Skip to main content

How to Prepare for an AI Consultation and Make the Most of the Hour

Most first calls with an AI consultant follow the same pattern, which is why knowing how to prepare for an AI consultation matters. The consultant spends forty minutes asking what software you use, how many staff you have and what a normal week looks like. You get ten rushed minutes at the end on what […]

A laptop on a desk showing a grid of app icons as you prepare for an AI consultation, next to a coffee mug, a smartphone, and papers.

Most first calls with an AI consultant follow the same pattern, which is why knowing how to prepare for an AI consultation matters. The consultant spends forty minutes asking what software you use, how many staff you have and what a normal week looks like. You get ten rushed minutes at the end on what you actually wanted to talk about. You leave thinking AI could probably help, with no idea where to start.

That hour wasn’t wasted because AI is overhyped. It was wasted because nobody did the homework first. Doing that homework is what separates a polite chat from walking out with a short list of jobs a system could take off your plate next month.

This guide covers what I ask owners to bring, the numbers worth digging up, a 30-minute version for when you’re flat out, and the questions that tell a real AI partner apart from someone reselling tools.

Key Takeaways

  • Knowing how to prepare for an AI consultation saves the first 40 minutes a consultant would otherwise spend asking basic questions about your business.
  • Log one normal week of your time before the call. Owners almost always underestimate the hours that go on admin, chasing and answering team questions.
  • Bring a written list of every recurring task, including the dull ones. Boring, repetitive jobs are where AI pays for itself fastest, not the clever stuff.
  • List every system you use and what it holds. Most founder-led businesses run on five to eight tools that don’t talk to each other properly.
  • A few rough numbers, like monthly lead volume, response time and missed calls, turn a vague AI chat into a conversation about real costs.
  • Pick one outcome you want from the hour. “Get two afternoons a week back” gives a consultant something to aim at. “Look into AI” does not.
  • Ask how they handle your data, what they wouldn’t automate, and who supports the system after it’s built. Vague answers are a red flag.

Why Most First AI Consultations Waste the Hour

I’ve been on both sides of these calls. I’ve sat across from consultants who wanted to sell me a platform, and I’ve run hundreds of discovery conversations with founders myself. The calls that go nowhere almost always fail in one of three ways, and the owner can prevent all three.

Arriving With a Tool Instead of a Problem

“We should probably be using ChatGPT more” isn’t a brief. Nor is “my mate set up an AI chatbot, and I want one too.” When the conversation starts with a tool, it stays on tools: features, subscriptions, which model is smartest this month. You come away knowing more about software and nothing more about your own business.

The useful starting point is the problem. Which job keeps landing back on your desk? Which part of the week do you dread? Where do leads, customers or money slip through the cracks? A consultant can work with any of those. They can’t do much with a product name.

Arriving With No Picture of Your Week

Most owners can’t tell you where their time goes. That’s not a criticism. When you’re running everything out of your own head, the week turns into a blur of messages, quotes, callbacks and fires. Ask for a breakdown, and you’ll get a guess, and the guess is usually wrong by several hours.

Without that picture, the consultant has to build it live on the call, one question at a time. That’s the forty minutes gone. If you arrive with it already sketched out, the conversation starts at “here’s where the hours go, what can we do about it” rather than “so, tell me about your business.”

Letting the Consultant Run the Whole Show

Some AI consultants are really salespeople with a deck. If you turn up with no agenda, theirs fills the space, and the hour becomes a demo of whatever they happen to sell. That’s how businesses end up paying for a chatbot when the real leak was unanswered phone calls.

This is a big part of why so many AI projects stall. Research out of MIT found that most enterprise AI initiatives deliver no measurable return, and the pattern behind the failures is consistent: they start with technology rather than process. Small businesses fall into the same trap, only with less budget to absorb it. If you’ve been burned already, my piece on AI automation after a bad vendor goes through what usually went wrong.

The fix is simple, and it’s not technical. Do a few hours of preparation so the call is about your business, on your terms. The rest of this guide is that preparation. If you’re still deciding whether AI fits a business your size at all, start with AI adoption for SMEs and come back.

A receptionist's desk in a busy clinic at the end of the day, a desk phone with a blinking light, sticky notes along the edge of the monitor, an empty waiting room behind, the kind of detail that shows how to prepare for an AI consultation.

