This site carries several bodies of writing — stories, essays, notes and other media — joined by one method and a set of commitments. This note explains both.
Who this is for
Mining Is Human and The Institutional Record are written for anyone interested in Southern African mining and the institutions that have shaped the region — readers who believe these stories deserve to be told.
Other essays, notes and media are written for General Counsel and executives, and board members in mining and other regulated sectors, and reflect the questions that surface in my advisory practice.
What each series is
Mining Is Human and The Institutional Record publish stories — of the mines, businesses, industries and institutions that have shaped Southern Africa. Each story assembles the public record into a single narrative.
Other essays, notes and media include notes, presentations, podcasts, and other media from my independent advisory practice — reflections from the work of sitting with leaders when clarity is needed and stakes are high.
What I write
Very little of what I write is wholly original. It is my own synthesis of ideas, experiences, sources and frameworks that I have found useful over fifteen years of leading legal teams, teaching and learning from in-house lawyers, and advising organisations under pressure.
Where I borrow concepts from management, psychology, organisational theory, history or other disciplines, I undertake sufficient research to ensure that I have understood and represented them fairly and in line with current thinking.
What AI-assisted means
My work is deliberately AI-assisted. I use AI extensively to search and research, challenge, organise, edit and sharpen my writing, and to create many of the accompanying graphics. I do not deliberately remove watermarks, provenance information or other indicators of AI generation from text or images, nor do I reprocess material for the purpose of concealing AI involvement.
AI assistance ranges from searching and analysing material available on the internet, through editing and fact checking, to drafting passages that I then revise, restructure and take responsibility for.
The selection of the subject, the narrative structure, the connections between events, the framings, the professional judgments and the conclusions are my own.
How I check the facts
Every factual proposition that survives into a story needs positive support. The absence of contradiction is not enough.
Sometimes the evidence is straightforward. A company may have published the production figure, technical result, transaction or investment that I am describing. For financial, operating and technical matters, I will sometimes rely on that published information, particularly where there is no practical means of independently establishing the underlying fact. Where it matters, the writing identifies it as the company’s figure or claim.
Often it is more difficult. The company may say nothing about the point. It may have changed ownership or name several times, or ceased to exist decades ago. Records may be incomplete and different accounts may conflict. In those cases I use AI search tools to work through whatever evidence I can find on the internet: contemporary company records and announcements, stock-exchange and regulatory material, government and geological records, legislation and judgments, academic and historical work, archived publications, contemporary journalism and other credible accounts.
I look for corroboration where it is available, particularly for facts that are uncertain, contested or important to the story. Repetition is not corroboration. Ten websites repeating the same unsourced statement do not turn it into a verified fact.
AI-assisted fact checking is a method of examining evidence, not independent verification in itself. I review and challenge the findings against my own knowledge and understanding, and consult underlying sources or seek independent corroboration where possible, particularly where a proposition is uncertain or consequential. Where I cannot independently verify a fact, I assess the evidence identified through AI-assisted research rather than treating the AI's conclusion as verification.
Ownership histories receive particular attention. Mining assets can pass through joint ventures, restructurings, nationalisations, mergers, disposals and changes of corporate name over many decades. I try to establish each link in the chain rather than infer one transaction from the next. My own previous writing cannot be used to verify my own writing, even when it has subsequently been reproduced elsewhere.
This process has evolved. Some of my earlier stories were researched and checked less rigorously than I would check them today. I have made mistakes, learned from them and progressively tightened the process.
Since mid-May 2026, every story has been subjected to successive fact checks as it develops. The last is run the night before publication against the story that will actually be published, not simply against the research from which it was written. That matters because compression can introduce error too: a complicated history can be accurately researched and then made inaccurate by reducing it to a sentence.
Where the evidence supports only a qualified formulation, I qualify it. Where accounts conflict and the conflict matters, I say so. Where the evidence cannot support the proposition, it comes out.
None of this makes the stories infallible. It is an attempt to make the confidence of the writing match the strength of the evidence.
Technical mining data — resources, reserves and production figures
Telling the story of a mine means engaging with numbers: tonnes milled, grades declared, ore bodies described, reserves estimated, production achieved over decades. These figures are part of the record. Without them, the story has no scale, no texture and no honesty about what these deposits meant.
I cite them for what they are: historical data, drawn from sources available online — my primary research surface — in the service of narrative. I am not a geologist, a competent person under the JORC Code, or an officer of a listed entity. I hold no professional credentials in mineral resource estimation or corporate disclosure. When I quote a resource figure, I am not making or endorsing a public report in the regulatory sense: I am telling you what the record shows.
I try to find confirmatory reporting, but the same figure — a JORC estimate, a production number, a declared grade — can be repeated many times across online sources, and a single misstatement replicates. My publication of any figure is neither an endorsement nor a verification: it is simply part of a story. If a figure is subsequently found to be wrong and I become aware of it, I will correct it.
Many of the figures in these stories predate modern resource classification standards entirely. Where a figure comes from a contemporary source — a listed company's announcement or technical report — I try to reproduce it accurately and attribute it to that source. I do not strip out context that the source included, and I do not add precision that the source did not claim.
Nothing in these stories is investment advice. Nothing should be read as a current assessment of any mineral asset, company or project. If you are making investment decisions, the announcements platform of the exchange where the company is listed, a licensed resource professional and your own independent advice are where you should start — not here.
What I do not claim
Unlike historians working from primary archives, experts whose authority has been earned through years of disciplined study and experience, or writers whose work is built on original reporting or literary craft, I make no claim to that kind of expertise or earned authority.
This is something different: a project driven by curiosity, sustained by discipline, and grounded in fifteen years of leading legal teams and advising organisations under pressure.
If I get something wrong
Tell me. I will investigate it, correct the article where appropriate, and gladly acknowledge your contribution.
Where a story is materially revised, the article on the publication site carries a dated note at the top setting out what changed, why, and who contributed.
Correspondence: writing@kevinlester.co.za