Training at the right altitude: execs, managers, ICs

Train everyone the same way and you lose everyone's time. Different audiences need different AI training. Here's the audience map and what each actually needs.

The audience tiers

1. Executives and senior leaders. Need to understand strategy and implications.
2. Managers and team leads. Need to enable and coach their teams.
3. Individual contributors. Need hands-on skill to use tools productively.
4. Specialists (engineers, designers, analysts). Need deep fluency in role-specific AI.
5. Support functions (HR, legal, finance). Need to understand AI impact on their domain.

Each gets different content. Each gets different time.

Executives: 1-2 hours total

What they need:

• Capability map. What AI can and can't do, in plain terms with examples.
• Organizational implications. How AI reshapes roles, workflows, and strategy.
• Governance landscape. Risks and how responsible organizations manage them.
• Decision frameworks. How to evaluate AI investments.

What they don't need:

• Prompt engineering tutorials.
• Tool-by-tool feature demos.
• Hands-on use cases.

Format: executive briefing, 90 minutes. Optional follow-ups with specific deep dives.

Managers: 3-5 hours

What they need:

• Everything from the executive briefing, compressed.
• Coaching skills. How to help their team members adopt AI.
• Change management. How to navigate resistance.
• Measurement. What to look for as signs of real vs. superficial adoption.
• Their own use cases. Drafting emails, summarizing meetings, reviewing work.

Format: half-day workshop, follow-up office hours.

Common failure: giving managers the same training as ICs. Managers need to lead adoption; they're the key leverage point.

ICs: 2-3 hours for basics, ongoing for depth

What they need:

• Specific workflows. Not "here's a tool" — "here's how your work changes."
• Hands-on practice. With real tasks from their role.
• Pitfall awareness. When AI fails and how to catch it.
• Where to get help. Internal champions, support paths.

Format: small-group workshops, cohorts of 8-15. Follow-up practice assignments. Ongoing office hours.

Specialists: 10+ hours, role-specific

Engineers need coding-tool deep dives. Designers need creative-AI workflows. Analysts need research-tool fluency.

Generic AI training doesn't cut it for people whose daily work involves AI as a core tool. They need deep, role-specific investment.

Format: cohort training, internal or external specialists. Ongoing communities of practice.

Support functions: 4-6 hours

HR, legal, finance, IT need training specific to their role:

• HR: how AI affects hiring, performance, compensation conversations.
• Legal: AI contracts, data handling, IP, regulatory landscape.
• Finance: AI procurement, ROI evaluation, budget implications.
• IT: AI architecture, security, operations.

Specialist training; don't lump them into general sessions.

Training formats that work

• Cohort-based, small group. 8-15 people, multiple sessions. Peer learning valuable.
• Hands-on, real tasks. Practice with actual work artifacts, not toy examples.
• Follow-up practice. One session doesn't stick. 3-4 sessions over 2 months better.
• Office hours. Ongoing support for questions.
• Peer communities. Slack channels, internal forums where people share.

Training formats that don't

• One-hour webinar for 500 people. Watched, forgotten.
• Self-paced videos, no follow-up. Low completion, no practice.
• Generic prompt engineering for everyone. Too specialized for most roles.
• Annual required training. Compliance, not capability.

The "executives must use it" point

No amount of training substitutes for executives visibly using AI themselves. Leadership that treats AI as "a thing my team should learn" while not using it signals: this is optional, not strategic.

Expect executives to adopt the tools. Make the training shorter if needed, but make it stick.

Measuring training effectiveness

Not: "training completed by X% of org." That measures seat-time, not skill.

Better:

• Application rate. 60 days after training, what percent of attendees are using the tools in measurable ways?
• Competency assessment. Quick task-based evaluations before and after.
• Retention surveys. "What from the training do you still use?" 90 days later.

If application rate is <40%, the training isn't working. Iterate.

The sequencing

Common sequence:

1. Leadership first. Executives trained; expected to use tools themselves.
2. Managers second. Trained to enable teams.
3. Pilot team ICs third. Deep, hands-on, cohort-based.
4. Broader rollout. Adjusted based on what early groups taught.

Training leaders after ICs is backwards: you lose top cover when ICs encounter friction and need advocacy.

The cost

Real AI training at scale is expensive:

• $500-2,000 per person for serious cohort-based training at external providers.
• Internal training: 100-300 person-hours of trainer time per 100 trainees.
• Plus the time for trainees themselves.

If you're planning AI transformation, budget training at 15-25% of total program cost. Orgs that underspend here underperform.