What AI actually changes for small advisory businesses (and what it doesn't)
There are two conversations happening about AI in small advisory businesses right now, and they rarely intersect. One is the hype conversation: AI will replace your advisors, your consultants, your bookkeepers, your account managers. The other is the conversation I actually have with clients doing this work day to day, and it is a lot less dramatic and a lot more useful.
The honest answer, from actual implementation work with small advisory firms in Norway, is that AI changes less than the hype claims and more than the sceptics assume. Both extremes are wrong in a way that costs businesses money if they act on either one.
What AI does not change
It does not remove the need for judgement. Advisory work, whether it is accounting, legal, HR, or operational consulting, is valuable specifically because a client is paying for someone to apply experience to an ambiguous situation and reach a defensible conclusion. AI does not do that. It does not carry professional liability, it does not read a client's tone in a meeting and adjust the advice accordingly, and it has no accountability when the advice is wrong.
The chatbot-as-magic narrative sells the idea that you can drop a language model in front of clients and replace the advisor entirely. In practice, every small advisory business I have worked with that tried this learned the same lesson: clients are not buying answers, they are buying confidence that someone competent has looked at their specific situation. A generic chatbot response, however fluent, does not provide that. It often actively damages trust, because it reads as exactly what it is - unaccountable and generic.
IPRESTANDA's AI advice doesn't apply to you makes this point directly: most AI-generated advice content is built for an average case that describes no one's actual business. Small advisory firms selling their expertise as customised cannot afford to then deliver an AI chatbot that behaves like an average case generator.
What AI actually changes
Where AI genuinely changes small advisory businesses is upstream and downstream of the judgement itself, not inside it.
Upstream: research, document review, drafting the first pass of an analysis, pulling relevant precedent or prior client history, summarising a long intake conversation into a structured brief. This is the work that used to eat hours of an advisor's week before the actual advisory judgement ever got applied. AI compresses this dramatically. An advisor who used to spend three hours preparing before a client call can now spend thirty minutes, with better preparation than before, because the AI-assisted first pass surfaces things a rushed manual review would miss.
Downstream: follow-up documentation, formatting deliverables, tracking action items, drafting the summary email that used to get written at 9pm after a full day of client meetings. None of this is where the client's value comes from. All of it eats capacity that could otherwise go into more clients or better thinking on the harder cases.
The cost of AI has changed. What used to require a development budget most small advisory firms could not justify is now accessible at a price point that removes the cost objection entirely. What remains is the same question it always was: which specific workflow, not AI as a category, are you actually trying to improve.
Where pilots actually fail
The pattern across IPRESTANDA's three reasons AI pilots fail in small businesses shows up constantly in advisory firms specifically: the pilot targets customer communication or advisory quality in the abstract instead of one narrow, well-defined workflow with a measurable before-and-after. A vague target produces a vague pilot, and a vague pilot produces the kind of underwhelming result that gets used as evidence that "AI does not work for us", when the real issue was scope.
How to choose the first workflow to automate is the practical fix: score your recurring workflows on frequency, friction, and output clarity, and start with the one that scores highest on all three. For most advisory businesses, that is intake summarisation or first-draft document review, not client-facing advice generation. Start there, prove the return, and build outward from a working example rather than a theoretical strategy.
The actual leverage
Real leverage from AI in a small advisory business looks unglamorous. It is a fifteen-minute reduction in prep time per client meeting, multiplied across every meeting, every week, every advisor. It is a first-draft document review that used to take an hour taking ten minutes, freeing that hour for the judgement work that actually justifies the fee. It compounds - not because any single use looks impressive, but because it happens dozens of times a week, every week, for the life of the business.
None of this replaces the advisor. All of it changes what the advisor spends their limited hours doing. That distinction is the entire argument, and it is the one that gets lost whenever the conversation drifts toward chatbots replacing people instead of workflows freeing them.
Murphy Alex builds operational AI systems for Norwegian SMEs from Frøya, Trøndelag. IPRESTANDA is at iprestanda.com.