Projects

2026 to present - Founder, sole director

An acquisition vehicle for agricultural and resort land in the Republic of Georgia.

What it was

A UK company incorporated in July 2026 to acquire, hold and develop land in the Caucasus. The underlying holding runs to 2,040 hectares across 17 registered parcels, the great majority of it long-leasehold pasture with a smaller freehold component including a resort plot.

What happened

Live and in progress. Most of the work so far has been legal rather than commercial: establishing that a foreign-controlled Georgian company may lawfully hold agricultural land under the 2019 Organic Law, which turned on a grandfathering provision and on a reading of which entities the restriction actually catches. The answer changed the shape of the whole plan.

What it taught me

In a frontier jurisdiction the title work is the deal. I wrote three versions of the land-restriction analysis and withdrew the first two, because each time I went back to the primary legislation rather than the commentary I found the previous reading was wrong in a way that mattered. The commercial thesis was never the risk. The risk was whether we were allowed to own it.

georgianland.co.uk

2019 to present - Founder and principal

A UK family investment company holding property in the Tees Valley.

What it was

A family investment company owned in equal thirds by me and my two brothers, holding residential and serviced-accommodation property through single-purpose subsidiaries. Three properties in Billingham, two of them on the same street.

What happened

Live. It has been through a full cycle of the thing family structures are actually for: raising capital against an asset by selling a subsidiary and buying it back, restructuring the board so the company does not depend on one person, and learning the difference between owning property and running a hospitality operation.

What it taught me

A family investment company is a governance instrument before it is a tax instrument. The moment it had three directors rather than one, decisions got slower and better, and the structure stopped being an extension of me. That is the whole point and I did not appreciate it when I set it up.

hwlegacy.group

2023 to present - Founder and managing partner

Cross-border investment and advisory. Real assets, physical commodities and special situations.

What it was

An advisory and origination business working where consensus capital does not go: physical commodity sourcing and broking, mining and real-asset transactions, investment banking advisory, and connecting operators with funders across jurisdictions.

What happened

Live and the most portable thing I do, which matters when the rest of the estate is fixed in one place.

What it taught me

Origination is a relationship business with a legal spine. The deals that close are the ones where somebody on the inside trusts you enough to tell you what is actually wrong with the asset before you find out yourself.

eranospartners.com

2026 to present - Founder

An automated research and information-product business for UK personal finance.

What it was

A pipeline that senses what people are actually asking about money on public forums, generates researched guides against those questions, puts each one through a quality gate, renders and lists it, and sells it. Built end to end so that no step needs a person once it is running.

What happened

Live and pre-revenue. Every stage of the machine works. The open question is the one that kills most information products, which is where the readers come from, and the current answer being tested is organic search plus an owned mailing list.

What it taught me

I killed an earlier version of this in July 2026 and the reason is worth repeating: every variant I looked at had the same load-bearing step, which was a human earning strangers attention repeatedly over months. I do not want to be a content creator, so the only version worth building was one where the machine does that part. Whether that works is still being measured, and I would rather say that than claim it.

fathommoney.com

2017 to 2023 - Founder and chief executive

Distributed-ledger infrastructure for trading and managing wholesale telecoms capacity. It failed.

What it was

A group of companies building a distributed operating layer for global telecoms networks: a trading arm routing transmission for carriers, and a technology arm building the ledger and settlement infrastructure underneath it. Offices in Newcastle, Birmingham and Barcelona.

What happened

It raised, it hired seriously, including senior engineers out of Red Hat and wholesale telecoms people with real books of business, and it was named distributed-ledger lead in a successful consortium bid to the European Commission Horizon 2020 programme. Then it ran out of road. The technology arm went into liquidation and the group wound down.

What it taught me

Two things, and the second one took years to admit. The first is that selling infrastructure change to incumbent carriers means selling to the people whose margin the change removes, and no amount of technical superiority survives that. The second is that raising well and building well are not the same skill as reaching escape velocity, and I mistook the first two for the third for longer than I should have.

2013 to approximately 2017 - Commercial Director (from January 2014; joined February 2013)

Wholesale telecoms routing and network access. My entry into the industry.

What it was

Logik Communications Pte Ltd, a wholesale telecoms operator running transit, wholesale and retail voice routing, SMS, roaming and MVNO solutions, with points of presence across several regions. I joined in February 2013 as senior business development consultant and became Commercial Director in January 2014, responsible for sourcing and structuring the financing behind a fast-growing book, and for corporate development across Eurasia.

What happened

I ran the commercial and financing side for around four years. It led directly into Bartr, which was an attempt to rebuild the same market on better infrastructure.

