Leading in the Age of AI
AI stopped being a chat window. It became a coworker. What that means for how you lead.
The coworker shift
AI stopped being a chat window and became a coworker. That shift changes what leadership requires now.
- 01The New Reality
Why the payoff hides, and why the work multiplies.
- 02Working With AI
You bracket the work. The AI does the middle.
- 03The Toolbox, Live
Models, modes, connectors, skills, schedules. Live.
- 04Guardrails
Your data, your policy, the trust threshold.
- 05The Adoption Map
Five stages. Where your organization actually is.
- 06The Decision
Which leader will you be a year from now?
The New Reality
Why the payoff hides, and why the work multiplies.
You bought the licenses. Where's the payoff?
Every leader in this room has asked some version of that question this year. Here is the uncomfortable news: we have been exactly here before.
Robert Solow, Nobel laureate in economics, 1987
Today you can see the AI age everywhere but in the productivity statistics. Closing that gap is the job in front of you right now.
Everyone bought it. Few are banking it.
McKinsey, late 2025: adoption is nearly universal. Meaningful bottom-line impact is rare.
The efficiency paradox
William Stanley Jevons, 1865: cheaper coal power did not reduce coal use, it exploded it. The same economics now apply to knowledge work.
- 01Efficiency
AI makes tasks faster and cheaper.
- 02Lower cost
Work that was not worth doing suddenly is.
- 03More experiments
More drafts, more analyses, more bets.
- 04More work
Demand for output expands to fill the new capacity.
The ROI equation flips
The old world: maximize Return
Big projects, big budgets, long paybacks.
High cost of trying means few bets.
You needed certainty before you started.
The AI world: minimize Investment
The cost of trying collapsed.
Run ten cheap experiments, keep the three that work.
Speed of learning beats certainty of planning.
When experiments get cheap, the advantage goes to whoever runs the most of them.
Canada's gap is the opportunity
A 5x jump in under a decade. The country is betting on it.
44th of 47
on AI training and literacy
KPMG / University of Melbourne, 2025
$200B
growth targeted by the national AI strategy
AI for All, 2026
250,000
AI jobs targeted over five years
AI for All, 2026
The computer was never the productivity gain. The reorganization of work around it was.
The same is true now. You are not waiting on a better model. You are deciding to change how the work gets done.
Working With AI
You bracket the work. The AI does the middle.
Treat it like your sharpest new hire
Brilliant at some things. Hopeless at others. Lean on it here, double-check it there.
Lean on it for
First drafts of anything: documents, plans, analyses.
Options when you are stuck staring at a blank page.
Summarizing the long thing you did not have time to read.
Keep your hands on
Facts and numbers: it can sound confident and be wrong. Verify.
Anything with your name on it. You own the output.
The final call. Judgment does not delegate.
The quality of what you get back tracks the quality of the brief you give.
You bracket the work. AI does the middle.
The skill is not handing everything to the AI. It is knowing which part is yours. You frame the work and you verify the result. The AI carries the load in between.
The shape of a strong ask: Role, Context, Task
A prompt structure is a repeatable shape you fill in. It forces you to name what the AI needs up front, so it stops guessing and answers what you actually meant. Role, Context, Task is the one to learn first.
Context is what makes it sharp.
Same question, three layers. Watch the answer climb as you give it a role, then the real specifics, then the exact output you want.
- 01Role
Give it a perspective
- 02Context
Hand it the specifics
- 03Task
Name the exact output
Make it argue with you
The first draft is rarely the best one. Before you accept anything, make it attack its own work: "What is weak here? What would a tough critic say? Now fix it."
AI stopped being a chat window. It became a coworker.
What comes next is live. Watch how much of the work you can actually hand over.
The Toolbox, Live
Models, modes, connectors, skills, and schedules. Shown for real, on stage.
Claude-focused, not Claude-exclusive
I will demo Claude, because it is the tool we run our own company on. The other major platforms have an equivalent for nearly everything you are about to see.
