CODITECT - the AI operations layer

Adopt AI across the enterprise with compliance, risk, and audit built in.

CODITECT records every AI-touched change against an approved task and a regulatory control today. Compliance evidence is produced by the work itself, not assembled before an audit. Below: a 3-minute demo of one engineering use case. The same model applies across the application stack - the harness analyses any target codebase regardless of language or framework per ADR-320.

Watch the 3-minute demo
The problem - in Mark Walker's words
"We and every other software company in the world are outstripping our ability to test what we're building."

Why now: the velocity of agentic coding has decoupled from the velocity of testing, auditing, and validation - the knowledge and proof that AI agents did what they were tasked to perform, i.e. testing, in this case. An AI agent can produce more code in a day than a team used to write in a sprint. The test, audit, and compliance layers did not get faster at the same rate. The gap is structural and widens with every model release.

Three deficiencies - in every company today - that no software addresses:

  • determining which tests need to run for a particular release
  • checking whether they ran
  • recording the outcome

Mark Walker, nue.io - meeting transcript [00:46:36]

CODITECT is the system that addresses all three deficiencies.

Live execution dashboard: State Bus on the left tracking the eight Blueprints from PENDING to DONE, terminal on the right running the agent.
Live execution dashboard - State Bus tracks each Blueprint from PENDING to DONE while the agent runs.

Who this is for

ICP

High-growth software companies doing agentic coding.

Mark recommended two ICPs - high-growth software companies and regulated industries - and added: “I don’t think you need to solve both of them.” This demo covers the first.

The problem - in Mark’s words

“We can code faster than we can test… there is no system for determining which tests needed to be run for any particular release or checking whether they were run or recording what the outcome was.”
Mark Walker, nue.io - meeting transcript [00:46:36]

Mark reframed this in the same conversation: it’s a code quality problem, not an engineering problem. The faster a team moves - especially agentically - the harder it is to know whether quality best practices are being followed, including by the agents themselves.

“Are the agents themselves following the quality rules?” [00:55:16]

The solution - what the demo shows

  1. An agent writes the new code - five files in a single burst.
  2. The system selects the four tests affected by the change, with a written reason for each.
  3. The system writes four more tests for situations the engineers did not write, including one designed to fail.
  4. All sixteen tests run. The deliberately failing test does fail, with the exact test, line, and change reported.
  5. Every change is recorded with the approving engineer’s name and a timestamp. Human or agent - same record, same review.
  6. The four audit documents an auditor would normally request are produced as a byproduct of the run. Each is timestamped and signed.
  7. Everything packages into one audit bundle.

Product - in Mark’s framing

CODITECT runs as “a quality management system for a software company that was doing agentic coding.” [00:46:36]

The system handles the work that nobody is doing today:

  • Determining which tests need to run for any particular change or release.
  • Checking whether they were actually run.
  • Recording what the outcome was - in an immutable audit trail.
  • Producing the IQ, PQ, and traceability documentation as a byproduct, not as paperwork on top.
“It just has to be a quality process that goes really fast.”
Mark’s hard success criterion - [01:11:21]

Velocity is the gate. The demo is built to show the QMS layer keeping pace with the agent - not bottlenecking it.

Watch the 3-minute demo

If you have not signed in yet, you will be asked for the access token from your email.

Watch the demo