What to review before a client advisory call

Most preparation advice for advisory calls tells you to pull the P&L, the balance sheet and the cash position, and to walk in with a view. That is the reporting half of the job. It is not the part that goes wrong.

What goes wrong is continuity. The client mentions something you agreed in July, and you cannot place it. They ask whether the thing you predicted actually happened, and you do not know yet. They ask a question you could have seen coming, and you promise to send something after the call. None of this is carelessness. It is what happens when the record of an engagement lives in four places that were never designed to talk to each other.

This guide is the sequence a good controller would run before a client advisory call if the client could afford to staff one. Every step can be done by hand. The last section is honest about what that costs.

The eight-step sequence

For a CPA or fractional CFO advisory engagement, preparation runs from prior decisions to verified outcomes, then to current reports and the questions that follow.

1. Block preparation time the day before.
2. Reconstruct decisions from the prior conversations.
3. Confirm when the books were last reconciled.
4. Verify each decision against the books and the bank.
5. Read current results in that context.
6. Anticipate the client’s likely questions.
7. Run the relevant scenarios before the meeting.
8. Open the call with what changed.

The order is most of the value. The rest of this guide explains why, and how to do each step.

This sequence was developed from AgentLink’s design-partner conversations with CPA and fractional CFO practices. It is an operating framework, not a professional standard.

Why this is a structural problem, not a diligence problem

With a handful of clients, you can hold most of the engagement in your head. Past that, memory stops being dependable.

At fifteen or thirty, reconstructing that context reliably becomes a capacity problem, regardless of how careful the practitioner is. Every call is a context switch. The notes exist: on Teams, in Fireflies, in Granola, in whichever recorder the firm settled on. The task list exists, in Notion or a spreadsheet. The books exist in the accounting system. Each of those tools does its job faithfully. None of them was built to hold the engagement across all three, and none of them was built to answer the one question that matters before a call: what did we say would happen, and did it?

A task list closes when you send the email, not when the money arrives. That is what task lists are for. Asking one to also track whether a prediction held is asking it to be something it is not. The gap is in the structure, not in the person using it.

So the sequence below is arranged around that gap. It is not a checklist of reports. It is an order of operations.

The sequence

1. Block the time the day before

Not the morning of. The morning of is when you discover the books are two weeks stale and there is nothing you can do about it.

Size it honestly. In our design-partner conversations, practitioners described preparation ranging from roughly two hours for a straightforward client to half a day for one with debt, multiple entities or a decision in flight. If you run a firm, this is a standing task for the associate on the engagement, not something the partner does in the car. The partner reviews what the associate prepared.

If that number looks large, it is. That is the point of the closing section.

2. Reconstruct the decisions, not the topics

Open the notes from the last two calls. Do this before you open a single number, because you want to read the numbers in the context of what was agreed, not the other way round.

Write down every decision that was made. Not every topic discussed. A decision has a subject, an expected outcome and a date: hire a second technician by September, hold the price increase until Q4, chase the overdue receivable this week. Most call notes are a transcript of topics. You are extracting the small number of commitments buried in them.

In a recent conversation with design partners, a fractional CPA described a client calling him two weeks after a meeting, talking for twenty minutes, and none of it landing anywhere: not in the meeting notes, not in the task system. That conversation contained a decision. This step is where you go looking for those.

3. Check when the books were last reconciled

Before you read the P&L, find the reconciliation date.

This came up more consistently than anything else in our design-partner conversations. Books go stale. If the bank reconciliation is three weeks old, every number downstream is provisional, and every decision made on top of it can be questioned later. You need to know that before the call, and the client needs to hear it from you, not discover it when a number moves.

If the books are stale, the call changes shape. Some of the numbers you were going to present become “as of the fifteenth.” Some of the questions you were going to answer become “I will confirm once reconciled.” Better to know that the day before.

4. Verify each decision against the books and the bank

Now take the list from step two and check each item, one at a time, against what actually happened.

Did the receivable land? Look at the bank, not the invoice status. Did the hire go through? Payroll will tell you. Did the price change ship? Revenue per unit will tell you, if the books are current enough to trust.

Three outcomes for each: it happened, it did not, or you cannot tell yet. All three are useful. “Cannot tell yet” may mean stale books, a missing categorization, or an outcome that has not become observable, and the client may not know which either.

A task list stops at completion of the activity: it records that you sent the email. This step records whether the outcome you expected followed.