The brands change and the buttons move. The habits do not: brief it well, pick the right model, connect it to your work, save what works, and verify before you ship. Learn the habits once, use them anywhere.
Pick the right vehicle for the trip
Every AI product offers a family of models. You would not take the Ferrari to get groceries. Claude's lineup, in car terms:
Haiku
High-volume, simple tasks and automations.
Fastest, cheapest
Sonnet
Your default: writing, analysis, most work.
Fast, balanced
Opus
Genuinely hard reasoning and big jobs.
Slower, pricey
Fable
The hardest, novel problems.
Specialized, top cost
Default to the Civic. Reach for the Ferrari only when the job is genuinely hard. Every AI product has the same ladder, whatever it names the rungs.
Your usage is a budget. Spend it well.
The 5-hour window: starts at your first message, then resets.
The weekly limit: resets every week.
What burns it fastest- 01One giant chat.
Every message re-sends the whole thread. Start a new chat per job.
- 02The Ferrari for groceries.
Reaching for the top model when the default would do.
- 03Heavy attachments
riding along in every message that follows.
Schedule your morning brief for 6:45am. The automation's first message opens your first 5-hour window while you sleep in, so it resets before noon and your peak working hours get a fresh one. Your usage, working your schedule.
Chat or Cowork? Pick the right mode
Same brain, two working styles. The toggle at the top is the most underused control in the product.
A conversation. You stay in the loop.
Quick questions, summaries, and explanations.
Drafting and editing where you want to steer each round.
Thinking out loud: pressure-testing an idea before a meeting.
A delegation. It works, you review.
Multi-step jobs: research, build, and assemble in one run.
Anything touching files: documents, spreadsheets, decks, folders.
Long tasks you would rather not babysit. Set it going, come back.
Rule of thumb: if you would forward it to a capable assistant, it belongs in Cowork.
Your new thought partner
A real question from a real week, with my calendar and inbox connected.
WHAT TO WATCH FOR- 01It pulls context from the calendar and inbox before answering, because I let it.
- 02I make it interview me instead of guessing what I want.
- 03I ask it to push back on my plan, not to agree with me.
Custom instructions gave it memory. Connections give it reach.
Once connected, the AI can read and act where you already work, with your permission.
- 01Email
Triage the inbox, draft replies in context, summarize a long thread.
- 02Calendar
See your week, find time, prep you for the next meeting.
- 03Documents
Pull facts from a real doc instead of you pasting it in.
- 04Files & Drive
Reach into the folder where the work actually lives.
- 05The web
Look something up live instead of guessing from memory.
- 06Your other tools
Notion, your CRM, Slack: where your team already operates.
Cowork does the work
I hand it a folder of messy meeting notes. It hands me back a finished deliverable.
WHAT TO WATCH FOR- 01It plans before it works, and shows me the plan.
- 02It reads the files and does the whole job in one run, while we talk.
- 03It delivers a finished document, not a wall of chat text.
A skill is a saved playbook
You know how a great onboarding doc means a new hire does the task right without you hovering? A skill is that doc, but the AI reads it and does the work.
Packaged once
Your format, your rules, your standard for good, written down one time.
Runs the same way
Same structure, same quality, every single time it fires.
Real deliverables
Skills produce finished files: a slide deck, a spreadsheet, a document, a PDF.
Skills are how one person's best week becomes the whole team's every week. The standard stops living in someone's head.
Teach it once. Put it on a schedule.
The step most teams never take. A skill that runs on a timer is work that does itself.
- 01Teach it
One good example of the finished product is the lesson.
- 02Save it
The playbook becomes a skill your whole team can run.
- 03Schedule it
Monday 6:45am, every week, without anyone remembering to ask.
The weekly recap is in your inbox before you pour the coffee. Nobody wrote it. Nobody forgot it.
Teach it once, schedule it forever
We save what Cowork just did as a skill, then put it on the calendar.
WHAT TO WATCH FOR- 01The finished example becomes the teaching material: this is the standard, match it.
