Preparing for a client meeting with AI

How the bot works

Nine steps. Reading the calendar, gathering the account's own records, assembling the sheet and citing every line are ordinary code and cost almost nothing to run; the model is only asked to summarize records it was handed and to phrase a question about an open item.

1. It starts from the calendar and the account. The meeting, the company, the people on the invitation. From there it reads only the records you pointed it at for that account — nothing else, and nothing from outside your systems.

2. One sheet per meeting, in the order you read. Where the relationship stands, what happened last time, what is open, what is owed, what is due. A fixed shape means somebody can read it on the way in rather than study it.

3. Every line names the record behind it. This came from that note, that invoice, that ticket, that agreement, as of that date. A line with no record behind it does not get written, which is the whole difference between a prep sheet and a pile of research.

4. The open items are the agenda. Unpaid invoices, open tickets, commitments with dates, renewal and notice windows, deliveries promised. Most of what goes on the sheet is something somebody is already waiting on.

5. What nobody wrote down is listed as unrecorded. A conversation nobody logged, a promise with no date, an account with no notes since the last renewal. The sheet says so plainly instead of reasoning its way to something plausible.

6. The questions come from the open items. Each question points at the record that raised it: the invoice past terms, the ticket nobody closed, the commitment with no owner. They are questions for you to ask, not advice about what to say.

7. It does not look the person up. No profile built from the internet, no posts read, no career history, no shared interests, no news about their company. The people on the invitation are names and roles from your own records, and that is the design rather than a setting.

8. The sheet waits for you, and writes to nothing. It lands where you asked before the meeting, and it stops there. Nothing is written back to the account, nothing is sent to the client, and no meeting, call or task is created.

9. It tells you which accounts have nothing recorded. The meetings where the sheet came back almost empty, the accounts whose last note is older than the last renewal, the commitments nobody wrote down. That list is the useful output in the first month, because an empty sheet is the only visible sign that nothing was written down.

What you configure

It can be configured to run in any business. This is what that actually means — the list of things that are yours, not ours.

Your meetings

Which kinds get a sheet — a client review, a renewal conversation, a first meeting, a site visit — and which are left alone.

Your sources

Which records it may read for an account, how they are read, and which source wins when two of them disagree.

Your sheet

The sections, their order, the length, and what gets dropped first when there is too much to fit on a page.

Your open-item rules

What counts as open, how far back it looks, which thresholds make an item worth naming, and which items always appear.

Your question rules

Which open items raise a question, how it is phrased, and the subjects that are never put in a question at all.

Your privacy rules

What the bot may never read — and the standing rule that it does not research a person, anywhere, in any configuration.

Your people

Who gets a sheet for which accounts, who owns an unrecorded item, and who sees the commercial sections.

Your timing

How far ahead of the meeting the sheet is built, where it is delivered, and what happens when a meeting is moved.

Your release switches

Which meeting kinds are prepared automatically from your calendar, which wait for somebody to ask, and whether a sheet is delivered anywhere besides the person attending. Writing back to your systems is off, and it is not a switch.

Transparent, and configurable by you

The dark band below is what the system will not do. This is what you can see and change — and it is how we build everything, not something invented for this page.

You see what it did, and why

Every line on the sheet names the record and the date behind it, every unrecorded item says what was missing, and each run records the sources it read with their timestamps, the items it raised, the questions it drafted and who it went to. A sheet can be checked against your own records rather than taken on trust.

Versioned prompts you can tune yourself

The source list, the sheet layout, the open-item rules, the question rules and the prompts behind summarizing a record are yours to tune. A change is a reviewable diff, replayed against meetings that have already happened so you can see what it would have put in front of you differently, and rolled back the same way.

A human in the loop wherever judgment actually matters

Reading the calendar, gathering the records, assembling the sheet and citing every line are automatic. Deciding what the meeting is for, what to offer, what to concede and what to say are a person's — that boundary is a property of the system, not a promise about how carefully we'll use it.

Every run logged, along with what it cost

One row per sheet and one per unrecorded item: every source with its timestamp, every line with its citation, the open items it raised, the questions it drafted, who received it and when — plus token usage and cost for that run.

It can be configured to run in any business. The records and the rules are yours: your sources, your sections, your open-item rules, your privacy rules, your people. The meetings on this page are one configuration of it, not the limit of it.

Works with your tools

This reads the systems you already run and hands you a page. We're not selling you a sales platform, and your records stay the record.

