Intelligent platforms · not an industry specialism
From discovering demand and generating leads to content distribution, websites, data analysis and trading systems, we design and build automation platforms that turn complex workflows into scalable systems. The industry a platform is deployed into is the configuration, not the expertise.
A live estate under active development, not a closed archive.
What we do
Each platform is a reusable engine. Capabilities, AI, automation, data, APIs, workflow orchestration, combine into a platform; the platform is then configured for a particular application, market or client. The industry is the deployment context, never the core expertise.
Automates the path from discovering demand to creating measurable opportunities. Demand discovery, content, distribution and capture as one connected system.
What we build toward A repeatable engine that finds demand and converts it, not a single campaign.
Discovers what people, markets or customers want and turns those signals into structured intelligence that feeds lead generation, content, product and distribution decisions.
What we build toward A single intelligence layer other platforms draw on, instead of one-off research.
Website generation, rebuilds and AI enhancement as part of a larger automation capability, not a standalone design service.
What we build toward A site that is one node in a bigger automation system, not the whole engagement.
Demand-led topic discovery through to content generation, video automation and multi-channel publishing, with attribution and feedback closing the loop.
What we build toward Content that is routed by evidence of demand, not a publishing calendar.
Not a claim of expertise in every market, an engineering capability for organising trading intelligence, comparing strategies and automating execution and monitoring workflows.
What we build toward Systems that keep their promises after the backtest is over.
Customer interaction, appointment booking, qualification and the operational workflows around them, automated end to end.
What we build toward Operational load taken off a team, with a visible audit trail.
How we work
Not a workflow we describe in a pitch deck, one we actually run, on every build, that leaves a record behind. Plan, design, implementation: each stage has its own discipline, and each one is enforced by a skill, not a good intention.
See our implementation workflowsEvery task becomes a timestamped, documented record with a required evidence section built in from the start, not filled in after the fact, and moves through a real lifecycle: backlog, in progress, complete.
Non-trivial work gets a real process map before it gets built: a top-down hierarchy of nodes, each with its own completion test, so the build has a spec to be checked against, not just a memory of what was intended.
Multiple models can work the same codebase without corrupting each other's changes, coordinated through a shared, append-only message board instead of luck.
Every claim of "done" needs typed evidence attached, a test result, a live URL, a log, graded against what it actually proves. A task isn't complete because someone said so; it's complete because the evidence says so.
Case studies
Not a list of apps, proof we know how to build an ecosystem. Every one below is standalone, but they share infrastructure: one strategy directory across the trading estate, one CRM across the directory & leads estate. Application screens are captured from local runs and presented as static case-study records, not a demo reel.
Finds real, server-derived demand before a campaign spends a penny, then runs the full lead lifecycle from opportunity gate to delivery. Standalone tool; also the front door that feeds qualified demand into the wider distribution engine.
Skills library: ep050-distribution-engine-skill · distribution_engine · agent-message-board
Every strategy ranked by net return after real costs, recalculated live from closed trades, never a backtest curve. Standalone research tool; also the shared source of truth every other app in the estate reads from.
Skills library: ep051-strategy-directory-skill · hosted-directory-data-platform
A trading floor where participant-owned agents allocate a simulated dollar between Directory strategies and watch each other do it, every decision logged, no real money at risk. Standalone spectacle; also where a Finder match or a fantasy portfolio actually goes to run.
Skills library: ep052-agentic-arena-skill · strategy-directory-client
Describe what you need in plain English; it searches the Directory, ranks the matches and hands the reference straight to an agent. It doesn't trade, it finds and hands off. Standalone search tool; also the entry point into the Arena and the Fantasy Challenge.
Skills library: ep053-strategy-finder-ai-skill · strategy-directory-client
Build a fantasy portfolio from real Directory strategies, back your own picks against everyone else's, and send the winner into the Arena as an agent. Human-first, preview-only demo, standalone game; also a second on-ramp into the same live strategy data.
Skills library: ep054-strategy-fantasy-challenge-skill · strategy-directory-client
A real local business directory, 1,498 businesses across 51 categories and 48 towns, paired with a local news feed that drives claim-and-quote traffic straight back into it. Behind the public site, a full CRM runs the pipeline end to end: import, field validation, a claims step that captures real owner consent, then generation, a live site built only once a business has claimed its listing. It's not a standalone product; it's shared lead-management infrastructure any of our apps can tap into to acquire and qualify leads, the same way the trading estate shares one strategy directory.
Skills library: ep047-directory-app-skill · leads-flywheel · site-generation-workflow
The infrastructure behind every build on this page: an append-only, timestamped record that lets independent models, Claude, Gemini, Codex, Hermes, work the same codebase without corrupting each other's changes. Not a demo: the Demand Hunt deployment above was root-caused and shipped through 21 of these messages between Claude and Hermes, live, the same day it went out.
Skills library: agent-message-board · workstream-task-lifecycle
A multi-tenant assistant that lives on a client site, answers real questions and captures leads. The owner gets a traffic dashboard behind it, summary stats, and a demand-signal view that flags above-baseline interest in specific services so they can react in real time: launch a promotion, and the assistant surfaces it live in the chat widget itself, with impressions, clicks and conversions tracked back.
Skills library: ep043-website-tracking · ep043-business-owner-access · ep044_common_site_blueprint
Next step
A short conversation, an honest read on whether we're the right studio, and a clear picture of what it would take. No pitch deck.