AI consulting is built for enterprises. If you're not one, there's a better option.

AI consulting is a legitimate industry that helps businesses navigate artificial intelligence. Consulting firms run discovery workshops, interview stakeholders, assess technology stacks, and produce strategy documents with implementation roadmaps. For large enterprises with complex needs, it works. For everyone else, it's the wrong tool for the job.

This guide covers what AI consulting actually is, what it costs, how engagements work, and why the consulting model breaks down for small and mid-size businesses. I also explain what I chose to do instead and why it matters for business owners who are searching for help with AI but don't need a 40-page report to get there.

The perspective here comes from daily work with Claude Code, running 14 specialist agents, and making AI implementation decisions in real time at Northbase and across other businesses. Advisory over consulting, because the model is faster for the businesses that need it most.

What is AI consulting?

AI consulting is a professional service where external experts help a business identify opportunities for AI, develop a strategy, and guide implementation. It follows the same model as management consulting, IT consulting, or strategy consulting. A firm comes in, studies your business, and tells you what to do.

A typical AI consulting engagement looks like this:

Discovery phase. The consulting team spends 1-3 weeks learning your business. They interview leadership, review your tech stack, map your processes, and identify where AI could add value. This phase is useful. Understanding the problem before proposing solutions is how good consulting works.

Assessment and analysis. The team evaluates your data infrastructure, existing tools, team capabilities, and competitive position. They assess your "AI readiness" using whatever maturity framework the firm prefers. This produces the first set of deliverables, usually a current-state assessment and gap analysis.

Strategy and recommendations. Based on the discovery and assessment, the team builds an AI strategy document. This covers recommended use cases, technology selections, implementation priorities, timeline, budget estimates, and risk considerations. It's typically 20-60 pages. It's thorough. It takes weeks to produce.

Implementation support. Some firms offer implementation as a separate engagement. Others hand off their strategy document and wish you luck. The best consulting relationships include ongoing support. The most common ones don't.

That's consulting. It's structured, thorough, and designed for organisations that need documentation, stakeholder alignment, and a defensible process for making large technology investments. There's nothing wrong with the model itself. The question is whether it's the right model for your business.

What does an AI consultant actually do?

Day to day, an AI consultant's work breaks down into a few categories.

Stakeholder interviews. Sitting down with department heads, IT leads, operations managers, and executives to understand pain points, workflows, and priorities. This is relationship management as much as it is research. A good consultant synthesises conflicting priorities across departments into a coherent strategy.

Technology evaluation. Reviewing AI tools, platforms, and vendors to determine what fits your specific needs. Comparing options. Testing capabilities. Assessing integration requirements. This is where hands-on experience matters enormously, and where many consultants fall short. Evaluating tools you've never used is different from evaluating tools you use every day.

Documentation. Reports, slide decks, roadmaps, architecture diagrams, cost-benefit analyses, risk assessments. Consulting firms produce a lot of paper. Some of it is valuable. Some of it exists because the engagement contract says deliverables are required and more pages feels like more value.

Workshops and training. Running sessions to upskill internal teams, align stakeholders on strategy, or demonstrate AI capabilities. This is often the most practically useful part of an engagement because it builds internal capability rather than just producing a document that sits on a shelf.

Implementation guidance. Advising on architecture decisions, vendor selection, integration approaches, and deployment. The best consultants stay involved through implementation. The most common outcome is a strategy document that gets handed to an internal team or a different vendor to actually execute.

Here's what's good about this model: it's thorough. A good consulting engagement covers angles you wouldn't have considered. It forces structured thinking. It produces documentation that helps with internal buy-in and board-level reporting. For a 500-person company making a $2 million AI investment, that rigour is justified.

Here's what's bad about it: it's slow. A typical engagement takes 4-12 weeks. In AI, the tooling shifts meaningfully every month. A tool recommendation made in week two might be outdated by week eight. The consulting cadence was designed for industries that move at annual cycles. AI moves at weekly cycles. That mismatch is real and it costs businesses money.