How to Prepare for an AI Consultation: Map Where Your Week Goes

This is the most valuable piece of preparation, and the one owners skip most often. Anyone can list their software. Very few have an honest picture of their own week. Get this right, and everything else on the call gets sharper.

Run a One-Week Time Log

Pick a normal week, not one with a holiday or a trade show in it. Keep a running note on your phone or a notepad on your desk. Every time you switch tasks, jot down what you’re moving on to and roughly how long the last thing took. Rough is fine. Fifteen-minute blocks are plenty.

Don’t tidy it up or leave out the embarrassing bits. If you spent forty minutes hunting for a quote you sent in March, write it down. If you answered the same question from three different staff members, write that down too. Those are exactly the moments a consultant is looking for.

By Friday most owners are surprised. The usual finding is that somewhere between a third and half of the week went on work that didn’t need them specifically. It just needed someone who knew the context, and right now the only person who knows the context is you.

Write the Recurring Task List

From the log, pull out everything that repeats. Daily, weekly and monthly. Chasing invoices, sending quotes, following up enquiries, booking appointments, writing the weekly team update, pulling numbers for the accountant, checking the job board, answering “where’s my order” emails.

Include the boring stuff. Owners tend to leave out dull tasks because they assume nobody would care, and dull is exactly what automates well. A good consultant will score each task on how repetitive it is, how much judgment it needs and how much time it eats. You can’t score what isn’t on the list.

Aim for at least thirty items. Most founder-led businesses with a few staff have fifty to a hundred recurring tasks once you count properly.

Write the “Only I Can Answer This” List

This is a separate list, and it matters just as much. Write down every question your team brings to you that they can’t answer themselves. Pricing exceptions. Which clients get priority. How a particular job should be quoted. What to say to an unhappy customer.

Each item on this list is a piece of knowledge that lives only in your head. That’s the real bottleneck in most businesses, and it’s the part a well-built AI brain is designed to capture, so your team can get the answer without waiting for you.

Run the Disappear Test

Last exercise. Imagine you were completely unreachable for two weeks. No phone, no email. Write down what would break first, what would stall, and what would quietly pile up for your return.

That list tells a consultant where the risk sits. It also tends to be the most honest document an owner has ever written about their own business.

Gather the Numbers and the Tool List

A clear picture of your week tells a consultant where the pain is. A few numbers tell them what it costs. Together they turn “AI could probably help” into “this one fix is worth roughly this much a month.” You don’t need accounting precision. Ballpark figures you can stand behind are enough.

The Systems List

Write down every tool the business depends on. CRM, accounting software, booking system, job management, email platform, phone system, project tool, the shared spreadsheet someone updates on Fridays. Next to each one, note what it holds and who uses it.

Most owners count five to eight systems and realise none of them shares data properly with the others. That’s normal and not something to be embarrassed about. It’s also one of the first things a consultant needs to know, because an AI system that can’t see your data is guessing.

Two more columns are worth adding. How often do you personally log in to check it, and what decision does it feed? If you check four dashboards every morning to work out how yesterday went, that’s a strong candidate for a single automated summary.

The Five Numbers Worth Finding

Pull these before the call, even if they’re rough:

  1. Monthly enquiries or leads. How many new people contact you in a typical month, across every channel.
  2. Response time. Roughly how long it takes for a new enquiry to hear back from a human. Be honest. “Same day, mostly” often means four hours on a good day and two days on a bad one.
  3. Missed calls. Most phone systems will report this. If yours doesn’t, ask whoever answers the phone for their best guess.
  4. Dormant database size. How many past customers or old leads are sitting in your CRM that nobody has contacted in six months or more.
  5. Hours on admin. From your time log, the total hours a week you spent on admin, chasing and answering questions.

These numbers tend to surprise people. A dental practice I worked with assumed it was answering most calls, with two receptionists on the desk. The real figure was that 47% of calls went unanswered. Once an AI receptionist was handling overflow and after-hours calls, missed calls dropped to zero, and Dr Claire’s booked appointments rose 44%. None of that would have come up without the missed-call number, so check yours. If after-hours calls are your weak spot, after-hours call answering covers the options.

