Endoxa's work takes different forms.

• A startup writing a white paper on what its technology actually does and why it matters.
• A foundation drafting the criteria it will use to evaluate AI safety grants.
• A research group preparing a grant for a funder who doesn't know what AI is, exactly.
• A company writing internal documentation to help its teams leverage a new AI system.
• A government office trying to understand a technical proposal well enough to respond.

Different briefs: the same craft, clarity, and care.

Endoxa's work is grounded in a sound understanding of AI capability and potential and in a disciplined conceptual and writing practice that makes AI use legible to a wide array of audiences, both inside and outside of an organization.

The other half of the work: prose that moves the people who read it.

Clarity is necessary but not sufficient. A document that is precisely correct and emotionally inert will not win a grant, will not convert a sceptic, will not be remembered after the meeting. The market has noticed: corporate executives mentioned storytelling on earnings calls 469 times in 2025, up from 359 in 2024¹ — a shift that tracks the rise of generative models almost exactly. As cookie-cutter prose has become free and infinite, judgment, taste, and the capacity to make a reader feel something have become scarce and expensive. Companies are hiring writers who think like strategists.

Most writing that targets feeling does not stand on firm argument. Most writing that argues firmly does not move anyone to care. The work that does both is rare, and it is the work Endoxa does. A well-built white paper should leave a sceptic unable to dismiss it and leave a believer unable to put it down. A grant narrative should withstand reviewer scrutiny and make the reviewer want the project to win. A founder's essay should hold up to argument and sound like something only that founder could have written. Rigor and resonance are not opposed; they are the two things careful prose has always been responsible for at the same time. Endoxa builds for both.

Increasingly, effective writing is also becoming an integral part of AI-powered systems and workflows themselves.

Ordinary language has become a technical language — a programming language — and lives in direct contact with compute. The AI researcher Andrej Karpathy calls this Software 3.0: where Software 1.0 was code humans wrote for machines and Software 2.0 was the weights of trained neural networks, Software 3.0 is the paradigm in which prompts, specifications, and ordinary prose are the program. As Karpathy put it in 2023, “The hottest new programming language is English.” Robust and clear writing is now an indispensable skill for the AI-powered workplace. New genres of optimal writing practices are emerging on this terrain, and Endoxa is developing a practice around them:

• An organizational charter that guides all AI behaviour.
• A system prompt that encodes a team's standards into every interaction.
• A specification an agent will follow as faithfully as a contract.
• A rubric that tells humans and machines alike what “good” looks like.

Endoxa brings clarity and coherence to AI work in ways that move work forward.

Endoxa welcomes new connections, conversations, and engagements. Contact us here.