For Chief AI Officers and executive teams
Many Chief AI Officers were appointed in a hurry, on top of an existing job, with no dedicated budget. They inherit a stack of generative AI pilots nobody can measure, and older models already making decisions that nobody has reviewed in years. We help you turn the title into a working mandate, and early pilots into results your executive and board can see.
The first six months
Authority
I'm accountable for AI across the organisation, but I don't control the budget, the teams or the systems.
Mandate
Nobody can tell me where my job ends and the CIO's, the CDO's or the risk team's begins.
Value
I've got a dozen pilots and no honest way to say which ones are working.
Production
The pilot worked. Six months on it is still a pilot, because nobody owns getting it into operation.
Data
The tools are serving up old policy and documents nobody owns. Staff stop trusting it after the first wrong answer.
People
Everyone has done the mandatory module. Half of them are wary of it and the other half are using it badly.
Tools
Every vendor is pitching me a pilot and divisions are buying their own. I don't know what's actually in use.
Capacity
This was added to my existing job, with no extra staff. The strategic work keeps falling off the list.
Obligations
Risk registers, impact assessments, board questions, and now the December Privacy Act changes. The goalposts keep moving.
Experience
I was chosen for my seniority, not my AI background. I can't always tell when a vendor is overselling.
Inherited systems
Everyone wants to talk about Copilot. Nobody can tell me what our older models are deciding, or whether they still work.
Direction
We have AI projects everywhere and nobody can say where it is all meant to take us.
For the executive team
Emeris gives executive teams an independent view of their AI capability, built with the people responsible for AI across the organisation. Your Chief AI Officer, or whoever carries that role, leads the work with us. Any executive can sponsor it: a chief risk officer, a chief operating officer or the head of a business area. You get a clear view of what to fix, fund or stop.
Which pilots are earning their keep, measured the same way each time.
Older models making decisions nobody has reviewed, tools divisions bought on their own, and obligations that are already arriving.
A prioritised plan, and the decisions that need your authority to stick.
The pattern
The pilot that never reaches production is usually a process or people problem. The governance framework nobody follows is usually an engagement problem.
Years of building and governing production AI where mistakes are public and audited taught us that AI programs succeed or stall across six connected pillars. Fix one in isolation and it rarely holds.
01
Accountability without authority. Decisions get made, then quietly ignored.
02
Training completed, but few people using AI well in their actual jobs.
03
Pilots that start easily and never finish, with no way to tell which are earning their keep.
04
Tools surfacing outdated content, and several divisions buying their own.
05
Vendors setting the agenda one pitch at a time.
06
No shared vision or strategy, so AI gets applied without a destination, and staff are unsure why it is happening or whether to trust it.
Start here
A quick self-check on your own, or the full review with your executive team. Both use the same six pillars, and both cover generative AI, the tools staff use every day, and the older models already making decisions.
Free · About 10 minutes
Twelve questions across the six pillars, plus a few about your role, for the person responsible for AI. You get an indicative scorecard, the constraint most likely to stall your program, a plain description of what to ask for if you commission help, and a written assessment from the principal within two business days.
Fixed fee · 2 to 3 weeks
An independent view of your organisation's AI capability for the executive team, built with the people responsible for AI.
Founding cohort
We are running the Capability Review with a small group of organisations at a reduced fee, in exchange for candid feedback on the method. Each needs a named executive sponsor alongside the leader responsible for AI. About 2 to 3 weeks; fee on request.
How we work
Most organisations get outside help with AI in one of two ways. We offer a third: defined services, each tied to a problem you would recognise, taken in the order that suits you.
Months of one person's time, made responsible for everything at once. What stays behind depends on the handover.
Attendance and a completion rate, with little change to how AI is governed or delivered.
Fixed-scope services, each with a defined deliverable, a fixed fee and a set duration. One framework sits behind them and is applied to your organisation. Take one, or build them up piece by piece.
Services
Every module is scoped, fixed-fee and designed to leave capability behind, not dependency. Each stands alone, so you can start with the one that matches your most pressing problem, or let the review suggest the order. The smaller ones are designed for a short internal approval path. Work commissioned by a business area is delivered in a form the central AI team can adopt. These three are where most organisations start.
Governance
Your decision rights written down and agreed at the top: a one-page charter, a map of who owns what across technology, data, risk and the business, and a paper asking for the authority the role needs.
2 to 3 weeks · fixed fee
Process
A defensible story on value. We triage your pilots into scale, fix or stop, and set up value measures, financial or public, that you can report against each quarter.
4 to 6 weeks · fixed fee
Governance
Exactly what applies to you, and by when. From 10 December 2026, organisations covered by the Privacy Act that use personal information in automated decisions that could significantly affect people must explain this in their privacy policy. Plus AI registers, impact assessments, and the expectations of your regulator or central agency.
Fixed scope · fixed fee
See all eleven modules, by pillar
Ongoing
The machinery run for you: committee secretariat, use-case intake, registers, reporting and horizon scanning. You lead; we keep it moving. Monthly retainer.
Included
Every engagement leaves your AI use cases documented against NIST AI RMF and ISO/IEC 42001, ready for transparency statements, audit and the next executive question.
Why Emeris
Emeris is led by a former senior executive who spent a decade building, funding and governing production AI in border protection and security vetting, after sixteen years in national security. High stakes, real consequences, and nowhere to hide when a model gets it wrong.
We have sat in your chair, argued for the budget, killed projects that should not ship, and lived with the systems long after launch. That is what the review is built from.
A machine learning capability that won the 2024 APS Data Science, Analytics and AI Award.
Through an organisation-wide AI and data program, later extended to partner organisations here and overseas.
A team of up to 45, with AI in production for targeting, vetting and detection. Not on slides.
Not tied to any vendor or platform. Our method names no product, so it works with the tools you already have. When you need design, training or delivery capacity, we bring in partners we have chosen and stand behind.
Get in touch
Leave your details and the principal will contact you within one business day to arrange a time that suits you, in person in Canberra or by video. No pitch, and no calendar links to chase.
Prefer to start on your own? Take the self-check. It takes about ten minutes.
Line of Sight
A fortnightly brief for people who have to answer for AI rather than announce it. One argument at a time, worked through properly: what is going wrong at the production boundary, why, and what it means for whoever is accountable. No news roundups, no vendor commentary, no predictions we would not put our name against.
Fortnightly · Unsubscribe any time · No list sharing