Get a team that runs on agentic AI.
The companies that build their agentic operating model this year grow. The ones that don't shrink.
What I do is very simple: I teach your team how to build with AI. It is more like a college class than done-for-you consulting. It takes four weeks, and each week gets progressively more technical.
Email ✷ NowGot a question? email grace@grace-clarke.com
The process ✶
Each week builds on the last: from your first skill, to your own automations, to systems your whole team shares.
Customize your training
- Founder calls and 1:1s so I design the training around how your team actually works.
- I draft three candidate builds per department. Your copywriter doesn't get "build a skill" - she gets her voice guide, her brief formatter, or her caption batcher. Pick one.
A curriculum built for your team and a build menu per department. Everyone preps a short doc about their role before day one.
Onboarding + your first builds
- I present the fundamentals: chatbot vs. agentic, and why delegation beats prompting.
- I onboard everyone live: skills, projects, scheduled tasks, connectors, memory.
- Everyone runs their first agentic prompts and builds their first skill, chosen from their prep doc.
Every person has built something for their own role - no technical gatekeeper. Plus 15 use-case ideas per person, personalized by me.
Skills + automations
- I present scheduled tasks and projects.
- Everyone writes the skill behind their first scheduled task and schedules it.
- Everyone sets up a project with real instructions and learns how to decide what's worth building.
A running automation per person and a personal build backlog. Everyone majors in their craft, minors in AI.
Team projects + shared skills
- I demo team-wide uses, starting with connectors into the tools you already run on (Google Sheets first).
- Each person builds a skill encoding their expertise: the copywriter's voice, the analyst's rubric, the designer's specs, the strategist's intake.
- We assemble them into a shared project the whole team can talk to.
Institutional knowledge, documented and usable by everyone. The hottest programming language is English.
Demo day
- Every person demos one thing they built. Learn-it-all over know-it-all.
- The team picks a ritual to keep building: a dedicated Slack channel, mini demos inside meetings you already have.
A team that teaches itself, and a ritual on the calendar instead of a workshop memory.
You build. I keep you current.
- You get the living deck: everything from the training, kept current as the models change.
- A monthly update from me on what's new and worth your attention.
A team that keeps building long after the training ends.
The thread through all four weeks: give context, iterate, encode what works.
My approach ✶
I teach.
This is a teaching engagement, not an engagement where I build for you. My goal is your team's fluency - not a deck, not custom infrastructure, not fractional leadership, not a roadmap. (Even Anthropic and OpenAI avoid roadmaps more than 60 days out - the models change too fast.)
My belief is any team needs to be able to build their own skills, stack those skills into workflows, and create their own connectors. That is a more valuable outcome than infrastructure you can use but never learn from.
If you want someone who comes in and sets it all up for you, that work exists and it's good. I can refer you to the people who do it.
This may be for you if ✶
Scroll →You're using AI as a chatbot and that's it
The team has been using AI like an ATM - you go to it when you need something. Nobody's built a workflow. You know there's more but don't know where to start.
You need to expand output, not headcount.
You don't need more people - you need more business impact. You need every person to expand what they can do, and someone (me) to build AI into how you work.
AI needs to show up on the P&L
You're not interested in productivity theater or making the team faster typists. You need an AI operating model to drive top and bottom-line impact.
You already did a training. Nothing's happening.
Without more side-by-side teaching, there is no reason a team should know what to do. You know a login is not leverage.
You have an enterprise contract and zero adoption.
Procurement signed the contract, the seats are paid for. People open it, ask one question, and leave. You're burning cash and damaging morale.
Your exec team is excited, and the rest of the team is lost.
Leadership has been pushing for "using AI" and sending around articles, but providing no support or direction.
You could do this yourself but have no time.
You need someone to sit you down to teach and lead with expertise from having done this many times before.
This may not be the right fit if
- You're already coding, scheduling tasks, or building automations.
- You're looking for fractional AI leadership. This is build-teach-exit work, not an embedded role.
- You want complex proprietary infrastructure. You need a software engineer - or at the very least, not a teacher (me).
- You want someone to build everything for you. I do not do that. I build a few things in support of long-term enablement.
- You need an AI roadmap. My output is your team's capability and curiosity, not a deck. (Models evolve so quickly, I don't know how to plan.)
Platforms I build for / with
Other offerings ✶
- AI for Everyone self-guided curriculum
- Codex training
- Claude Code training
- Managed agents training (agents that live in your Slack and act as part of your org chart)
These are custom. Email grace@grace-clarke.com.