Global consultancy or specialist?
Two kinds of firm implement enterprise AI in Australia: global consultancies and specialists. They are built for different jobs. The difference that matters is not size or day rate — it is who does the building, how long before anything reaches a real user, and what your own team can do once the engagement ends.
The differences that change the outcome
Who does the building
Partner-led, with delivery staffed by a pyramid of consultants. The people who sold the work and the people who do it are usually different people.
The people who scoped it build it. On our engagements the person in your workshop is the person configuring the connector the following week.
The shape of the engagement
Typically a multi-year transformation programme, with a strategy phase before anything is deployed.
An eight to ten week proof of concept on your own systems and data, with agents live inside the first few weeks. If it is not working you find out in month two, not year two.
What you are left with
A capability that generally continues to require the firm that built it.
Your own staff trained to build and extend agents themselves. If we are still the only ones who can change it when we leave, we have failed.
How success gets measured
Milestones delivered, workstreams completed, phases signed off.
Adoption and hours returned. We report daily active users and time saved against a measured baseline, because a platform nobody opens is not a success.
None of which makes one model wrong. If you are running a genuinely global, multi-year transformation, the large firms are built for exactly that. Most of the work we are asked about is not that.
Ask for adoption numbers
Whoever you choose, ask what proportion of invited users were still using the thing ninety days later. It is the question that separates a deployment from a deck, and it is asked far less often than it should be. Ours, from one of Australia’s largest media organisations, measured against a baseline rather than estimated:
Forty per cent of all platform usage went to agents rather than search, and the deployment scaled from 200 users to 500. We publish these because most engagements report contract value and milestones instead, which tell you what was bought rather than what changed.
Questions worth asking
Should we use a global consultancy or a specialist to implement enterprise AI?
It depends on what you are buying. If you need a multi-year, multi-country transformation programme with hundreds of people, a global consultancy is built for that. If you need working AI agents in production inside a quarter, and your own team able to build more of them afterwards, a specialist will get you there faster and for less. Most Australian enterprises asking the question are in the second category and buying as though they are in the first.
Is a specialist cheaper than a Big 4 firm for AI implementation?
Usually, though the more useful difference is what the money buys. A specialist engagement is typically a fixed-price proof of concept measured in weeks, where most of the cost is people building things in your environment. A large transformation programme carries strategy phases, larger teams and longer timelines before anything is deployed. Compare time-to-first-working-agent rather than day rates.
Can a smaller firm handle an enterprise rollout?
Scale of deployment and size of supplier are different things. We have taken a deployment from 200 users to 500 inside one of Australia’s largest media organisations, with more than 100 agents built in eight weeks. What matters is whether the team has done it before in an environment like yours, not how many people are on their bench.
What happens when the consultants leave?
This is the question worth asking early, and the one most engagements answer badly. Our model is to train your staff to build agents alongside us, so the capability stays. At our largest deployment, client teams were building their own agents before the engagement finished, and 40% of all platform usage had shifted to agents rather than search.
Do we need an AI strategy before we implement anything?
Less than you would be told. A strategy built without touching a real system tends to describe an organisation that does not exist. We would rather connect your systems, put three or four agents in front of real users, and let what happens inform the strategy. If you do want the groundwork first, that is what the AI Readiness Blueprint is for — it is weeks, not quarters.
Who implements enterprise AI in Australia besides the large consultancies?
There is a small group of specialist implementation partners in Australia working directly with the platform vendors. JOURN3Y is a certified implementation partner for Glean and WisdomAI, and builds on Snowflake and Google Cloud, with agents in production across media, banking, telecommunications and asset management.
Tell us the job, and we will tell you what it takes
Including, occasionally, that we are not the right people for it.
.png&w=96&q=75&dpl=dpl_ARd9DoC4YLepaQM16NqynQbv4cJm){kind=logo}