What we have shipped, what we have changed our minds about.

v0.6   2026-05-08

AI Growth Department offer and readiness check added.

• Published the AI Growth Department Blueprint offer page for the first customer acquisition sprint.
• Added the local-first AI Department Readiness Check to recommend the first safe growth role to build.
• Updated homepage, navigation, footer, and privacy copy to reflect the new customer-facing acquisition path.
• Kept the tool deterministic and browser-local. No backend storage or AI call is active.

v0.5   2026-05-07

Workflow Sketchpad beta added.

• Added a client-side Workflow Sketchpad for building a simple process map in the browser.
• Added editable JSON export, a PDF-ready flat report view, and an email handoff that tells users to attach the JSON if they want a project review head start.
• Updated the privacy page to explain that Sketchpad data stays local unless the visitor chooses to download or email it.
• No backend storage is active yet. That will only ship with explicit consent, validation, and a clear data-handling note.

v0.4   2026-05-07

Public guardrails, self-check, and Workflow X-ray launched.

• Published the How we work page covering operating principles, work we will do, work we will refuse, safe pilot principles, and data handling.
• Released a client-side Governance self-check: 8 questions, red/amber/green result, runs entirely in the browser.
• Released a beta Workflow X-ray intake. Deterministic logic only, no AI claims, no data sent anywhere unless the user emails the result.
• Added a sample managed-operations dashboard as a labelled lab page so prospects can see the shape of what we track.
• Published a privacy and data handling page in plain English.
• Started Field notes with four practical posts on AI failures, decision mapping, what should stay human, and chatbot pragmatism.

v0.3   2026-04-22

Refined the offer ladder and removed "AI strategy".

Replaced vague AI strategy framing with a five-step ladder: Governance Starter Pack, Process Digitiser, Assistant Opportunity Blueprint, Assistant Pilot, Managed AI Operations. Each one has a defined outcome and a clear exit. We deleted the catch-all "consulting" line because it lets us drift into work we would not stand behind.

v0.2   2026-03-31

Adopted a written kill-criteria template for pilots.

Every assistant pilot now ships with a written set of conditions that would pause or disable it: error rate threshold, drift in tone, customer complaint, vendor outage, audit failure, or owner request. The template is given to the client and lives next to the runbook. This was driven by a near-miss where a draft assistant kept producing valid-looking but stale references after a vendor model update.

v0.2   2026-03-12

Stopped using free-tier consumer AI tools for any client work.

After confirming retention behaviour was inconsistent across consumer tiers, we moved every workflow to paid services configured to not retain or train on client data. The data-handling commitments on the privacy page reflect this.

v0.1   2026-02-04

First public site, first written principles.

Published the initial Automation Nation site with positioning, offers, and a short principles section. Internal version of the engagement runbook, kill-criteria template, and data-handling note went into use with the first three pilot clients.