5 steps to make gen AI deliver value
Gen AI pilots are easy; building enterprise value requires a workforce and operating model built around AI as a capability, not a tool. I outline a five-step change playbook: set a North Star, build trust with data and governance, reimagine workflows, redesign structures, and empower people at scale. The prize is a blended human–AI organization where agents handle the grind and employees shift to higher‑value work.
Points clés
- Two-thirds of global companies already use gen AI, yet employees are using it three times more than leaders realize, underscoring the gap between experimentation and enterprise value.
- Step 1 centers on a clear North Star focused on outcomes, plotting a path from discrete agents to agent swarms and minimum viable organizations (MVOs) while acknowledging uneven rates of change.
- “Gen AI high performers” attribute at least 10% of EBITDA to gen AI and are more likely to invest in trust; companies that build AI/digital trust are nearly 2x more likely to see 10%+ revenue growth.
- Morgan Stanley trained an OpenAI-powered assistant on 100,000+ research reports and, after rigorous evaluation and guardrails, achieved 98% adoption among wealth management teams.
- Step 3 reframes gen AI as a capability within workflows, evolving from stand‑alone agents to agent groups and then agentic swarms; 48% of US employees would use gen AI more with formal training, and 45% if tools are embedded in daily workflows.
- McKinsey’s internal platform Lilli has been used by 92% of staff, with 74% using it regularly; it saves 30%+ time on information gathering and synthesis and has answered nearly 19 million prompts.
- Step 4 introduces mixed operating models: MVOs for repetitive, logic-based processes (for example, invoice processing) alongside augmented human teams in high-touch functions; new roles emerge such as AI workflow optimizer and automation product owner.
- Step 5 stresses participation at scale: involving at least 7% of employees in transformation initiatives doubles the odds of positive excess TSR, compared with the 2% typically involved.
- Adoption skews middle-out: 62% of millennial managers report high AI expertise versus 50% of Gen Z (18–24) and 22% of baby boomers (65+), making millennials potent change champions.
- Singtel launched an AI Acceleration Academy with NTU and NUS to train 10,000+ employees in data and gen AI skills, exemplifying a skills-first rollout.
À retenir
Start by picking a simple North Star (no, “deploy a chatbot” won’t save Q4). Open the data taps with real governance, rewire one end-to-end workflow where value is obvious, and train people like you actually want them to use the tools. Recruit millennial superusers, keep humans in the loop where customers matter, and let agents chew through the boring bits—because sarcasm aside, the fastest ROI here is making drudge work disappear.
Sources