Access and Permissions

Find out now who holds the logins for each system and whether you have admin access. A surprising number of owners find their website, domain or CRM sitting under a former staff member’s account or an old supplier’s email. You don’t need to fix that before the call, but it’s useful for the consultant to know, because it affects how fast anything can actually be built.

Data and Privacy

You don’t need to hand over customer records for a first conversation. Summary numbers do the job. When you do get to the point of sharing access, ask where your data is stored, which AI providers will process it, and who on their side can see it.

In New Zealand, the Privacy Act 2020 applies whether the work is done by a person or an AI system. The Office of the Privacy Commissioner publishes plain-English guidance on using AI tools with personal information, and it’s worth ten minutes of your time before the call. A good consultant will bring this up unprompted.

A business owner at a kitchen table on a Sunday evening, writing a long handwritten list on a legal pad as a first step in how to prepare for an AI consultation, a phone and a calendar diary open beside them.

How to Prepare for an AI Consultation in 30 Minutes

Not everyone has a spare week to log their time. Maybe the call is tomorrow, and you’ve only just remembered it. Here’s the short version. It won’t be as good as the full preparation, but it’ll still put you well ahead of most people who book an AI strategy session.

Ten Minutes: Rebuild Yesterday

Open your calendar, your sent emails and your phone’s call log for yesterday. Walk through the day in order and write down what you did and roughly how long each thing took. One day isn’t a full week, but it’s real data rather than a guess, and it’ll show the pattern.

Ten Minutes: The Top Ten Annoyances

Write down the ten recurring jobs that annoy you most. The ones you put off, the ones that keep coming back, the ones you’d happily pay someone to make disappear. Don’t overthink the order. Your gut is usually right about what’s draining you.

Then circle the three that eat the most time. Those three are what the call should be about.

Five Minutes: Three Numbers

Of the five numbers above, find the three you can get quickest. For most businesses, that’s monthly enquiries, missed calls and dormant database size. If you can’t find a number, write down your best guess and say on the call that it’s a guess. Consultants work with estimates all the time. What they can’t work with is nothing.

Five Minutes: One Outcome and One Fear

Write one sentence about what you want from the hour. Make it specific. “I want two afternoons a week back.” “I want every enquiry answered within five minutes.” “I want to take a holiday without the phone ringing.”

Then write one sentence about what worries you. “I don’t want my customers talking to a robot that sounds like a robot.” “I’ve been burned before, and I don’t trust anyone in this space.” Say it out loud early in the call. It saves the consultant tiptoeing around it, and any honest consultant will answer it directly.

Decide Who Joins the Call

One more thing, and it takes no time at all: decide whether anyone else should be there. If your office manager or ops lead handles most of the day-to-day admin, they often know more about where the hours go than you do. Ten minutes of their input can change the whole conversation.

If the problem you’re trying to solve is admin piling up on you specifically, admin overload in a small business is a good read before the call. It’ll help you put words to what’s happening.

Questions to Ask an AI Consultant (and What a Good Hour Looks Like)

Preparation gets the most out of the hour. The right questions tell you whether this is someone you want to work with. You’re interviewing them as much as they’re assessing you.

Questions to Ask

What would you not automate in my business? This is the best single question you can ask. A consultant who says “everything can be automated” is either inexperienced or selling. A good one will name the judgment calls, the relationship moments and the high-stakes decisions that should stay with a human, and explain why.

Where would you start, and why there? You want reasoning, not a product pitch. The answer should tie back to something you said about your week or your numbers.

How do you handle my data? Where it’s stored, which AI providers process it, who can see it, and what happens to it if you part ways.

Who supports the system after it’s built? Most AI systems need tuning in the first few months as they meet real customers and real edge cases. Find out whether you’re getting ongoing training, optimisation and support, or a handover and a goodbye.

Can you show me something running? A real system in a real business, not a slide. It doesn’t need to be in your industry, but it needs to be live.

For the contract stage, my piece on the questions your accountant should ask about an AI automation contract covers the commercial side.

Red Flags

Watch out for anyone who talks about tools for the whole hour and never asks about your process. Watch out for big percentage promises with no explanation of where the number comes from. Be wary of pressure to sign on the call, and of anyone who goes vague when you ask about data. If they can’t explain what they’d build in plain English, they probably can’t build it well either.

