Most organisations already use AI in some form; few use it on purpose. We help small and mid-sized businesses, charities and public bodies decide where AI will pay off, build it properly, and govern it so that it stands up to scrutiny.
15 questions · Instant results · No sign-up needed
AI reached most workplaces through the back door: staff tried it, found it useful, and carried on without anyone deciding what it was for. The result is familiar, whatever the sector.
"Half the team uses ChatGPT, and nobody knows what goes into it."
Unmanaged use is the most common starting point: staff adopt free tools on their own initiative, and client, donor or personal data leaves the organisation with no policy to govern it.
"We ran a pilot. Everyone was impressed, and then nothing happened."
Pilots stall when they are chosen for novelty rather than value; without a baseline, an owner and a route into daily work, even a good demonstration quietly expires.
"We know we should be doing something. We just don't know where to start."
The market is loud and the advice contradicts itself. What most organisations need first is not a platform but a clear view of the three or four tasks where AI will save real time.
We did not come to AI through a sales deck. Our founding research used machine learning to forecast Islamic State attacks six months in advance, and today we build and run our own AI pipelines that read, classify and map large volumes of multilingual text for clients such as the Konrad Adenauer Stiftung.
Our advice comes from systems we have built, tested and maintained ourselves: we know where models fail, what they cost to run, and how to check their work before it reaches a client.
Every recommendation rests on a baseline and a measure: what the task costs today, what AI changes, and how you will know, in hours and pounds, whether it worked.
We come from risk intelligence, so governance is never an afterthought; data protection, bias, accountability and the UK and EU rules are built into the work from the first week.
No software company pays us to recommend its product. We choose Claude, ChatGPT, Copilot, open-source models, or no AI at all, according to what the task actually needs.
Four kinds of work, which most clients combine: a plan worth following, tools that do real work, people who use them well, and rules that keep the whole effort safe.
We map where AI can save time or improve quality across your organisation, rank each opportunity by value and risk, and turn the result into a costed, sequenced roadmap your leadership can stand behind.
We design and build the tools themselves: assistants grounded in your own documents, automated workflows for drafting, triage and reporting, and analysis pipelines for large volumes of text and data.
Tools change nothing unless people use them well. We train teams on real tasks drawn from their own work, and leave behind playbooks that make good practice the default rather than the exception.
We put the guardrails in place so that your use of AI is safe, lawful and defensible: to your board, your clients, your donors and, where it matters, your regulator.
Take the free readiness assessment below. Your score, a breakdown across six areas and a tailored 90-day plan appear instantly.
We talk through your results, your priorities and your constraints, and whether we are the right partner. No obligation.
A Readiness Sprint turns your priorities into a costed plan, with a measurable target for every initiative.
We build, train and govern, then hand everything over with the documentation your team needs to run it.
A practical day for up to 20 staff on using AI safely and well in their own roles, built around tasks they bring with them.
Best for: teams starting out
Interviews, a process review and a data audit, ending in a prioritised roadmap and a briefing for your leadership or trustees.
Best for: leaders who need a plan they can defend
One high-value use case, built, tested against a baseline and put into daily use, with the numbers to show what changed.
Best for: teams with a clear use case
Ongoing senior advice, governance review and new builds as your needs grow, without the cost of a full-time hire.
Best for: organisations scaling AI
Every engagement is scoped and priced after a free consultation. If the likely benefit does not justify the spend, we will say so.
Fifteen questions across six areas, built on the latest research into what separates organisations that get value from AI from those that do not. The moment you finish, you will see your score, what you could gain, quick wins and a 90-day plan: free, with no sign-up.
Our work began in humanitarian early warning and the protection of minorities, so we understand the constraints of mission-led organisations: budgets are tight, the data often concerns vulnerable people, and every decision may one day have to be explained to trustees, donors or the press.
We design with those constraints rather than around them; the aim is to give your staff back the hours lost to reporting and administration, without putting the people you serve at risk.
Small, fixed-price steps that pay for themselves before the next one begins; no licences you will not use.
Beneficiary and personal data stays out of public models, with safeguards documented from the start.
Policies, records and plain-language explanations your trustees and donors can read and trust.
We begin every engagement by establishing what data you hold, how sensitive it is and where it is allowed to go. We then choose tools and settings that keep sensitive information out of public models, and we document those decisions so that you can show them to a client, donor or regulator.
No. Most of the organisations we work with have no in-house developers. We design around the systems and skills you already have, and we hand over every tool with the documentation and training your staff need to run it.
Whatever fits the task. We are not paid by any software vendor, so we recommend Claude, ChatGPT, Microsoft Copilot, open-source models, or no AI at all, depending on your data, your budget and the work itself.
Workshops, sprints and pilots are fixed-price and scoped after a free consultation. If the likely benefit does not justify the spend, we will tell you in that first conversation.
A well-chosen pilot should show measurable time savings within six to eight weeks. If a use case cannot show value in that time, we would rather say so before you commit the budget.
Governance is built into our work from the first week. We map your AI use against UK GDPR, the EU AI Act where it applies to you, and relevant sector guidance, and we write the policies and records that make your use of AI defensible.
Start with the five-minute assessment, or speak to us directly: the first consultation is free, and it is with the person who will do the work.