The same wiring that sees connections makes sequencing feel unnatural
Iain McGilchrist describes two modes of attention. The right hemisphere sees wholes — patterns, relationships, the gestalt of a situation. The left hemisphere sequences, categorises, and communicates step by step. Most creative leaders operate predominantly in the first mode. But the world communicates in the second.
The MIND strengths that Brock and Fernette Eide describe in The Dyslexic Advantage— Material, Interconnected, Narrative, Dynamic reasoning — come with a cost. Spelling is harder. Sequencing is slower. Linear documentation fights against you. The ideas are there. Getting them onto paper in a form that a linear thinker can follow? That's the translation problem.
In meetings: the vision is clear in ten seconds. Writing the spec takes ten days. By the time it's done, the moment has passed, the priorities have shifted, and the idea is filed under “backlog.”
“I communicate badly yet have great ideas.”
The thing most creative leaders won't say. I'll say it.
Every time the bridge existed, the idea lived
Looking back, the pattern is obvious. The work that succeeded wasn't better work — it was work where someone or something translated the vision into language the decision-makers could act on.
Jessica Vickers
Jessica translated the NorthStar vision to the board in language they trusted. She understood what I was building, believed in it, and reframed it as clinical strategy. The outcome: CDO appointment. The idea lived because she existed.
CEO direct alignment
Rare. The CEO understood the design-first methodology without translation. He saw the connection between discovery, clinical workflow, and patient outcomes directly. No intermediary required. Design-First was approved as methodology from week one.
YouTube walkthrough
I documented the multi-brand token system as a walkthrough video. The global team watched it. They finally understood what the design system was for — not because I presented it differently, but because I had time to think it through before speaking. Async solved synchronous communication.
When the bridge was missing, the idea was right but still failed
These are harder to write. The idea was right. The execution was right. The communication was wrong — not factually, but structurally. The gap wasn't competence. It was translation.
Middle management translation
The NorthStar was right. The sprint board had no ticket for it. Middle management understood neither the vision nor how to break it into deliverables. The idea sat in a prototype file for months while the team shipped features against the old pattern.
Documentation timing
I wrote the spec. It took ten days — longer than I wanted, slower than my thinking. By the time it was done, priorities had shifted. The audience had moved on. The spec was right but it arrived late. The moment had passed.
The perpetual deferral
The idea was right. The communication was wrong — not factually, but politically. I stated the vision. I should have framed the outcome. I named the problem. I should have named the risk of inaction. The room agreed in principle and the sprint board never moved.
“The idea was right. The communication was wrong. Not factually — structurally. I named the problem. I should have named the cost of ignoring it.Approved in theory, blocked in practice — every time
CLAUDE.md isn't a context file. It's an assistive communication device.
It translates how I think into a format that agents — and by extension, teams — can consume. The dyslexia that made traditional documentation fight against me? AI doesn't care about spelling. It reads intent. It tolerates connection-rich, non-linear thinking and responds with structured, sequential output.
The spec that used to take ten days now takes an hour. Not because the ideas are simpler — but because I no longer have to translate them manually. I speak in connections. The agent produces sequence. That's the trade.
More importantly: the communication loop is tighter. I can test how an idea lands before it enters a room. I can reframe for a technical audience, then immediately reframe again for executives. The translation that used to require a Jessica Vickers in the room — a trusted interpreter with domain knowledge and political credibility — now has infrastructure.
Since adopting AI-assisted workflows, I do more creative work, not less. The mechanical labour is handled. What's left is judgement, vision, and connection — which is what I was always better at anyway.
One idea, four translations
The concept: a dedicated agent that takes raw thinking — passionate, fast, connection-rich, possibly misspelled — and translates it in real-time for any audience. Not a crutch. A bridge. The same way glasses aren't “artificial seeing”, this isn't “artificial communicating.” It's assistive.
Technical spec language
Acceptance criteria, API contracts, component API design, error states, edge cases
Strategic framing
ROI, competitive positioning, risk of inaction, timeline, business outcome
Visual thinking language
User empathy, design principles, pattern rationale, intent over instruction
Professional, measured language
Competencies mapped to role requirements, evidence-based claims, no jargon
The agent connects directly to the Stakeholder Simulator — which models how different stakeholders will respond. This agent crafts the message. That one predicts the reception. Together: a communication system that closes the loop between vision and understanding.
The agent doesn't just translate — it leads from behind
Translation alone isn't enough. The way a message is delivered shapes how it lands, independent of the content. The Communications Agent encodes servant leadership principles — not as a management philosophy, but as a communication protocol.
Wait for the other person to finish. Resist the urge to correct mid-sentence. The idea improves in the silence.
Optimise for the recipient's understanding, not your own clarity. Write for them. Translate for them. Not for yourself.
The translation must preserve the truth of the idea. Diplomacy is not dilution. Reframe — never distort.
These aren't abstract principles. They're structural. Lead with questions, not statements. Acknowledge before advising. Frame solutions as options, not directives. Make the other person feel heard before being guided. These are rules the agent can apply — not simulate, but actually follow as a communication pattern.
The 10 perfections encoded into a system. Patience (Khanti), loving-kindness (Metta), and truthfulness (Sacca) as infrastructure — not personality, not mood-dependent. Always on.
I don't design products. I design how products get designed. The NorthStar, the token system, the component architecture, the design system as shared infrastructure — that's the work. But I couldn't always say that in a way others understood.
The interpreter problem was always mine to solve. For most of my career I solved it by finding the right human bridge — someone who trusted the vision and could reframe it for the room. That worked when the bridge existed. It failed when it didn't.
Now the bridge is infrastructure. It doesn't depend on who's in the room. It doesn't get tired. It doesn't need a budget. It just translates — with patience, with precision, and with the truth of the idea intact.
The system that makes this possible
The Communications Agent is one piece of the team. See the full agent system — and the infrastructure behind it.