What You Should Leave With

A good first hour should end with three things. A short list of the jobs worth tackling first, ranked by how much time they’d recover versus how hard they are to build. A plain-English description of what the first piece of work would look like. A clear next step, even if that step is “not right now.”

You shouldn’t leave unsure what was recommended or why. If you do, ask before you hang up. Once something is built, a structured review a few months in keeps everyone honest. I wrote about that in the 90-day AI automation checkpoint.

Two people at a cafe table mid-conversation, one sliding a handwritten page of notes across the table to explain how to prepare for an AI consultation, coffee cups and an open notebook between them.

Conclusion

Learning how to prepare for an AI consultation isn’t about becoming technical. It’s about turning up with an honest picture of your business: where your week goes, what keeps landing on your desk, what lives only in your head, and a few numbers showing what that costs. Bring that, and the hour becomes a working session instead of an interview.

There’s a bigger payoff too. The preparation you do for the call is the same raw material a proper system gets built from. Your task list becomes the automation plan. Your “only I can answer this” list becomes the knowledge your AI brain holds, so your team stops waiting for you. Your numbers become the daily view of the business you’ve never had in one place.

That’s the real goal. Not one clever tool, but an AI brain and AI workforce that runs the repeatable work so you can stop being the bottleneck.

Ready to Use the Hour Properly?

If you’ve done the preparation and want to put it in front of someone who’ll give you a straight answer, book a 30-minute discovery call with me. Bring your task list and your numbers. We’ll look at where your week actually goes and work out whether there’s a clear first win. If there isn’t, I’ll tell you.

Frequently Asked Questions

What should I bring to an AI consultation?

Bring a rough log of one normal week, a list of your recurring tasks, a list of the software you use, and a few numbers such as monthly enquiries, how fast you respond and how many calls go unanswered. Add the one outcome you most want from the call. Nothing needs to be polished. A page of honest notes is worth more than a slide deck.

How long does an AI consultation usually take?

No. A good AI consultant translates between your business and the technology, rather than expecting you to speak their language. What you need is a clear picture of how your business runs day to day: who does what, where the delays are, and which jobs keep landing back on your desk. If someone makes you feel slow for not knowing the tools, find someone else.

Do I need to be technical to work with an AI consultant?

No. A good AI consultant translates between your business and the technology, rather than expecting you to speak their language. What you need is a clear picture of how your business runs day to day: who does what, where the delays are, and which jobs keep landing back on your desk. If someone makes you feel slow for not knowing the tools, find someone else.

What questions should I ask an AI consultant before hiring them?

Ask what they wouldn’t automate in your business, how they handle your data, who supports the system once it’s built, and what results they’ve actually delivered for similar businesses. Ask to see something running, not a slide. Pay attention to how specific the answers are. A consultant who only talks about tools, and never about your process, is selling software rather than solving problems.

Is it safe to share business data with an AI consultant?

It can be, if you share the right things in the right way. For a first call, you rarely need customer records at all, because summary numbers do the job. Before handing over access, ask where data is stored, which AI providers process it and who can see it. In New Zealand, the Privacy Act 2020 still applies, so a decent consultant will raise this before you do.

How do I know if my business is ready for AI?

You’re ready if you have a few people, a few years of trading and recurring work that eats hours every week. You don’t need clean data or a tech team. The real warning sign isn’t being too small. It’s not being able to describe how your business runs, because until you can, nobody can build a system around it.

What happens after an AI consultation?

After a good consultation, you should have a short list of the jobs worth automating first, a rough sense of effort and payoff for each, and a clear next step. That might be a deeper strategy session, a first bounded build, or simply a decision that now isn’t the time. You should never leave unsure what was recommended or why.

About Octavius

Titus Mulquiney is the founder of Octavius AI, where he builds AI brains and AI workforces for founder-led businesses stuck running everything out of their own head. Twenty years in marketing, ex-Sony product manager, ex-GM Zeal NZ. Based in Auckland, working with operators across NZ, Australia, and the US. Connect on LinkedIn.

The Business Stress Test

Stuck running your business out of your head? Three minutes on your phone tells you what the chasing costs you a year, where it leaks, and what to hand off first.

Brass Butler
Brass tacks, deftly handled.

Strategic automation for founders who want their time back. Starts with a free 30-minute call to find out if there's a fit.

Book a discovery call

Not ready for a call? Take the three-minute Business Stress Test.