AI Strategy Consulting
A costed, sequenced AI plan for your business — written by the people who will build it. Independent of every vendor, and honest about what you should not do.
Most AI Strategies Die in the Deck
Not because the thinking was wrong. Because nobody in the room had to build the thing afterwards.
The pattern is familiar to anyone who has been through it. A firm arrives, runs a set of workshops, and produces a handsome document with a maturity curve, a set of horizons and a portfolio of initiatives. Everyone nods. The document goes in a drawer. Eighteen months later the same business is still doing the same manual work, and the only measurable output of the exercise is the invoice.
The failure is structural, not moral. A strategy written by people who will never implement it has no feedback loop. Nothing in it is disciplined by the question could we actually ship this next quarter, with this team, on top of these systems? So the initiatives are described at an altitude where they all sound feasible, and the hard parts — the integration that does not have an API, the data that is riddled with blanks, the approval step that legally cannot be automated — never surface until someone tries.
We work the other way around. We are a build shop that does strategy, not a strategy shop that subcontracts builds. Every recommendation we make has to survive the fact that we may be the ones delivering it, which is a remarkably effective filter. It kills the initiatives that only work in a slide, and it forces us to be specific about sequence, cost and dependency in a way that is uncomfortable to write and extremely useful to own.
It also means we are willing to write down the unwelcome conclusion. Sometimes the highest-value recommendation is to buy a $200-a-month product instead of building anything. Sometimes it is to fix a process first. Sometimes it is to wait six months. An independent adviser with no software to sell can say those things plainly, and that independence is the entire point of hiring one.
What an AI Strategy Engagement Covers
Six areas. Each one produces something you can act on, argue with, or take to a board.
Opportunity Mapping
We interview across your functions and inventory every process where AI could plausibly help — then score each one on value, feasibility and time-to-value. Most businesses end up with 25 to 40 candidates and are surprised by which ones rank.
- Function-by-function process inventory
- Value and feasibility scoring on a consistent scale
- Explicit rejection list, with reasons
- Quick wins separated from structural plays
Business Case & Economics
Every shortlisted initiative gets a costed model: build cost, run cost, expected benefit, payback range and the assumptions that drive it. If an assumption is a guess, we mark it as a guess.
- Build, run and token/inference cost modelling
- Benefit stated in hours or dollars, not percentages
- Sensitivity on the two or three assumptions that matter
- Payback ranges, not single-point fantasies
Build vs Buy vs Wait
For each opportunity we make a call: buy an off-the-shelf product, build something custom, or wait because the market will solve it for you within a year. Naming the product is usually cheaper advice than building it.
- Named products where buying clearly wins
- Custom build recommended only where it earns its keep
- "Wait" as a legitimate, argued recommendation
- No reseller margin, no partner tier, no commissions
Data & Systems Reality Check
The honest constraint on most AI plans is not model capability — it is that the data lives in six systems and one spreadsheet. We audit what you actually have and what each opportunity truly requires.
- System-of-record and integration surface review
- Data quality assessment against specific use cases
- Blocked opportunities identified early, not mid-build
- Costed remediation where the fix is worth it
Operating Model
Who owns AI once the consultants leave? We recommend a model that fits your size — which usually is not a centre of excellence, because most Australian mid-market businesses do not have the people for one.
- Internal ownership and accountability model
- Capability gaps and how to close them
- Where to use partners vs hire vs upskill
- Realistic for your headcount, not a Fortune 500 template
Governance Guardrails
Enough governance to move safely and not one page more. Privacy Act 1988 obligations, acceptable-use rules for staff, and a risk position for each initiative — set before the first build, not after an incident.
- Australian Privacy Principles mapped to your use cases
- Staff acceptable-use policy covering public AI tools
- Risk rating per initiative with mitigations
- Aligned to Australia’s Voluntary AI Safety Standard
What You Walk Away With
Concrete artefacts, not a summary of the workshops you already attended.
Scored Opportunity Register
Every candidate process, scored on value and feasibility, with the rejected ones and the reasons they were rejected. This is the artefact clients keep using two years later.
Costed Business Cases
Build cost, run cost, expected benefit and payback range for the top three to five initiatives — with the assumptions exposed so your CFO can attack them properly.
Sequenced Roadmap
What to do first, second and third, and why that order. Dependencies mapped, so nobody starts a build that is blocked by a data fix nobody scheduled.
Build vs Buy Decisions
A recommendation on each initiative, with named products where buying wins. We would rather point you at a $200/month tool than invoice you to rebuild it.
Governance Baseline
An acceptable-use policy for staff, a risk position per initiative, and your Privacy Act 1988 obligations mapped to the specific use cases in the plan.
Executive Presentation
We present the findings to your leadership team and take the questions live. The document is the record; the conversation is where the decisions actually get made.
