Directing Your AI Ensemble
How to Operate Strategically While Executing Flawlessly.
After teaching hundreds of knowledge workers how to build custom AIs, a pattern emerged.
People fall into two camps.
Camp 1: The Generalists
They build one custom AI that spans their entire role. It’s their thinking partner, their strategist, their development coach, and their execution assistant—all in one.
But because it’s a generalist, execution suffers. When it comes to your most important tasks, it delivers well, but not perfectly. Mistakes happen. Outputs need reworking. It knows everything about you but masters nothing specific.
Camp 2: The Specialists
They build many different custom AIs. One for email campaigns. One for proposal writing. One for client delivery. One for employee coaching. Each one works impressively well.
But using them feels like managing freelancers who never talk to each other. You lose context jumping between conversations. Constant copy-pasting. No compounding intelligence. No shared memory. Your knowledge stays fragmented across isolated tools.
The paradox: You need both—the deep context of a generalist AND the precision execution of specialists.
But having both meant choosing between scattered tools or middling performance.
This is not an issue anymore, but hardly anyone knows it.
A quiet feature update earlier this year in ChatGPT, and more recently Microsoft Copilot and Claude (called Skills) has solved it:
You can @mention any custom AI to bring it into a conversation thread.
No new window. No copy-paste. Just @CustomAIName and voila, it joins the conversation.
So your generalist AI and your specialist AIs can seamlessly work together.
And you can operate as a confident AI conductor.
Working ON + Working IN
There’s a distinction in business that every manager knows:
Working ON the business: Strategy, vision, reflection, planning.
Working IN the business: Execution, delivery, operations, tasks.
Most people can only do one at a time. And often great strategists struggle with execution, and great executors lack strategic acumen.
Teams solved this through collaboration—but at the cost of meetings, alignment overhead, and communication friction.
AI orchestration changes the game.
Your General Strategist AI = Working ON
Helps you think strategically about your role
Reflects your goals, values, and growth areas
Provides coaching, frameworks, and strategic insight
Challenges your thinking and expands possibilities
Your Task Specialist AIs = Working IN
Execute specific tasks with precision and expertise
Handle repetitive workflows consistently
Apply domain-specific knowledge and best practices
Deliver polished outputs that match your standards
The @customAI workflow bridges them.
You stay in your strategic conversation—working ON.
And when you need something executed—working IN—you call in the right specialist. It contributes with full context.
Then you return to your generalist strategy.
Now you can think strategically while executing tactically—in the same breath (or conversation thread).
You stop choosing between thinking and doing. You do both, simultaneously, through orchestration.
You become both executive and operations team—not by doing everything yourself, but by learning to conduct your AI ensemble.
How This Works: My Personal Stack
I used to have one generalist ChatGPT Project that handled everything.
It knew my Coachfully.AI work, my voice, my offerings, my value proposition.
But when I asked it to write a client proposal or draft my weekly newsletter, the output was... good, but rarely great.
So when I discovered the @CustomAI feature, I rebuilt my stack.
I transformed my generalist AI into my strategic partner and Co-CEO. It doesn’t write my newsletter or proposals anymore. Instead, it:
Thinks through business strategy and client dynamics
Proactively challenges my assumptions and exposes blind spots
Plans the work for my specialist AIs when I need execution
Then created three AI specialists for my most important recurring work:
@ProposalWriter - Turns discovery call transcripts into tailored coaching proposals with clear value propositions and next steps.
@EmailNewsletter - Crafts weekly story-driven emails that connect personal narrative to actionable insights.
@ArticleWriter - Acts as editorial partner and thought leadership writer—sharpening ideas, challenging weak arguments, structuring narratives that move from insight to action (Note: this AI already existed and I use Claude because it’s a superior writer).
Now when I’m thinking through client strategy or content direction with my Co-CEO, mid-conversation I can type:
“@EmailNewsletter take this Substack article and draft this week’s email about the orchestration concept—lead with the two camps problem, connect to my personal transformation story”
It joins the conversation with full context, drafts the email in my voice and style, and I’m back to strategic thinking when I’m done the email.
Or if I’ve just finished a discovery call for a new client engagement:
“@ProposalWriter take this transcript and create a follow-up proposal for the executive team—emphasize the AI orchestration framework and include the organizational intelligence progression in a phased approach.”