What it taught me

Margin in a commodity trade comes from information and from settlement, not from the product. That observation is what Bartr was built on, and it was correct. It just was not sufficient.

2020 to 2023 - Owner and managing director

A bar and restaurant in Billingham. The most public and most misunderstood thing I have done.

What it was

A traditional real-ale bar and restaurant at 40 Mill Lane, Billingham, acquired in August 2020: hand-picked guest ales, Sunday roasts, a small local team, and during the pandemic a takeaway and online beer-delivery operation built from nothing.

What happened

In January 2021 I opened the premises during lockdown restrictions as part of a national protest about what the closures were doing to small hospitality businesses. Cleveland Police applied for a review of the premises licence, Stockton Council revoked it in July 2021, and I appealed. The appeal was dismissed at Teesside Magistrates Court in March 2022 and again in the High Court in February 2023, with costs. The business was sold in 2023.

What it taught me

I have written a full account of this separately, because the press coverage is the first thing anybody finds about me and it is not the whole story. Read that rather than this paragraph.

to early 2025 - Former director and former interim chief executive

A Singapore minerals-technology company. A former role, ended.

What it was

A company commercialising a thiosulfate polymer process for extracting gold from ore as an alternative to cyanide. I served as a director and as interim chief executive through a turnaround, covering company direction, investor strategy and commercial development.

What happened

My involvement ended in early 2025. I hold no current office, mandate or obligation with the company.

What it taught me

An interim chief executive brought in for a turnaround is being asked to make decisions the permanent management could not, with less information than they had. The job is to be decisive on the things that are reversible and extremely slow on the things that are not.

2026 to present - Builder

A personal operating system: an always-on agent running my businesses, compliance and research.

What it was

A cloud-resident set of agents working against a versioned knowledge base of everything I am responsible for. It produces a daily briefing, tracks statutory deadlines across a group of companies, runs research, writes code, and increasingly publishes. It runs whether my laptop is on or not.

What happened

Live and improving. It has a written constitution governing what it may and may not do on my behalf, and the interesting engineering is almost entirely in those constraints rather than in the capability.

What it taught me

The hard part of an autonomous system is not making it capable, it is making its actual permissions match its stated permissions. I once relaxed a rule in writing and left the enforcement layer untouched for two months, then read the resulting refusals as evidence the system was working correctly. It was working correctly against the old rule. Now nothing counts as in force until something enforces it and the test is logged.

2025 to present - Builder

Automated identification and claiming of ownerless UK land and assets.

What it was

A system that finds UK land and assets that have fallen to the Crown as bona vacantia, cross-references the public registers, and works up claimable cases.

What happened

Infrastructure built, in development.

What it taught me

Most of the value in a data product of this kind is in the joins between registers that were never designed to be joined.

2025 to present - Builder

A multi-model interface querying several frontier models in parallel.

What it was

A React application that puts the same question to several large language models at once and lays the answers side by side.

What happened

Built and used personally. Not commercialised.

What it taught me

The disagreements between models are more informative than any single answer. Where they all agree the question was easy; where they diverge is where the real uncertainty is, and that turns out to be a useful instrument in itself.

2026 - Builder

A rules engine for what may lawfully be communicated to whom, and in which jurisdiction, when raising capital.

What it was

An issuer-agnostic tool encoding eight routes across the UK, the US and offshore, built on a certify-before-communicate principle: the rule is not what you say, it is who you are allowed to say it to and what you established about them first.

What happened

Version one shipped in July 2026. The rules core has since been lifted into a separate system as the component that decides whether a piece of published content is a financial promotion.

What it taught me

The dated, certified investor list is the engine and everything else is plumbing. Almost every compliance failure in a small raise is a sequencing failure: the communication went out before the certification came back.

2025 - Founder

A consumer drug-purity indicator built on a proprietary compound.

What it was

A harm-reduction product: a consumer-grade indicator for checking the purity of a substance before it is taken.

What happened

Parked. It is a genuine research and development project and it needs a dedicated workstream rather than a corner of one.

What it taught me

Some ideas are not slow because they are wrong, they are slow because they need a laboratory. Knowing which kind you have saves a year.

2025 - Founder

Multilingual AI-generated audio narratives for walks, runs and journeys.

What it was

A mobile application that tells you where you are, in history, as you move through it, in your own language.

What happened

Parked. A mobile build is capital-intensive and it did not survive a ranking against everything else.

What it taught me

The idea I most enjoyed was the one with the weakest route to revenue. Enjoying an idea is not a signal, it is a bias.