wardley-maps.sgit.ai / Method
Four claims, and what would falsify each
This is the opinionated part of the site. Four places where the standard doctrine got pushed, contradicted or extended by six months of applying mapping to an agentic software project — each stated in the words it was first written in, each with a map, and each naming what evidence would change it.
Naming the falsifier is not modesty. If a map is a claim, not a picture is going to be this site's first sentence, then a page asserting a claim has to say what would settle it. Four pages, four falsifiers, and two of the four are currently open questions rather than settled positions.
De-commoditisation
"a thing can be commoditised at one phase of evolution while still being meaningfully custom-built at the next phase… I am going to call this pattern de-commoditisation." Explicitly framed as a gap in doctrine — and the strongest counter-argument, that ILC already covers it, has not been engaged with.
Read → Original · 28 July 2026The custom-axis verdict
You can relabel the evolution axis, and here is the rule for when you may: relabel the axis when the thing genuinely evolves; use a maturity model when the thing merely improves. The sharpest methodological contribution in the corpus.
Read → Original · 28 July 2026You cannot map a gap
"you cannot map a gap, because a gap has no evolution." Stated absolutely — and the map that provoked it exists and is good, which makes this the site's most interesting unresolved contradiction. An original map shape, not just an original map.
Read → Original · 7 February 2026Maps are graphs with position
"a Wardley Map is not a separate artifact from the graph. It is a graph with positional metadata." The strongest architectural idea here, and the one with no implementation at all. Published as method, labelled as design.
Read →Two more that carry a paragraph each
The commodity illusion
what is interesting when you map like this is that you see a lot of things that feel like a commodity or a product to the user are actually very weak and very immature behind the scenes.
A named reading technique rather than a new axis. It is what you get by mapping the user's experience and the implementation as separate value chains and then looking at the vertical distance between them. A login box is a commodity to the person typing into it and frequently Genesis behind the glass. 23 June 2026.
The agent as a user
the agent is also a user… Execution is mature; safe authorization via an LLM is custom-built and immature; the LLM is commodity, but the uninjected, on-task safe decision is not.
Anchoring a map on an agent rather than a person is a small move with large consequences, because the agent's needs decompose differently: it needs the same capability, and separately it needs the capability to be safe to invoke without supervision — and those two sit in completely different places on the axis. The second half of that quotation is the whole argument for the machine surface. 23 June 2026.
Two reading techniques these pages assume
User needs first. "the thing about Wardley maps that is always important is to start with the user needs, so we start with what they actually want." Standard doctrine, and stated here because every map on this site is anchored that way.
Paired before/after maps. "Each map is drawn twice. The reality map shows how immature the path is today… The with-our-service map shows the same value chain once our service supplies the missing data. The distance between the two is the case for the service." That pairing is what turns a diagram into an argument, and it is the format this site would use for any claim about a product — including its own.