Work with the practice
Two shapes of engagement: fixed-scope packages you can start from a DM, and custom advisory for organizations already living inside decision systems that affect real lives.
The system becomes easier to see, and harder to excuse.
Advisory turns a strained workflow, vendor claim, grievance route, or governance principle into something teams can inspect, discuss, improve, and hold accountable.
Find your door
Describe the live AI-shaped pressure you are facing and get routed, in real time, to the fixed-scope offer that fits it.
Pick the door by the pressure, not the product name.
Entry-level - $99 USDVendor Interrogation PackA fixed-scope, asynchronous first rung for vendor pressure: six questions, templates, and one response review.
Vendor Interrogation Pack
Six questions your AI vendor cannot refuse to answer, with one forensic read of what they say back. A fixed-scope, finite, asynchronous artifact: 12-page PDF, two email templates, and one written response review within five working days. Use it before the procurement meeting, not after.
- A live AI vendor selection sitting in front of procurement, legal, or ESG.
- You felt your gut tighten on “95% accuracy” and want the structural questions the deck did not answer.
- You want a tool to send your vendor this week, with no discovery call to schedule.
If the pack uncovers something live and you need a structured outside read, the next rung is the Systems Briefing ($997).
A defined deliverable, quickly.
Productised offerings with a defined deliverable, a fixed price, and a short turnaround. Start here when there is one messy workflow, live decision, vendor claim, or governance question that needs an outside read.
A confidential 1:1 working session on the AI-shaped pressure in your system - built around a written brief delivered before the call.
Stress-test the alwayses and nevers in an AI policy, pledge, covenant, vendor assurance, or governance document before hostile readers do.
Stress-test a resettlement plan, ESIA chapter, grievance procedure, or conformance mapping for where AI in the workflow has weakened it before a lender or an inspection panel does.
Stress-test a results framework, evaluation report, or evidence synthesis for where AI in the MEL workflow has broken the chain from reported result back to source record.
Clarify what an AI vendor is asking your organization to trust before those claims become workflow, contract, or institutional fact.
Turn dense expert material into a clear, publish-ready narrative for AI governance, systems thinking, ESG, and institutional design.
A deployable hub for the briefing offer: visual essays, companion notes, and a recorded walkthrough.
The free anchor under the reads: AI in Safeguards Work: An Interim Protocol is the discipline the Defensibility and Evidence Chain Reads enforce, free to adopt and citable in bids.
Pick the door by the pressure, not the product name.
- Systems Briefing when the live issue is ambiguous, the deadline is real, and someone needs outside judgment before the next meeting.
- Always/Never Audit when the issue is a specific AI governance document whose promises need an adversarial read before buyers, regulators, workers, or counsel test them.
- Evidence Chain Read when the issue is a results framework, evaluation, or evidence synthesis heading into funder verification with AI somewhere in the pipeline.
- Safeguard Defensibility Read when the document is a resettlement plan, ESIA, or conformance mapping that has to survive lender due diligence and AI has been in the workflow.
- AI Vendor Review when a vendor claim is about to become contract, workflow, or policy and the procurement question is sharper than the strategy question.
- Content Sprint when the argument is already settled in your head and the bottleneck is turning it into a usable public or internal asset.
- Custom Advisory (see below) when the issue spans multiple actors, workflows, or governance layers and the right shape is not yet a single deliverable.
Specific clients and outputs stay confidential unless cleared for publication, but the deliverable categories are stable across engagements: a scoped recommendation in writing, a usable artifact for the next meeting (brief, memo, vendor-question list, decision map, or escalation route), and a clear next move with named owners. Where useful, a follow-up review checkpoint is built in before the work hardens into operating habit.
Small enough to start, concrete enough to use.
The first step is not a giant discovery exercise. It is a short scoping pass around the system, deadline, and decision pressure already in front of you.
You send the live question
Share the workflow, vendor claim, governance issue, or reporting channel that is starting to matter.
Scope the right shape
You get a short written recommendation: fixed-scope briefing, vendor review, pilot, or custom advisory.
Align before work begins
A focused call confirms the deadline, useful evidence, decision owner, and what the output must help you do.
You leave with usable artefacts
The output is built for the meeting, procurement decision, governance discussion, or workflow fix that comes next.
For decision architectures already under strain.
Scoped in conversation around the system you are assessing and the nearest deadline that matters. Useful when the process looks orderly, but the consequences underneath are starting to lose their shape.
Map the decision architecture around a live system: who sees what, what gets preserved, where accountability sits, and where judgment needs support.
Design grievance flows that preserve voice, protect callers, create usable records, and make escalation meaningful when a complaint reveals operational risk.
Contract-level mechanisms for AI vendor agreements: evidence, override, provenance, refusal rights, and accountable continuation.
Ask about advisory
Answered from the practice’s own material, with sources.
Ask about advisory
Answered from the practice's own material, with sources.