AI in Daily Consulting: Where It Helps — and Where It Destroys Trust

In almost every conversation with banks, insurers, and retailers right now, I am asked how much advisory work can be automated in the future. The honest answer is: a lot — but not where most people expect. And defining that exact dividing line is precisely the capability advisory organizations need to build right now.

My rule of thumb is simple: AI handles everything around the conversation. Humans handle the conversation itself. Everything that involves preparation, follow-up, research, and structure should be automated. Everything that involves trust, judgment, and commitment should not be automated — not out of nostalgia, but because it simply doesn't work there.

Where AI Makes a Real Difference in Everyday Advisory Work

Meeting preparation in five minutes instead of thirty. Before every important appointment, an advisor needs the full picture: existing contracts, life events since the last contact, open action items, potential conversation starters. That is aggregation work — and today it can be done in a fraction of the time. The crucial point: the time saved shouldn't be spent on more meetings, but on better ones. Anyone who automates preparation and simply converts that time into higher volume has misunderstood the leverage.

Meeting notes and documentation. The biggest silent time-sink in advisory services is post-meeting documentation — and the biggest quality killer is when it gets skipped due to time pressure. Automated summaries that extract client needs, objections, and agreed next steps solve both problems. In regulated industries, there's a second benefit: advisory logs are mandatory anyway, and something that has to be created regardless shouldn't be entered manually late at night.

Follow-up drafts. AI writes the first draft, the advisor writes the final one. This saves five to ten minutes per email and lowers the barrier to reaching out within 48 hours — which is worth more in advisory work than any literary eloquence.

Objection preparation. Before difficult appointments, I have AI generate the strongest counterarguments against my own recommendation. This is one of the most underrated use cases: a sparring partner that isn't polite and has no regard for internal sensitivities.

Pattern analysis in your own data. Which meeting types lead to closed deals, when do clients book, which life events trigger which needs, where do processes drop off? These are questions nobody used to have time for. These evaluations yield the numbers used to manage capacity — such as the insight that 38 to 42 percent of bookings occur outside business hours.

Where AI Has No Place

In the advisory conversation itself. No system whispering live answers to the advisor. A conversation where one participant is reading off a script is not an advisory conversation. The client notices — not from the content, but from the rhythm.

In personal initial outreach. Mass outreach that merely appears personalized is the fastest way to devalue a brand. Especially with existing clients who have had the same point of contact for years, the difference between genuine and generated is immediately palpable — and the damage to trust outlasts any open rate.

With recommendations, terms, and commitments. Everything that creates binding commitments belongs to a human being who stands behind them. In financial and insurance advisory, this isn't a matter of style, but a matter of liability and compliance.

In reading people. Whether a client hesitates because they don't understand the product, because they still need to convince someone at home, or because they distrust the institution — that is judgment work that requires presence. AI can provide the hypothesis. Only a conversation can confirm it.

The Three Rules I Give to Teams

First: AI drafts, humans decide. No output goes to a client unverified. Anyone who softens this rule saves minutes and risks mandates.

Second: Automate the preparation, not the relationship. The test is simple: would it bother the client if they knew this task was automated? For a meeting reminder: no. For a personal assessment following a retirement planning meeting: very much so.

Third: Invest the saved time in greater humanity, not higher volume. This is the core strategic decision. If automation frees up two to four hours of administration per advisor per week — which we regularly see in implementations — you can convert that into more meetings or better ones. Those who bet on volume compete with everyone else doing the same. Those who bet on quality differentiate themselves in the very area AI cannot occupy.

Why This Is a Leadership Decision

The knee-jerk reaction of many organizations is to introduce AI as a cost-cutting program: same output, fewer people. In advice-intensive businesses, I believe this is strategically wrong. The more routine tasks get automated, the scarcer non-automatable work becomes — and scarce goods get more expensive, not cheaper. In the coming years, personal advisory will shift from the default baseline to a premium asset. Anyone who cuts it now is cutting away their own differentiation — leaving them to compete against providers where price becomes the only point of comparison.

Takeaway

AI belongs anywhere around the conversation: preparation, documentation, drafts, data analysis. Not in the conversation itself, not in recommendations, not in reading people. And the time saved belongs in better meetings, not more meetings — otherwise, you automate away your own differentiation.