Do you need an AI consultant, or can your team do it yourselves?

Plenty of businesses don't need one. If you have someone keen with a few spare hours a week, you're happy to pick one tool and stick with it, and your first jobs don't involve sensitive client information, your team can get a long way on its own.

A consultant earns the fee when several teams do very different work, the information is sensitive, an earlier attempt stalled, or nobody has the time to lead it.

Below are the signs you can do it yourselves, a plan for the first month if you do, the free help worth using first, and the signs that outside help will pay for itself.

Signs you can do it yourselves

You probably don't need a consultant yet if most of these are true:

  • Someone wants to lead it. One person who's curious about AI and trusted by the team, with two or three hours a week to try things and show others.
  • The work is fairly similar across the business. If most people write the same kinds of emails, quotes and reports, what works for one person will work for the rest.
  • The first jobs are low risk. Drafting internal notes and summarising long documents you wrote yourselves are good places to learn, because a mistake costs very little.
  • You've already picked a tool. Or you're happy to use our Claude vs ChatGPT vs Copilot comparison to choose one in an afternoon.
  • You'll check the results. You're prepared to compare how long a job took before and after, and drop anything that doesn't save time once checking is counted.

Signs outside help will pay for itself

Outside help usually makes sense when one or more of these apply:

  • Nobody has the time. The person who'd lead it is the owner, and the owner is already flat out. A plan that sits in your head for six months costs more than help would.
  • Different teams do very different work. What helps sales won't help accounts, and building separate training for each team is a real job.
  • Your clients trust you with sensitive information. Health, financial or legal details need clear rules and the right settings before anyone puts them into an AI tool.
  • You tried, and it stalled. A handful of people still use it and nobody can say why the rest stopped. An outsider can usually see the reason faster, partly because staff will tell them things they won't tell the boss.
  • You're about to spend real money. If you're weighing up seats for the whole team, or someone is pitching you a custom build, a second opinion before you sign is cheap by comparison.

Free tool AI Readiness Check: ten quick questions, and you'll see the one step that would help your business most. About two minutes.

A do-it-yourself plan for the first month

If the first list sounds like you, this is the order we'd follow.

Week 1: pick one tool and one owner. Put everyone who'll use it on a business plan so you get admin controls and a central bill. As shown on their Australian pricing pages in November 2026, including GST, Claude Team costs A$34 a person a month on yearly billing and ChatGPT Business costs A$28.

Australia's Guidance for AI Adoption suggests naming a senior leader as the overall owner of AI as one of the first steps, so name that person now.

Also in week 1: write down what's safe to put in. A one-page policy covering the approved tool, what client information can go in, and who checks work before it goes out. Our free AI use policy template takes about twenty minutes to adapt.

Week 2: pick two or three jobs per team. Choose jobs people repeat every week that take longer than they should. Time how long each takes today. That number is your baseline.

Weeks 2 and 3: practise on real work. Get people trying the tool on those jobs with their own examples, and have the owner share what works in a short weekly catch-up. Save the instructions that work so nobody starts from scratch, using Projects in Claude or ChatGPT.

Week 4: check what changed. Time the same jobs again. Keep what saves time and drop what doesn't, then pick the next jobs. Ask people to show you their best use rather than tracking who logs in.

  1. Week 1One tool, one owner, and a page on what's safe to put in
  2. Week 2Pick two or three jobs per team and time them
  3. Weeks 2 and 3Practise on real work and save the instructions that work
  4. Week 4Time the same jobs again. Keep what saves time
The first month, if you're doing it yourselves.

Free help worth using first

Before you pay anyone, these cost nothing:

  • Our AI readiness check. Ten questions, about two minutes, and it points to the one area to work on first.
  • The AI Time Savings Calculator. It estimates the hours your team could get back, and what that time is worth.
  • The Guidance for AI Adoption foundations, published by the National AI Centre on ai.gov.au. It's written for organisations starting out and includes a policy template.
  • The Australian Government's AI Adopt Centres. They offer free training, one-on-one advice and roadmaps to eligible small and medium businesses that trade interstate or overseas in the National Reconstruction Fund's priority sectors. Details are on business.gov.au.
  • The vendors' own help centres. Anthropic and OpenAI both publish setup guides for admins.

The middle path: buy only the piece you're missing

Doing it yourselves and hiring a consultant for everything aren't the only options. Many teams do most of the work in-house and pay for one piece:

  • Only the plan. If you're not sure where to start, an assessment shows which jobs to tackle first, and your team runs with it. Our AI Opportunity Assessment starts from $1,000 + GST.
  • Only the training. If you have the tool and a plan but the team isn't confident, one hands-on workshop on their own work gets them over the first hump. Our AI Workshops start from $3,500 + GST for a half day.
  • Only a sounding board. If you're leading it yourself and want someone to check in with, a monthly advisor costs less than a project. Our Ongoing AI Advisor starts from $1,500 + GST a month.

These are starting prices, not a quote. Every business gets its own fixed price once we know what it needs.

What doing it yourselves really costs

Doing it in-house swaps a consultant's fee for your own time, and owner time is usually the scarcest thing in the business. Weigh it against what's at stake. In Thomson Reuters' 2025 survey of 2,275 professionals, respondents predicted AI would save people about five hours a week within a year. That's a forecast, not a measured result. For a team of eight, that would be around 40 hours a week.

If doing it yourselves gets you there in a month, keep going. If six months pass and you're still at the starting line, the time lost is worth more than help would have cost.

Common questions

Want a hand?

Not sure which side of the line you're on? Take the free AI readiness check, or book a free chat and we'll tell you straight whether you need us, and if so, for which piece.