AI consultancy and build partner for independent PR and marketing agencies.
AI is moving the margin in agencies. We build the operations that keep it in yours.
A £499 audit in ten days that names the margin you are leaving on the table. The first build in six weeks. Then we run and govern it with you, inside your own accounts, so your team owns what we make.
Not ready to talk? Take the free 10-minute readiness assessment and get a scored report within days.
Dan O'Sullivan scopes and builds every engagement himself. We serve few clients at a time on purpose.
First employee and COO of an education startup Roblox (NYSE:RBLX) later acquired. Fifteen years making new technology speak the operator's language.
Commercial Director at an a16z-backed studio whose job is agencies: selling into them, co-pitching with them, working in their teams.
The agents we build run the studio's commercial operations: pitch briefs turned into a costed first response, a morning pipeline briefing, conference and event tracking. Built and maintained by Dan.
Audit. Build. Maintain.
Three stages, each priced before it starts, each one optional. If the audit finds nothing worth building, we'll tell you and you keep the roadmap.
AI readiness audit
We benchmark you against comparable independents and identify the margin you are leaving on the table.
- 5 to 10 specific opportunities, each rated
- 90-day roadmap
- AI usage policy fitted to your agency
- Client-facing AI transparency statement
Priority build
The highest-impact agent or workflow on your roadmap, built inside your stack. Priced on the systems it connects to, the data it draws on and how much it runs on its own.
- Documented so your team runs it without us
- Projected monthly run-cost agreed before go-live
- Your people sign off wherever judgement is needed
Operations retainer
We put what we build into daily use: set up for everyone who needs it, the team trained, and fixed as real work tests it. Then we keep it running and governed, with a monthly hour with the MD so the team sees leadership using it.
- Maintain · rollout and training, upkeep, governance, a monthly MD hour£3,500
- Operate · plus ongoing development: every month we improve and extend what we've built, and your requests go first£5,000
- Lead · plus new builds as the roadmap grows, and a strategy day with your leadership team£7,000
The same budget, spent the usual ways
Annual prepay 10% off any retainer Referral credit 10% off your own build or retainer Prices exclude VAT
- A return-on-investment method for AI at a creative agency. Time data in, a per-tool number the board can act on out, so the agency knows which licences pay for themselves.
- An audit for an independent communications agency. The best build turned out to be the task the MD liked least: two hours every week assembling a media-mentions summary for the biggest client.
- A Claude setup for a B2B sales team. Built from the assistant Dan runs his own commercial work through: a morning briefing, pipeline and follow-up sweeps, a weekly pipeline report, all wired to the team's email, calendar and CRM. Handed over as a pack the sales lead installs himself.
- Advisory for a school on AI-drafted pupil reports. Special educational needs profiles are the most sensitive documents a school writes, so the design started with the constraints: pupil data never leaves the school's own Google environment, a signed-off language guide anchors every draft, and a person reviews each report before a parent sees it.
- An AI visibility panel. It asks ChatGPT, Gemini and Claude the questions a client's buyers ask, and counts who gets named and who doesn't.
of agency staff now use AI at work
of agencies have redesigned a workflow end to end
have no formal position on telling clients how AI is used
Most agencies share the same gap
Your team already uses AI: nine in ten agency staff do. What's missing is the system around it. Two thirds of agencies have no documented AI strategy, fewer than one in ten have redesigned a workflow end to end, and where a policy exists, nobody checks it. Some people lean on the tools heavily, others not at all, because client work always wins.
So the hours AI gives back leak out: nearly half of agencies say the time saved is absorbed or hard to account for. Meanwhile the team ships what HBR named workslop in 2025: AI output that looks polished but creates rework, dilutes your voice and quietly erodes client trust. Most MDs never see it. Your clients are reading the same headlines you are, and they will ask where the AI savings are. The agencies with an answer keep the margin.
- Independent UK PR and marketing agencies, £3m to £6m revenue, 15 to 50 staff
- Founder or MD-led, with operational decision-making authority
- The team uses AI every day but there's no owner, no playbook, and a policy nobody checks
- You'd rather pay someone to do this properly than fit it around client work
Who you'll work with
I'm Dan O'Sullivan and I run Daio. I started as a secondary science teacher through Teach First and by 2013 was building AI-powered learning tutors before generative models existed. Then a decade in commercial roles, most of it taking new technology into organisations that needed it to earn its keep.
Today I'm Commercial Director at an a16z-backed game studio, where the job is agencies: selling into them, co-pitching alongside them and working inside their teams. Daio is small on purpose. I scope and build the work, so you deal with the person solving your problem, and I take on few clients at a time.
- First employee and COO of an education startup Roblox (NYSE:RBLX) later acquired
- Co-author of a peer-reviewed best paper on conversational AI tutors
- Brought Paramount, Universal and Cartoon Network onto a kid-safe platform, the level of scrutiny your clients' legal teams will bring to AI
- Model-agnostic across Claude, OpenAI and Gemini; every build ships with a run-cost
The questions every MD asks
We’ve tried AI and got very little back. Why would this be different?
That is the common experience and a fair place to start. An MIT study in 2025 found 95% of company AI pilots returned nothing, and the latest UK agency benchmark says fewer than one in ten agencies have redesigned a workflow end to end. The reason is usually the same. The licences got bought, nobody owned the change, and fitting the tools to how you actually deliver never made it onto anyone’s calendar, because client work came first. The technology is not what changes the outcome. It is whether someone owns that fitting work and keeps it running once the novelty wears off. That is the job we do, and if the audit finds nothing worth building we will say so.
What happens to our client data?
Nothing without your sign-off. We work from documents you choose to share, under NDA as standard. Anything we build lives inside your own accounts and your stack, so client material never sits on our systems, and what AI touches is written into the client-facing AI transparency statement you get from the audit.
Which AI do you use?
Whichever fits the task. We’re model-agnostic across Claude, OpenAI and Gemini, and every recommendation comes with a projected monthly run-cost: model selection, usage caps, caching strategy. You get a clear number to approve before anything goes live, and a straight call on which model fits, with nothing riding on which provider you pick.
What does this mean for my team?
The parts clients pay you for stay with your people. The brief, the judgement and the relationship, at the front of every piece of work and at the back. We build the systems that compress the production middle: the rework, the reporting, the retrieval, the admin. And we put them into daily use rather than leaving them on a shelf: set up for everyone who needs them, the team trained, and a monthly hour with the MD so the team sees leadership using them.
Who owns what you build?
You do. Builds live in your accounts, and every build ships with documentation written so your team can run it without us. The retainer exists because most agencies want the systems maintained and governed, not because you’re locked in.
Do we have to tell clients?
More of them are asking. Over half of agencies have no formal position on telling clients how AI is used, and procurement teams at larger clients have started to ask for one. The audit gives you a client-facing AI transparency statement, the document you hand a client who asks what was AI-assisted in their work, and an internal usage policy that backs it up. Our view: say what you use, where a person signs off, and what never goes near a model. Said early, it reads as confidence. Said after a client asks, it reads as an admission.
Three ways to start. Pick the one that fits.
The assessment is free and takes ten minutes. The audit is £499 and takes ten days. The call is twenty minutes with Dan and costs nothing.
AI Operations Readiness Assessment
10 minutes, 17 questions. Your agency's score across five dimensions, three margin opportunities and a 90-day outline, back within days. Name, agency and email only.
Form not loading? Open the assessment in a new tab.