Make AI coding agents follow your engineering standards - at much lower cost

Three things in one engagement: the complete context supply chain kit (a git repo of scripts, CI checks, and staged releases - full source, not slides), the live one-day workshop where your people operate it on a worked example, and the architecture behind it so your team can adapt it to your own standards afterwards. The kit isn't sold separately - the workshop is the only way to get it. Full breakdown in the “The Assets & The Know-How” section above.

No. Courses teach individual habits; this installs the governed pipeline around your company standards - ownership, CI validation, staged rollout, freshness, rollback, and measurement. You still learn hands-on, but you leave with the working kit, not just tips.

It is a 1-day live workshop, online. Private company workshops are scheduled with you on the planning call - usually a few weeks out. The first public cohort (September 8, 2026) is closed to new enrollments.

Not right now. Individual seats and the team bundle are no longer sold. The second cohort will be announced separately - join the list and you'll hear first. What is bookable today is the Private Company Workshop: delivered for your organization, tuned to your stack, with a 2-week implementation pilot included.

It will be announced separately, with its own dates and pricing - we are not taking seat sales for it yet. Add your email in the pricing section (or subscribe to the newsletter) and you will hear about it before it goes public.

It is live and hands-on - you operate the kit yourself in guided labs, on a worked example. All templates, lab material, and the kit you take home are yours to keep. Sessions are not recorded unless we confirm otherwise in writing before the workshop begins, and participants may not record, stream, capture, or transcribe the session (including screenshots and AI note-taking bots) without prior written consent. The lasting value is the kit and materials you keep, not a video.

Full refund up to 14 days before the workshop date. Inside that final 14-day window it switches to rescheduling - we move your private workshop to the next date that works for your team. Individual participants can always be swapped for colleagues at no cost.

You can - and most teams start there. The problem isn't writing the rules, it's everything after: copies drift apart across repos and tools, nothing validates changes before agents consume them, there's no staged rollout or rollback, and nobody can measure whether agents actually follow the rules. The workshop gives you the operating model around the files - versioning, validation, distribution, telemetry - in one day instead of months of trial and error.

Foundations workshops teach engineers how to use agents better: prompting, context windows, hallucination prevention, MCP, and local project setup. This workshop starts one layer higher. It teaches platform and engineering leaders how to run a context supply chain as company infrastructure: owned, validated, released, rolled back, and measured across teams, repos, and tools. If your team is still learning the basics, foundations training is useful. If you already have agents in use and need consistency, governance, and proof, this is the operating model.

It absolutely can - we've watched a senior developer install a stack of skills, watch every prompt get slower and more expensive, delete them all, and go fast again. That's the trap - and it isn't only about cost: attention is finite even when the window isn't, so the more you pile in, the more the rule that actually matters for the task at hand gets buried and ignored. This system is built around scoped guidance instead: shared rules stay small, domain guidance is routed by the work at hand, only the skills that changed get regenerated, and the pilot tracks whether the right guidance is actually activating. You stop guessing which rules help.

Two layers, and neither is a black box. Scoping - which domains of standards apply in a repo - is deterministic and signal-based: the same repo resolves the same way every time, with an explicit override for monorepos. No embedding lottery, no per-run drift. Which specific rule then activates for a given prompt is the agent's own call from each rule's description - the probabilistic layer - which is exactly why we measure activation and eval it rather than assume it.

Three layers, and we're honest about each. Structure is gated automatically before anything ships, so a change that would break the corpus fails fast. Behaviour is checked with agent evals - multiple trials, rubric-graded - for the rules that justify the cost, run on a schedule rather than as a blunt per-commit gate (agent evals are slow and non-deterministic). And in a pilot, redacted telemetry shows the trend on real work. We don't pretend CI proves quality: CI proves the change is valid, evals catch regressions, telemetry shows the trend.

Real, and newly formalized. A July 2025 academic survey (Mei et al.) synthesized over 1,400 papers into a defined discipline: context as a finite, structured resource you assemble and optimize, not a static prompt. What the literature has barely touched is the operational layer - governing, versioning, and distributing that context across a whole team - which is exactly what this workshop and kit build. Reference: arxiv.org/abs/2507.13334.

Different axis. MCP gives agents capabilities: tools and live resources at runtime. This system governs standards: the company guidance agents should follow across repos and tools. Most mature teams eventually need both. MCP is a capabilities layer; this is the standards-governance layer.

Control stays with the platform team. Changes move through reviewed releases and staged channels, while engineers avoid per-repo update work. The workshop teaches the rollout and rollback motions so the system is governed centrally without becoming busywork for every developer.

Almost nothing. Once, they run a single setup script - or you push it to their machines through your existing MDM / managed-device tooling, so it's already there and they do nothing at all. After that it's invisible: the right standards load into their agent automatically, refreshed before each session. There's no system for them to learn, no versions to bump, no per-repo files to maintain. The platform team owns the pipeline; everyone else just codes.

It is designed to avoid modifying application repositories. The install path is user-level, controlled by your organization, and can be disabled if your rollout policy requires it. The workshop covers the operational boundaries so platform and security reviewers know exactly what is in scope.

Every participant gets a 30-day follow-up q&a call with the instructors and access to the private alumni channel, on top of the templates and the kit you take home. The private workshop additionally includes a rollout roadmap and a 2-week implementation pilot for your organization; we can also set the kit up on your standards as a paid add-on.

Yes. The Private Company Workshop includes a 2-week implementation pilot: we help your owners select a small repo set, define baseline metrics, install the pilot channel, review activation telemetry, and turn the result into a leadership-ready rollout recommendation. The point is not to promise generic ROI; it is to prove whether the system works in your environment.

Comfort with git, a terminal, and JSON, plus a recent Node.js (v20+) installed - every lab tool is a single dependency-free script, so there is nothing else to set up. No specific coding agent or IDE is required.

No. Smaller teams can still get the core value: standards written once, current across their agent setup, and easier to review. Larger organizations use more of the governance, rollout, and measurement layers as they scale.

The Private Company Workshop is the format: from €15,000 (launch price), sized to your company and scoped on a 30-minute planning call - tailored to your platform and including a context architecture review, a rollout roadmap, and a 2-week implementation pilot to prove the rollout in your repos. It works from a handful of platform owners up to a company-wide rollout. We also take direct consulting engagements: an on-site audit of your AI coding setup, a customized adoption approach, and team training.

Private workshops are invoiced: we send a quote after the planning call, then a VAT invoice with your company details and VAT ID. Bank transfer, purchase orders, and vendor onboarding are supported, and card payment is available on request. All listed prices are excl. VAT; EU businesses with a valid VAT ID are reverse-charged (0% invoiced, you self-account).

Not at the moment: we only sell the private company workshop, and individual seats are closed until the second cohort is announced. Put your email in the pricing section and you'll be told when seats open again. If you're employed, the other route is your employer booking a private workshop - the manager brief is written to be forwarded for exactly that conversation.