The human gate stays
Everything is an assist to the person who signs your content off - never a replacement. Every output says so, every time, in so many words.
AI writing help built the way a regulated firm needs it built: your compliance sign-off stays in charge of every word that goes out - it just gets drafts that have already been argued with once. Information-not-advice by design, and honest about what a tool can and can't claim.
Most advisers don't - and in a regulated firm, that scepticism is doing its job. So nothing here asks you to trust a machine's judgement. Every skill we build for this sector works to the same three rules, stated on the page and baked into the outputs.
Everything is an assist to the person who signs your content off - never a replacement. Every output says so, every time, in so many words.
Nothing we make claims to be “compliant”, “approved”, or endorsed by anyone - because nothing that runs in a chat window honestly can. Rule citations are never presented as verdicts.
Where a fix needs firm-approved language - risk warnings, disclosures, status lines - the skills mark the spot rather than inventing wording your compliance team never agreed.
A £39 skill you can try this afternoon, or a pack built on your firm's own voice and approved wordings - scoped by conversation, because that's how a regulated firm buys.
Flags what your compliance sign-off is likely to catch - advice-vs-information drift, claims that need balancing, missing risk warnings - before the draft leaves your desk. Runs in your own claude.ai; team licence for up to 5.
Two firm-specific skills: a voice profile built from your own writing with the regulated register held in (so drafts sound like your firm, not a careful robot), and a branded document skill that produces your recurring client documents to your layout - the same craft as the demo below. Delivered into your own workspace as files you keep.
The mock-up alongside is for a fictional fund-governance firm - Okenshaw Fund Governance doesn't exist, and any resemblance to a real firm is coincidental. The craft is real: it's the same process behind the branded document skill we run in live client work - a skill that knows your sections, your tables, your typography, and your register, and drafts the recurring document from this quarter's inputs.
The draft lands with your team for review and sign-off, like any draft - it just starts from your format instead of a blank page.
Two sentences is plenty. Darren reads every enquiry and replies personally - no list, no sequence, and an honest "not yet" if that's the right answer.
No, and we'd be wary of anyone who says otherwise. Everything here is an assist to the person who signs your content off - it prepares drafts, flags what sign-off is likely to question, and holds your firm's register. The judgement, and the approval, stay human.
No - it sits alongside them. The established adviser content agencies leave compliance with your firm - their own terms say so - which means their drafts still go through your sign-off. Skills that flag what sign-off will likely question, in your own voice, make that relationship faster - whoever writes the first draft.
The ready-made skills are tools for client-facing marketing content - personal client data shouldn't go through them. What the firm-specific document skill handles, and how, is agreed as part of scoping. Either way, drafts run in your own claude.ai workspace under Anthropic's data terms and don't come to us - point your data-protection review at those terms, the same as for any Claude use.
Don't take it on trust - take it in stages. The £39 flagger shows you the register we work to before any bigger conversation. The firm-specific work is scoped on a call, delivered into your own workspace as files you keep, and references are available on request.
Walk us through what your firm writes and what sign-off keeps catching. We'll tell you honestly whether a £39 skill covers it, the pack fits, or neither does yet.