AI and automation
We help you pinpoint where AI actually makes sense for your business, train your team to use it with confidence, and build simple automations around the way you already work.
You don't need to jump in at the deep end. We offer four straightforward ways to work together, whether you're just scoping out the possibilities or ready to build. A simple review or team training is usually the best place to begin.
A realistic assessment of your operations to spot repetitive bottlenecks. We map out where information is manually moved or repeatedly requested, evaluate your existing internal knowledge, and deliver a short report that cuts through the noise: what will save you time, what isn’t worth the hassle, and what to set up first.
Training focused on your staff’s actual workload, not hypothetical examples. We show your team what AI handles brilliantly, where it gets things wrong, and how to build daily habits that keep your client and patient information completely secure.
AI assistants that draw exclusively from your verified information. Designed for customer-facing sites or internal team platforms, they operate safely within set guardrails: acknowledging their limits and directing users to human support when needed.
We connect your internal platforms, software, and website so information flows smoothly without manual copy-pasting. We sort incoming enquiries to the right team member, draft initial replies for quick approval, and automate repetitive reporting tasks.
Examples
Assistants that answer from your own information, and automations that bring the right information to you, so nobody has to go looking for it or build it by hand.
Read the Western Rheumatology case studyThe clinic’s assistant answers patients’ common questions using the clinic’s own information, so fewer people need to ring the practice. Patients can also order repeat prescriptions online.
Every enquiry from your website, email and socials lands in one place, sorted by what it is and how urgent it is, with a reply drafted for checking. Follow-ups go out on time, and you can see at a glance which leads are new, waiting or booked.
Before the day starts, an email arrives with what needs doing: enquiries waiting on a reply, today’s meetings, upcoming deadlines and anything that’s been left too long. You know where to start before you open anything else.
A short research note, ready each morning or before an important meeting. What competitors are doing, news in your sector and what people are searching for, so you get the headlines without the hour of digging.
Website traffic, enquiries and social media results in one place, updated automatically. A plain-English summary tells you what changed this week and what’s worth acting on, so nobody has to build the report by hand.
The questions people ask in comments and messages, answered from your own information, like opening hours, availability or how to book. Anything new or sensitive goes to your team to reply to themselves.
No AI for the sake of it. We keep every project focused and measurable, so you can judge whether it actually saved the time it was built to save.
We map out where your team’s hours actually go before building anything. If a job doesn’t need AI, or just needs a simpler fix, we’ll tell you straight.
We start with the task that offers the clearest payoff. We keep the scope tight and measure it directly against the time it was meant to save.
The people doing the job test the system while we build it. Their feedback shapes how it behaves, and they decide when it’s genuinely ready to use.
We document every workflow so it runs smoothly without us. If you’d prefer us to keep looking after it, we can do that too.
Routine answers can be automated, but anything new or sensitive is prepared as a draft so a real person always has final approval.
We lay out every provider and their privacy terms before building. Confidential data is always protected and excluded from model training.
Assistants operate strictly within the boundaries of your internal information. If an answer isn’t in your docs or needs a human touch, it passes the question straight to a team member.
Who leads the work
Susan Ryan leads our AI work. Before founding the studio she supported critical infrastructure systems for London Underground and utility companies across the UK. Since 2023 she has trained clients’ teams to use AI in their everyday work.
Most of our day-to-day AI work happens in Claude Team Projects. Each client and each website we look after has its own project, holding the briefs, documents and instructions that work needs, so every conversation starts from the right context rather than from scratch.
We use the Team plan rather than personal accounts because of how it handles data. Under Anthropic’s commercial terms, our conversations and the documents in our projects are not used to train its models by default, and projects are visible only to people inside our studio account. As owners, we control the data settings for the whole account, including whether conversations can be sent to Anthropic as feedback.
Around those projects we have built workflows for the jobs that repeat: replying to enquiries, drafting quotes and proposals, writing care reports for the sites we maintain, and the admin behind each project. Each one produces a draft for one of us to check before anything goes out.
We are now setting up Hermes, an open-source agent that can carry out longer, multi-step tasks on its own, and running AI models locally on our own machines. Local models mean some work never leaves the studio, which matters for clients whose information should not go to an outside provider.
Tell us what takes up your team’s week. We will tell you honestly whether AI would help, and where we would start.