For firms whose deliverable carries a name on it. Drafted holds the standards you work to and the way your practice works, so the drafting collapses and the judgment stays exactly where it belongs.
Before AI, the artefact and the expertise were sold together. You could not get the drawings without the engineer, so clients learned to think they were buying documents. AI has separated them, and once a report can be drafted at nearly no cost its price falls toward nothing.
For a firm that sells judgment that’s an opportunity rather than a threat, provided the drafting genuinely gets cheaper and the review genuinely stays rigorous. Drafted is built for that split: the machine drafts against your standards, and a qualified human reviews and signs.
Illustrative, from a fee proposal. What matters is the direction, not the exact numbers.
Speed without a standard is just faster mistakes. The reason a principal can put their name on Drafted output is that the standard is enforced before the draft exists, not audited afterwards.
Before an agent may touch the work it has to search your organisation’s knowledge: your standards, your client history, the decisions you already made.
Your procedure for this kind of work: the checklist, the evidence bar, the house style. Not loaded means the write is refused.
The output lands as frames your team watches appear, compares side by side, and reviews before anyone signs anything.
What you fix goes back into the Wiki and the Skill, so it’s already fixed for every future job and everyone who runs one.
An Australian engineering consulting firm keeps the standards its practice runs on in its Drafted wiki: AS 5342, the CIBSE commissioning codes, AIRAH DA27, and its own written handover protocols. An agent working one of their jobs doesn’t get the standards pasted into a prompt. It has to go and read them.
They haven’t agreed to be named yet. When they do, their name will be here.
| The concern | The answer |
|---|---|
| Will this replace our consultants? | It replaces the typing, not the accountability. Your people get more client-facing hours, not fewer. |
| Who is liable when the AI is wrong? | Exactly where liability sits when you sign a graduate’s draft. The agent drafts; a qualified human reviews and signs. |
| If AI does the junior work, how do graduates learn? | Drafting boilerplate never taught judgment. Juniors move up to reviewing drafts against the standards, with the sources one click away, and get into client rooms earlier. |
| Why not just use ChatGPT? | Chat produces text. A firm produces signed, standards-compliant deliverables in house style, with a trail showing what they were checked against. |
Start with one live proposal or one review, and see where the hours land.