5. Read the current numbers in that context

Only now open the reports. P&L, balance sheet, cash, the variance report if you produce one.

The difference from reading them first is that you now know what you are looking for. If the receivable slipped, the cash line is not a surprise, it is the consequence of something you already identified. If the hire went through, the payroll increase has a reason attached. You are reading the numbers as a story you already know the first half of, rather than as a list of figures to explain.

6. Anticipate the questions

You have just checked what did and did not happen. That tells you what they are going to ask.

If the receivable slipped, they will ask whether payroll is safe. If the hire went through, they will ask what it did to burn. If a supplier invoice doubled, they will ask why. Write down the two or three questions that follow from what you found, and have the answer ready.

This is the difference between reporting and controllership. Reporting answers the question when it is asked. Controllership arrives with it already answered.

7. Have the model already run

If you know the client is weighing a decision, do not just anticipate the question. Run the scenario before the call.

They mentioned bringing on two contractors. Arrive with the cash picture for that. They are considering a second location. Arrive with the fixed-cost step and the months of runway it consumes. You are not trying to make the decision for them. You are making sure the meeting produces a decision rather than a follow-up.

Practitioners described this as the most time-intensive step in the sequence, which is why it is the first to go when capacity is tight. A meeting that ends with “I will model that and send it over” has pushed the decision by two weeks, and two weeks later the numbers have moved.

8. Open the call with what changed

The visible output of all of this is the first two minutes.

Do not open with “how is the business going?” You already know how the business is going. You have the books, the bank and the last two calls in front of you. Open with what changed: “Last time we agreed to chase the Henderson invoice. I can see it landed on the twentieth, so that takes the pressure off the September payroll. The thing I want to spend time on today is the contractor question, because I ran the numbers.”

That opening does two things. It signals that the engagement has continuity, which is what the client is paying for. And it sets the agenda around the decision, not around the reports.

What this costs

Run properly, and using the range practitioners described to us, this sequence is two hours to half a day per client per cycle. At five clients that is a heavy day a month. At thirty, it is a job.

When capacity is tight, steps two, four and seven are the first to be compressed, because each requires reconstructing context across systems that were never connected. Not because practitioners do not see the value, but because the time is not there, and the tools they have do not carry any of it. The recorder recorded. The task list held the task. The books held the numbers. The part that connects them, the part that says “we predicted this and here is whether it happened,” has to be done by hand every time, and it is the first thing to go when the week gets busy.

The sequence still works if you do it by hand. It just does not scale past the number of clients you can hold in your head.

Our product

AgentLink (agentlink.finance) is designed to perform steps two, three, four, six and seven continuously, for every client in a firm’s portfolio, and hands the partner a brief before the call with the decisions listed, each one checked against the bank, the reconciliation date stated, and the likely questions surfaced. It is built on the premise that this is the controller’s job, and that most small business clients cannot afford a controller. It is sold through accounting and fractional CFO firms, who keep the client relationship and make the judgement calls.

Frequently asked questions

How far back should I go in the notes?

Start with the previous two calls. Go further back when an unresolved commitment, forecast or decision is still active.

What if the client does not use a recorder?

Then the reconstruction step is harder and more important. Capture the decisions at the end of every call, however you do it, with the subject, the expected outcome and the date. A short decision record at the end of the call removes most of the reconstruction the day before the next one.

What financial reports should I review before a client advisory meeting?

The standard set: P&L, balance sheet, cash position, and a variance report if you produce one. The point of this guide is that the reports are step five, not step one. Read the prior decisions and the reconciliation date first, then the reports.

How long should a fractional CFO spend preparing for a client call?

The practitioners we work with described two hours for a straightforward client and up to half a day for one with debt, multiple entities or a live decision. The number depends on how much of steps two, four and seven has to be reconstructed by hand.

Should the client see the preparation?

Not the working. The output, yes. Opening the call with what changed is the client seeing it.

Does this apply to a monthly cadence only?

It applies to whatever cadence the engagement runs on. The gap between calls is where things happen that never enter the engagement record, and the longer the gap, the more there is to reconstruct.

About the author

Nilanjan Raychaudhuri is the founder of AgentLink (Tublian LLC, Columbus, Ohio), which builds controller-layer software sold through CPA and fractional CFO firms. He has spent the past year interviewing practitioners at fractional CPA and CFO practices about how advisory work is actually delivered. Team page

The practitioner observations in this guide come from design partner conversations and are used with permission where attributed. Corrections to nilanjan@agentlink.finance.