How much does AI consulting cost?

This is what most people searching for "AI consulting" actually want to know, so here are real numbers.

Big 4 firms (Deloitte, PwC, EY, KPMG) charge enterprise rates because they serve enterprise clients. A basic AI strategy engagement starts around $50,000 and scales quickly into six figures. Multi-department implementations regularly exceed $500,000. You're paying for brand credibility, structured methodology, and a team of 5-15 people. The partner who sold you the engagement is not the person doing the work. Junior analysts and associates handle most of the research and documentation.

Boutique AI consulting firms typically charge $25,000-$100,000 per engagement. They're smaller, often more technical, and can move faster than the large firms. The quality varies enormously. Some boutique firms are staffed by people who were building AI systems before it was fashionable. Others formed in 2024 when every consulting firm decided they needed an "AI practice."

Independent AI consultants charge $200-500 per hour in Australia (higher in the US). The advantage is direct access to the person with the expertise. No junior analysts doing the legwork. The disadvantage is capacity. An independent consultant can only serve a few clients at once, and many of the best ones are booked months ahead.

For Australian businesses specifically: the market is smaller than the US and UK, so pricing can be less transparent. Expect to pay a premium for consultants with real AI implementation experience rather than just strategy credentials. There are fewer of them.

The pattern across all these tiers is the same. You're buying time. The consulting model charges by the hour, by the day, or by the project. The more complex the engagement, the more hours it takes, the more it costs. That's the model working as designed. The question is whether you need all those hours or whether you need something different entirely.

Why AI consulting doesn't work for most businesses

If you're running a business with 10-200 employees and you're googling "AI consulting," you're looking for the right thing in the wrong category. Here's why.

The consulting model assumes you need discovery. A consulting firm spends weeks understanding your business before they can tell you anything. That makes sense if you're a 2,000-person company with 15 departments and a legacy tech stack that's been accumulating for 20 years. It doesn't make sense if you're a 30-person company with a clear problem. You don't need someone to spend three weeks discovering that your sales team is drowning in manual follow-ups. You know that. You need someone to tell you what to do about it.

The deliverables don't match the need. A 40-page strategy document is a liability for a small business, not an asset. You don't have a strategy team to interpret it. You don't have a PMO to execute the roadmap. You don't have six months to implement recommendations in phases. You need answers. Short, clear, actionable. "Use this tool. Set it up this way. Skip that one. Here's why." That's a conversation, not a document.

AI moves faster than the consulting cadence. I've written about this in detail, but it bears repeating. A consulting engagement that runs for 8 weeks will produce recommendations that are partially outdated by the time they're delivered. New models ship monthly. New tools launch weekly. Capabilities that didn't exist at the start of an engagement may fundamentally change the recommendation by the end of it. The consulting model was built for industries that sit still. AI does not sit still.

Junior-heavy delivery is a structural problem. Large consulting firms sell on the reputation of their partners and deliver with their graduates. The partner who impressed you in the pitch meeting shows up for the kickoff and the final presentation. Everything in between is handled by analysts who may have strong general consulting skills but limited hands-on AI experience. You're paying partner rates for associate work. In a field where practical experience is everything, that gap matters.

The cost doesn't scale down. Consulting firms have overhead. Offices, support staff, methodology development, sales teams. That overhead gets baked into every engagement. A $50,000 minimum might represent reasonable value for a large enterprise. For a 40-person company, it's most of the annual technology budget spent before a single tool gets implemented. The economics don't work at the SMB scale.

AI consulting vs AI advisory

I chose to build an advisory practice instead of a consulting practice. Not because consulting is bad. Because advisory is a better fit for the businesses I want to serve.

Here's how the two models compare:

The fundamental difference is what you're paying for. Consulting sells hours. Advisory sells judgment. A consultant's value is proportional to how much time they spend on your problem. An advisor's value is proportional to how quickly they can solve it.