How the Engagement Runs
Three to six weeks for most businesses. Delivered remotely across Australia, in person around Melbourne where it helps.
Free Consultation
An hour with you, no deck. We work out whether there is anything here worth doing and tell you honestly if there is not. Plenty of these conversations end with a recommendation and no invoice.
Discovery & Interviews
Time with your leadership, your system administrators and — critically — your frontline staff. We look at your actual systems, not a diagram of them. The truth about your data lives with whoever maintains it.
Modelling & Prioritisation
We score every opportunity, build the cases for the shortlist, and make the build-buy-wait call on each. This is where most of the unwelcome findings surface, and we bring them to you early rather than at the presentation.
Presentation & Handover
We present to your executive team, take the hard questions live, and hand over the working models. If you want us to build the first initiative, we can. If you want to take it in-house or to another firm, the document is written so you can.
Where This Fits
Strategy is one piece. Most clients arrive via the audit and leave via an implementation.
AI Readiness Assessment
The $3k AI Opportunity Audit. A compressed, scored read on where AI would pay for itself in your business — and where it would not.
See the auditAI Roadmap Development
Turning a set of good initiatives into a sequenced twelve-month plan with owners, dependencies and budget by quarter.
Build the roadmapAI Implementation
The part most strategy firms subcontract. We build, integrate, evaluate and hand over systems that run in production.
Ship itFrequently Asked Questions
What Australian executives ask before commissioning AI strategy work.
Two structural differences. First, we are vendor-independent — we do not resell any platform, we take no referral commissions, and we have no partner tier to protect. When we tell you a build is a bad idea, there is nothing in it for us. Second, we build. The people writing your strategy are the people who will ship the systems, which changes what gets recommended. Teams that never have to deliver tend to recommend things that sound good in a boardroom and fall apart on contact with your actual CRM. A large firm will typically field a partner for the pitch, a manager for the workshops, and analysts for the deck. You get a small senior team from start to finish.
A written strategy document, but the useful part is what is in it. You get a prioritised opportunity register scored on value and feasibility; a costed business case for the top three to five initiatives with realistic assumptions and payback ranges; a build-versus-buy call on each, with named products where buying wins; a data and systems reality check that says which opportunities are blocked and what unblocks them; an operating-model recommendation covering who owns AI internally; and a governance baseline covering privacy, risk and acceptable use. It usually runs 30 to 50 pages. We also present it to your executive team and leave you with the working models, not just the PDF.
For most businesses, three to six weeks from kick-off to the executive presentation. That covers stakeholder interviews across your functions, a review of your systems and data, opportunity workshops, business-case modelling and the write-up. Complex groups — multiple business units, several ERPs, regulated data — run longer, typically six to ten weeks. If you need a faster read, the AI Opportunity Audit is the compressed version: around three weeks and roughly $3,000, aimed at businesses with about 20 or more staff who want a prioritised shortlist before committing to a full strategy.
No, and waiting until your data is tidy is how organisations lose two years. Data readiness is one of the things the strategy assesses, not a prerequisite for it. In practice most Australian mid-market businesses have exactly the same picture: a decent transactional system, a CRM that is only partly populated, and a large amount of institutional knowledge trapped in email, PDFs and people’s heads. A good strategy works out which opportunities are viable on the data you have today, which need a defined and costed data fix first, and which are genuinely years away. That sequencing is most of the value.
Less time than you expect, but from the right people. We need roughly two hours from the CEO or managing director, an hour each from the heads of the functions in scope — typically operations, finance, sales and customer service — and two to three hours from whoever actually administers your core systems, because that person knows the truth about your data. We also want time with two or three frontline staff per function. The people doing the repetitive work always know where it is, and they are consistently the best source of opportunities in the entire engagement.
Then we say so, and that is a legitimate outcome. Sometimes the real bottleneck is a broken process, an unfit system of record, or a pricing problem that no amount of AI will fix — automating a bad process just produces bad outcomes faster. We have told businesses that their highest-value next step was replacing a piece of software, fixing their quoting process, or doing nothing at all for six months. You are paying for an independent read, not for permission to spend money. A strategy that talks you out of a bad build has paid for itself several times over.
The initial consultation is free — an hour, no deck, no obligation, and you will get a direct opinion on whether there is anything worth doing. From there, the AI Opportunity Audit is around $3,000 and suits businesses with roughly 20 or more staff; it gives you a scored opportunity register and a recommended first build. A full strategy engagement is quoted on scope after that conversation, because a single-site services business and a four-division group are not the same job. We quote fixed price, not day rates, so the scope is agreed before anyone starts.
Start With an Honest Hour
The first consultation is free and costs you nothing but the time. You will get a direct opinion on whether AI is worth your money right now — including if the answer is no.