Founder-led · London & DACH
Senior, AI-accelerated research, origination and workflow support for advisory teams, charities and specialist businesses. From public-record intelligence to practical AI workflows and rapid prototypes, the aim is simple: do better work faster without losing judgement, traceability or context.
Founder-led · Evidence-first where it matters · No mass-database exports · No junior research layer
A documented ownership, funding or operating change has occurred.
Verified — named primary sourceThe timing and surrounding facts are consistent across the underlying record.
Verified — supporting documents checkedThe change may create a credible reason to engage now rather than later.
Inferred — from the verified facts aboveWhether an adviser, partner or competing bidder is already involved.
Unresolved — not established from available evidenceV verified · I inferred · U unresolved — the gaps stay visible
[V] Verified against a named primary source where possible.
[I] Inferred, with the facts and reasoning made explicit.
[U] Unresolved. Reported rather than quietly dropped.
The work starts with the question, not the tool. For public-record research, that means finding and checking the documents that actually carry the claim. For workflow work, it means understanding the process before automating it. For prototypes, it means building enough to test whether the idea deserves further investment.
Where evidence does not support a conclusion, the answer is not padded. Where client information is involved, handling and tool choices are agreed before anything is shared.
For teams where one well-evidenced opportunity is worth more than a generic list of hundreds.
Improving how existing work gets done, rather than adding AI for its own sake.
Small working products that test whether an idea deserves real investment.
Things built to test how AI, public data and lightweight automation handle real research and decision-making problems.
Finds, enriches and monitors high-value prospects and targets using evidence from the public record, with a pursue / monitor / reject view rather than a lead feed.
Open ProspectFitLab →An experimental dashboard placing housing, crypto and major commodities in historical context against a deceptively simple question: is now a good time to buy?
Open TimeToBuy →Turns an address into structured insight on neighbourhood, schools, transport, planning and the other factors buyers actually weigh.
Open Jollo →Nearly a thousand IMF Article IV reports, structured and classified by theme so a specific policy question can be answered in minutes rather than days.
Open the Explorer →Recommendations that begin with context beat recommendations that begin with a product list. The point is to understand the organisation first, then decide where AI or deeper research would genuinely help.
Roland gave us tangible AI suggestions that we were able to act on immediately. We were also impressed at the research and AI work he had accomplished before meeting us, ensuring he had valuable insight into who we are and what we do — meaning he could tailor his suggestions to our charity.
No transformation programme by default. The right starting point depends on whether the problem is research, workflow or an idea that needs to be made tangible.
A focused piece of work around one company, sector angle, prospect set or business problem — designed to show whether the approach creates useful signal.
Continuous monitoring, enrichment and periodic briefs where the value comes from catching meaningful changes early and repeatedly.
A working tool, interface or AI-enabled process built quickly enough to test with real users before committing a larger budget.
Focused proof-of-value work typically starts around £1,000–£1,250. Ongoing managed support is scoped once the first piece of work has shown where the value really sits.
A London-based, founder-led practice run by Roland Arlt, working across the UK and DACH markets in English and German. AI accelerates the research and build process; senior judgement remains on every piece of work.
Background spans banking, software, cybersecurity, sales, business development, fundraising and early-stage technology businesses.
Previous employers / professional background. Not clients and not endorsements.
Founder-led. No junior research layer and no outsourcing. The person who does the reading writes the brief.
Evidence-first. Public-record work uses named sources; where the evidence does not support a conclusion, the brief says so.
Bilingual. English and German, with DACH research and outreach delivered in German where appropriate.
Practical. The output is a decision, brief, workflow or prototype — not AI theatre.
Bring one target, one workflow or one question. If there is no useful fit, the call should make that clear quickly.
No sales theatre — just a practical conversation about the question you are trying to solve.
Leave an address and I will send a short note on where AI or evidence-led research could apply to your organisation.
Advisory and M&A teams, charities and fundraising teams, and specialist businesses where better research, sharper targeting or a practical AI workflow can improve decisions.
For public-record work: Companies House filings, PSC and charges registers, procurement notices, regulator registers, published accounts and other named sources. Where a project uses client information, handling and tool choices are agreed first.
Verified means a named source supports the claim. Inferred means the conclusion follows from verified facts and is labelled as interpretation. Unresolved means the point could not be established and is stated rather than omitted.
No. Deliverables are designed for advisory, fundraising and leadership teams to use directly, whether they are briefs, workflows, prototypes or lightweight web tools.
Yes. Work can be delivered in English or German, including DACH-focused research and outreach.
Public-record research can often be done without any internal data. If client material is needed, the scope, handling and tools are agreed before anything is shared.