An analogy that works: consulting is like hiring an architecture firm to design your house from scratch. They survey the land, draw plans, manage approvals, and charge by the hour for all of it. Advisory is like hiring an architect who's already built 50 houses on similar land, walking through your site, and telling you exactly what to build and what not to. Same expertise. Different delivery model. Very different cost and timeline.

I'm not pretending this comparison is neutral. I chose advisory because I think it's the better model for the businesses I work with. But the comparison is honest. If you're a large enterprise with complex needs, consulting might be the right call. For most businesses under 200 people, it's overkill.

When you actually do need an AI consultant

Being fair about this matters more than being persuasive. There are real situations where consulting is the right model.

Large-scale implementations. If you're deploying AI across multiple departments in a 500+ person organisation, you need a team. One person can't manage the stakeholder alignment, change management, technical integration, and training simultaneously. A consulting firm brings capacity that an individual advisor can't.

Regulated industries. Healthcare, financial services, government. If your AI implementation needs compliance frameworks, audit trails, risk assessments, and documentation for regulators, the consulting model's thoroughness is a feature, not a bug. The paperwork isn't overhead. It's the deliverable.

Multi-department rollouts. When AI adoption needs to happen across sales, operations, marketing, finance, and customer service simultaneously, the coordination alone is a full-time job. Consulting firms have project management infrastructure built for this. An advisor gives you the strategy. A consulting firm gives you the strategy and the people to execute it across the organisation.

Board-level justification. Sometimes the deliverable isn't the strategy itself. It's the credibility of who produced it. A Big 4 firm's name on a strategy document carries weight with boards and investors in a way that an independent advisor's doesn't. That's not about quality of advice. It's about organisational politics. It's real and pretending otherwise would be dishonest.

You need ongoing implementation support with dedicated headcount. If the project requires 5-10 people working for 6 months to build and deploy, that's consulting. Advisory can set the direction. Consulting can provide the team to execute it. And if your business is deploying AI agent systems specifically, an agent operator is the role that bridges strategy and execution.

How to evaluate AI consulting or advisory for your business

Before you spend money on either model, answer these questions honestly:

What's your actual budget? If it's under $25,000, traditional consulting firms won't give you their best work. You'll get a junior team or a templated deliverable. At that budget, independent consultants or AI advisory will give you more value per dollar.

What's your timeline? If you need answers in weeks, not months, consulting's discovery and delivery cycle won't fit. Advisory is built for speed. If you have 6-12 months and a complex implementation ahead, consulting's structured approach has merit.

How complex is your situation? A single business unit adopting AI tools is a focused problem. An advisor can handle that in a session. Ten departments with different needs, legacy systems, and compliance requirements is a complex problem. That might need a team.

Do you need a document or a decision? If you need a strategy document for your board, hire a consulting firm. If you need someone to tell you what to do so you can start doing it, hire an advisor. Know which one you actually need before you start looking.

Who will execute? If you have an internal team that can execute once they know the direction, advisory gives them the direction. If you don't have the internal capability and need external people to build and deploy, you need consulting or a development partner.

Here's a decision framework that cuts through the noise:

Three questions to ask any AI consultant or advisor before hiring them:

1. What did you build or deploy in the last 30 days? This is the only question that matters. AI changes so fast that experience from six months ago is historical. You want someone whose knowledge is current because they're using these tools right now, not because they read the changelog. I work with Claude Code and agentic AI systems every day. That's not a selling point. It's table stakes for anyone claiming to advise on AI in 2026.

2. Who will actually do the work? If you're buying a consulting engagement, the person in the pitch meeting should be the person doing the work. If the answer is "our team of analysts will handle the research phase," you're paying for a brand name and getting a graduate's best guess.

3. How does your advice stay current? If the answer involves "we update our frameworks quarterly," that's too slow. AI doesn't wait for quarterly updates. The advisor or consultant should be building, testing, and deploying AI tools as part of their own work. Not just reading about them.