- 02One correction sticks: I fix it once and the skill remembers.
- 03Scheduled for 6:45 every Monday. This job no longer has a human in it.
Speed is not the finish line
The fastest way to lose the room is to ship AI output nobody checked. The pros work a loop:
Instruct
A real brief: role, context, the standard for good.
Plan
Make it show its plan before it starts the work.
Inspect
Read what it made like you would a new hire's first draft.
Verify
Check the facts and numbers before they leave the building.
You own the output. If you would not put your name on it, it is not done.
Guardrails
Your data, your policy, and the trust threshold.
Where your data goes
- 01On business plans, your work is not training the model
Claude Team, ChatGPT Business, Copilot: on proper business plans your conversations stay in your workspace. On free personal accounts, assume otherwise.
- 02Still, use good judgment with sensitive data
Follow your data and privacy policies, and think twice before pasting personal, customer, or financial details when the task does not truly need them.
If you would not put it in an email to someone outside the company, pause before putting it into any tool.
One page of policy prevents the expensive surprise
- 01Approved tools.
Which AI tools are sanctioned, and how new ones get vetted.
- 02Data rules.
What can and cannot be put into AI. Client, financial, regulated data.
- 03Who owns it.
The implementation lead, and how to raise a question.
- 04The roadmap.
Where you are headed over the next few quarters.
63%
of breached organizations had no AI governance policy
IBM Cost of a Data Breach, 2025
$670K
added to the average cost of a breach when shadow AI is involved
IBM Cost of a Data Breach, 2025
Agents cross a trust threshold
Up to now, people drove the tools. Agents do delegated work. That is not a technology upgrade, it is a trust threshold.
- 01Start where mistakes are cheap and reversible.
- 02A human approves anything that leaves the building or spends money.
- 03Log everything, so trust is earned on evidence.
- 04Widen autonomy as the track record earns it.
Manage it like a capable new hire. Nobody gives a new hire the company credit card on day one.
The Adoption Map
Five stages. Where your organization actually is, and what moves you.
The Work Smarter AI Adoption Map
Every organization is somewhere on this map. Most are earlier than they think. Stage is about the organization, not the calendar: what moves you is deliberate work in the right order.
Why stages get skipped, and why it fails
Buying build tools before building capability. The tools land on an untrained base with no governance, and they die as shelfware. Or worse, as a security incident.
Skills before builders
Builders emerge from a trained base. Champion programs launched into an untrained org select for hobbyists, not the people closest to the valuable workflows.
Guardrails alongside, never after
The use policy lands on day one. Governance bolted on later arrives as a crackdown, and crackdowns kill momentum.
Measure from day one
The CFO's question arrives around month six. The answer has to start in month one: who uses what, and hours saved.
The map exists to make the next stage the ambition, not the last one.
The Decision
What you do with this decides what the next year looks like.
- 01The cost of intelligence just collapsed.
- 02So the work doesn't shrink. It multiplies. Your competitors are already running the experiments that weren't worth it last year.
- 03The gap between teams that built this capability and teams that didn't is widening every quarter. And it compounds.
- Shadow AI everywhere, no rules
- Best people quietly leaving
- Chasing a target that moved
- Trained team, real guardrails
- Work shipping every week
- Setting the pace for the market
Tomorrow morning, one move
Open the tool. Say: "Interview me, then set up my profile so you always know who I am and what I care about." Then ask it one real question from your actual week.
Ten minutes. No committee, no platform decision, no strategy document. That is the whole assignment.
You cannot lead a change you have not lived. The strategy is the output of starting, not the prerequisite.
Keep the whole deck
Every tool I showed you today is a full training session on our site, including this deck. Free, no email wall.
worksmartercompany.com/resources
Start a conversation
Thirty minutes, free. Tell me where your organization is on the map, and I will tell you the highest-leverage moves. No deck, no pitch.
worksmartercompany.com/book
Colin Cox · The Work Smarter Company · colin@worksmartercompany.com
You're at the trailhead.
Now lead.