• HubSpot
• Salesforce
• A PSA or professional services tool
• A helpdesk or ticketing tool
• QuickBooks Online
• A calendar or booking tool
• Google Drive, SharePoint or Dropbox
• A spreadsheet of accounts

Works well for

• Managed service providers — A client review per quarter, where the open tickets and the contract terms are the conversation.
• Software companies — Account teams walk into renewal conversations needing the history, the tickets and the commitments in one place.
• Financial advisory firms — Review meetings run on a schedule, and the file somebody has to read first is spread across systems.
• Insurance agencies — Annual reviews per account, where the policy terms and the open claims decide the agenda.
• Accounting firms — Client conversations sit on open items, deadlines and what was promised at the last one.

How a rollout works

Nothing is sent on day one and nothing is written back, ever. The first milestone is meetings that have already happened, prepared from your own records as they stood that morning, so you can read what you would have walked in with.

1. 01Discovery

   The quote call, then a working session. We go through the meetings you prepare for, which records exist for an account, what counts as an open item, which sections you want on the page, what the bot may never read, and who gets the sheet. You get the source map, the sheet layout and a written scope.

2. 02Dry run

   Meetings that have already happened, prepared from the records as they stood at the time. Nothing is delivered. You read each sheet against what actually came up and read the unrecorded list — and the unrecorded list is where the work is, because an account with no notes looks exactly like one that is going fine.

3. 03Assisted mode

   Live on one meeting kind. The sheet is built ahead of the meeting and lands with the person attending. Nothing is written back to your systems and nothing goes to the client; who runs the meeting does not change.

4. 04Optional limited automation

   Optional, and narrower than it sounds. A short list of meeting kinds you name, prepared from your calendar, delivered only to the person attending, with every hard stop in force and the kill switch. Plenty of teams stop at step 03, and that's a good place to stop.

What we don't claim

• No numbers we haven't earned. No win-rate claim, no retention claim, no hours-saved figure, no meetings-per-week number. We have not prepared your meetings. What a run takes and what it costs come out of the dry run, and they will be your numbers.
• No customer claims on this page. No names, no case studies, no logos.
• Research is not preparation, and this does the second one. A pile of material about a company is easy to produce and hard to use. What this hands you is what your own records say is open, owed and promised.
• A prep sheet does not make the meeting go well. It can make sure nobody is surprised by an invoice or a ticket that was already in your system. What you offer and what you agree to are still yours, and so is the conversation.
• If nothing is recorded, the sheet will be thin. The design reads your records rather than guessing around them, so an account nobody has written notes on produces a short page and an honest list of what is missing. That list is worth having, and it is not a substitute for writing things down.

Questions people ask

Can AI prepare us for a client meeting?

Yes, from your own records. An agent can read the account history, the notes from last time, open invoices, open tickets, commitments with dates and the contract terms, and hand you a one-page sheet with a citation on every line and a list of what nobody wrote down. What it will not do is research the person you are meeting or tell you what to say.

Where does it get the information?

Only from sources you name in the configuration - the CRM or account history, the helpdesk, the accounting system, the PSA, the agreement in your document store. Nothing comes from outside your systems, and each line names the record behind it.

Does it look the person up on LinkedIn?

No. It builds no profile of a person from the internet, reads no posts, and collects no career history, interests or connections. The people on the invitation appear as names and roles from your own records, and there is no configuration that changes that.

Is this the same as meeting notes afterward?

No, and they are deliberately separate. This runs before the meeting and creates no task. Turning what was said in a meeting into commitments and follow-ups is a different bot with its own page, and so is scoring a call that has already happened.

Does it build the quarterly deck?

No. A recurring pack built from reports for an audience is a different bot with its own page. This produces one page for the person about to have one conversation, and it presents nothing.

Will it write to our CRM?

No. It writes to nothing: no note, no field, no activity, no ticket. The sheet is a document that lands in front of a person and changes nothing behind it.

What if the account has no notes?

Then the sheet is short and says so. Unrecorded items are listed as unrecorded rather than filled in with something plausible, and the accounts that keep coming back empty are reported as their own list.

Will it tell us what to charge or what to offer?

No. It makes no recommendation about price, discount, renewal or tactics. It assembles what is known and names what is open, and the questions it drafts point at records rather than at a strategy.

Is there a record of how a sheet was built?

Yes, per sheet: every source it read with its timestamp, every line with its citation, the open items it raised, the questions it drafted, who received it and when - plus the token usage and cost for that run.