How to prepare for an AI consultation

Why preparation matters

A paid AI consultation is a working call, not a sales pitch. Thirty or sixty minutes goes quickly, and the quality of what you leave with depends heavily on what you bring in. An owner who arrives with one real workflow and a few rough numbers will get far more value than one who arrives with a vague worry about being left behind by AI.

This guide is written for small business owners preparing for a call like the AI Clarity Call, but the advice applies to any practical AI consultation.

The aim is not to arrive with perfect documentation. The aim is to bring enough real evidence that the call can move from "AI might help somewhere" to "this is the first useful thing to do next".

What to bring

You do not need a strategy deck. You need the everyday detail of how your business actually runs.

• The one task or workflow that eats the most time each week, described step by step.
• Rough numbers: how often the task happens, how long it takes, and who does it.
• A list of the tools you already pay for, such as your booking system, accounts software, inbox, and spreadsheets.
• Your website address, and access to Google Search Console if search visibility is on the agenda.
• Anything you have already tried with AI, even if it went nowhere. Failed experiments are useful evidence.

If you are unsure which workflow to choose, pick the one that causes the most friction when you are busy. Good candidates are repeated enquiries, inbox triage, booking admin, quoting, reporting, stock checks, content drafting, handoffs between team members, or any task where the same information gets copied from one system into another.

Bring examples rather than descriptions where possible. A real email thread, a sanitised spreadsheet row, a screenshot of the current form, or the exact steps you take in a tool will usually beat a polished summary. If anything is sensitive, redact it before the call. The consultant needs the shape of the workflow, not private customer data.

What not to bring

You do not need to bring a finished AI tool idea. In fact, arriving too attached to a specific tool can make the call less useful. The job of a good consultation is to test whether the idea fits the business problem, not to praise the first idea back to you.

You also do not need a technical specification. If a build comes later, the specification should be produced after the workflow is understood. A better starting point is a short note like this:

• We receive 20 to 30 quote requests per week.
• Each one takes roughly 10 minutes to sort.
• The hard part is deciding which ones need a fast reply.
• The current tools are Gmail, a spreadsheet, and our website form.
• We want fewer missed opportunities without losing human judgement.

That is enough to have a useful conversation.

What a good 30 minute call covers

A 30 minute call suits one focused question. That might be a quick scan of where AI could genuinely help in your business, a first pass on website and search visibility, or a sense check on a tool you are about to buy. Expect the person on the other end to narrow the conversation quickly. That is a good sign, not a brush-off.

A strong 30 minute call normally ends with a short decision. That might be "do this first", "do not buy that tool yet", "your website needs clearer pages before AI search will help", or "this is automation work, not an AI agent project". The value is clarity. It should prevent waste.

What a good 60 minute call covers

A 60 minute call has room to map a problem properly. That usually means walking through the workflow, separating what should stay human from what can be automated, comparing options against the tools you already have, and putting the candidate fixes in a sensible order. Repeated admin, handoffs, reports, and inbox triage are common ground here, and the automations capability page describes what that work looks like when it turns into a build.

A strong 60 minute call should also rank the work. Not every useful idea should be built first. The best first slice is normally small enough to test, close enough to revenue or saved time to matter, and boring enough to survive real use. That might be a better intake form, a reporting helper, an inbox triage loop, a clearer consultation page, or a small internal dashboard.

Website and search preparation

If the call includes website or search visibility, bring access to the evidence that search systems use. Google Search Console is the most useful starting point. You do not need to understand every chart, but the consultant can learn a lot from crawl status, indexed pages, queries, impressions, and which pages Google has chosen as canonical.

Useful website prep includes:

• The main pages you want customers to find.
• The search terms people already use when they contact you.
• The pages that explain your strongest offer, product, or service.
• Any old URLs, redirects, or pages that still get visits.
• The last time key pages were rewritten.

For AI-search visibility, the same rule applies: bring real pages and real questions. AI systems need clear answers, not only technical files. A page that plainly explains what you do, who it is for, what happens next, and what the limits are is a better asset than a page packed with vague AI keywords.

Automation and agent preparation

If the call is about automation or agents, prepare the workflow rather than the prompt. Write the task as a sequence of steps. Note who does each step now, where information comes from, and what could go wrong if the system acted without approval.

The most useful detail is often the exception. What happens when a customer does not provide enough information? What happens when a booking clashes? What happens when a spreadsheet row is missing a value? What happens when the AI is not confident? These edge cases are where good systems are designed.

What you should leave with

Before the call ends, make sure you can answer three questions. What is the first thing to do next? What did we decide not to do, and why? What would a sensible second step look like if the first one works?

In practice that means a short, ordered action list, a clear note of anything parked as not worth it yet, and enough plain-English reasoning that you could explain the plan to a colleague afterwards. If a build makes sense, scoping it is a separate conversation, so there is no pressure to commit on the call.

Booking details, durations, topics, and prices for the studio's own calls are on the consultations page.

The best result is not a giant roadmap. It is a small, ordered next move that you trust enough to act on.

Related capability

Automations

Workflow automations that save time across handoffs, repeated admin, and reporting.