The proposal gets drafted while I have all the context and stay in strategic mode with my Co-CEO.
What changed?
The work got better. The proposals are more tailored and need no editing. The newsletters are written faster, and follow best practices to a tee. The articles are tighter and more aligned with my insights and vision.
Also my generalist is creating more strategic value for me. It only does what it does best: strategic thinking, problem-solving, and planning the work for my specialist AIs.
But more importantly, I got better.
I stopped being a task-switcher and became a better AI Director. I spend my time in creative, strategic mode—thinking ON my business—while my specialists handle execution IN the business.
From Personal to Organizational Intelligence
Right now, this feels like a personal productivity breakthrough.
But a bigger pattern is emerging that will reinvent how we use AI to get our best work done.
Stage 1: Individual.
Every employee builds their own generalist AI—a personal strategist for their role. Over time, they add AI specialists for key recurring tasks.
Stage 2: Peer-to-Peer.
Sarah in marketing built an amazing competitor analysis GPT. Jake in sales tries it, and loves it. He shares his customer researcher AI with her. She has something similar, and they compare notes.
Stage 3: Team.
The team realizes they’re solving similar problems differently. They collaborate on shared specialists and publish them for all: a company-branded proposal writer, a customer service responder, an internal knowledge retriever that knows company processes.
The instructions for these shared specialists then get refined and improved through collective use.
Stage 4: Organizational Intelligence.
Over time, the best specialist AIs become living company assets—trained on company knowledge, brand voice, and best practices.
Continuously refined through team usage. Accessible to anyone who needs them.
A form of collective, compound intelligence.
This transforms AI adoption from ad hoc tool adoption into structured capability evolution:
Personal Intelligence → Team Intelligence → Organizational Intelligence
The implications are extensive, and profound.
Your company’s collective expertise becomes encoded and accessible. New employees gain immediate access to institutional knowledge. Best practices spread organically through shared AI tools.
The organization gets smarter over time, not just individuals.
This is knowledge compounding at scale. Every use makes the system more intelligent. Every refinement benefits everyone.
The organization develops a form of collective superintelligence, which is amplified individual intelligence working in concert.
While You Were Prompting
Here’s what’s happening right now:
A quiet revolution is underway. While most people are still learning to prompt AI, a small group is learning to direct it.
They’re building custom AI ensembles—a strategic generalist surrounded by precision specialists—and using @mentionAI to orchestrate them seamlessly.
The solopreneur who used to struggle with execution? Now operating with capabilities that once required a full team.
The small marketing team competing against enterprise budgets? Punching above their weight class through coordinated AI intelligence.
The large company watching productivity stagnate despite AI investment? Being outmaneuvered by competitors who’ve mastered collective AI orchestration.
The defining capability of the next decade won’t be prompt engineering or general AI literacy.
It will be orchestration—the ability to conduct multiple specialized intelligences into coherent, compounding outcomes.
And the window to become an early adopter is closing faster than you think.
From Dabbler to Director
If you’re an individual contributor:
Start building your ensemble today:
Create your Strategic Partner (generalist AI) that knows your role, goals, and context
Identify your most important or most repetitive tasks that need consistent excellence
Build specialist AIs for each one with clear instructions and examples
Practice orchestration by using @mention to bring specialists into your strategic conversations
Watch what happens when your intelligences start working together instead of in isolation.
If you’re a team leader:
Your competitive advantage isn’t just having AI—it’s having a team that knows how to conduct it. This means:
Helping your team build custom AI ensembles (generalist + specialists)
Creating shared specialist AIs that become team assets
Establishing practices for refining and sharing custom AIs
Measuring orchestration capability, not just AI usage
Your team’s collective intelligence—amplified through orchestration—becomes your most defensible competitive advantage.
We Can Help
At Coachfully.AI, we’ve helped hundreds of knowledge workers and leadership teams build their AI ensembles and master orchestration.
Whether you’re an individual looking to 10x your capabilities or a leader building AI-amplified teams, we’ll help you make the shift from AI dabbler to AI director.
Because the future doesn’t belong to people who use AI. It belongs to those who learn to direct it.