflow design, and change discipline—is what compounds.

---

Music Stack

• Lyrics: @OpenAI
• Music: @SonautoAI
• Mastering: @DistroKid
• Images: @midjourney and @adobe
• Video: @canva

References

• McKinsey & Company. The state of AI: How organizations are rewiring to capture value (2025). 71% of organizations regularly use gen AI; enterprise-level EBIT impact remains limited; functions with the most adoption include marketing/sales, service ops, product development, software engineering, and knowledge management. McKinsey & Company
• McKinsey & Company. AI in the workplace: A report for 2025 (Jan 2025). Only ~1% of C-suite respondents describe gen-AI rollouts as “mature”; leadership underestimates employee usage; skills gaps cited by 46% of leaders. McKinsey & Company
• Brynjolfsson, E., Li, D., & Raymond, L. “Generative AI at Work.” NBER Working Paper 31161 (2023). Field evidence from a contact center shows ~14–15% productivity gains on average, with the largest gains for newer agents. NBER
• Instacart. “Bringing inspirational, AI-powered search to the Instacart app with Ask Instacart.” Company announcement (May 31, 2023). Instacart
• Boston Consulting Group (BCG). Where’s the Value in AI? (Oct 2024). Over the past three years, AI leaders achieved ~1.5× revenue growth vs. peers, with stronger TSR and ROIC. BCG Media Publications
• Gomez-Uribe, C. A., & Hunt, N. “The Netflix Recommender System: Algorithms, Business Value, and Innovation.” ACM TMIS 6(4), 2015. Recommender influences choice for ~80% of streamed hours. Ailab UA
• Prosci. “The Correlation Between Change Management and Project Success” (June 7, 2023). Projects with excellent change management are ~7× more likely to meet objectives than those with poor change management. Prosci
• Sonauto (YC). Company page: “Sonauto is an AI music editor that turns prompts, lyrics, or melodies into full songs.” Y Combinator
• DistroKid Mixea. Product site & support: Unlimited online mastering for a flat annual fee; “first track free, then $99/year” plan. MixeaDistroKid Help Center
• Accenture. Why Artificial Intelligence Is the Future of Growth (2016). Long-run macro estimate that AI could boost labor productivity by up to ~40% by 2035 (forward-looking scenario). ICDST+1

Note: A July–Aug 2025 MIT Media Lab NANDA report (“The GenAI Divide: State of AI in Business 2025”) has been widely reported as finding ~95% of enterprise gen-AI pilots lack measurable ROI. It’s a fresh study and still being debated; preliminary versions and coverage are available. AI NewsVirtualization Reviewfortune.com