Business analytics AI is useful to a growing company for one reason: it shortens the time between something going wrong in the pipeline and someone senior noticing. This article answers a narrow question. How do you set up a weekly business review on the levelpartner.ai platform where the numbers come from live records and AI writes the first draft of the commentary? If you run a 10 to 250 person firm whose turnover has stalled, that review is where growth problems usually become visible first.
The weekly review most firms do not have
In a typical firm with flat revenue, the leadership meeting looks at last month's invoices and a sales forecast built from optimism. By the time a drop in enquiries shows up in revenue, it happened two or three months earlier. A weekly review built on leading indicators catches it while you can still act: new enquiries, speed of first reply, meetings booked, show rate, proposals sent and deals won.
All six of those live in the platform already if your forms, calendars and pipeline run there. The AI part is the commentary that explains which of them moved and why.
What the AI does and does not do
The AI reads the figures and the underlying records and writes short explanations: enquiries fell in one service line, reply times rose on Thursday when two people were out, the show rate dropped after the reminder sequence was paused. It is very good at noticing a change and pointing to the records behind it.
It does not decide anything. It can mistake correlation for cause, and it cannot see what happened outside the system, such as a competitor launching or a key client going quiet. A person still owns the conclusions.
Before you start
- Forms, calendars and pipelines running on the platform for the parts of the business you want to review.
- Opportunity stages that mean the same thing to every salesperson, written down.
- A named owner for the weekly review, usually the operations or sales lead.
- An agreed list of no more than eight figures for the review.
Building the weekly review
- Agree the eight figures and write a one-line definition for each.
- Go to Dashboard and open the Dashboard Selection Menu.
- Click + Add Dashboard, name it Weekly business review, then Confirm.
- Click + Add a widget manually for each figure.
- Pick the pipeline in widgets such as Conversion Rate to match definitions.
- Click Save Changes, then set seven days in the Date Picker.
- Add the previous seven days as the comparison range.
- Click Summarize Dashboard with AI each Monday morning.
- Send the summary to the leadership team with the dashboard link.
- Run the review for four weeks, then remove any figure nobody discussed.
What it replaces
It replaces the Monday spreadsheet and the person who builds it. In a 60-person company, that is often a senior operations manager spending half a day pulling exports from the CRM, the calendar tool and the ad accounts, then reconciling them by hand. The dashboard removes the exports and the AI draft removes most of the writing.
It also replaces the argument about whose numbers are right. Sales, marketing and finance look at the same records. For the specific widgets marketing usually wants on its own page, see choosing the metrics for a marketing metrics dashboard.
Picture the first week: enquiries are steady, but meetings booked are down by a quarter. Suppose the AI commentary points out that the booking link in the follow-up email changed two weeks earlier and now opens a calendar with no slots until the following month. Fixing the link would take minutes. Without a weekly review, a drop like that might not reach the revenue line for another two months. The scenario is hypothetical.
Weekly time
Building the dashboard takes a day. Running it takes the owner twenty minutes on Monday morning to read the commentary, check one surprising figure against the records and decide what to raise. The leadership discussion itself should be fifteen minutes. If it takes longer, there are too many figures.
People ask the AI for insights in general and get a page of generic observations. Ask it specific questions instead: which figure moved most, which records explain the change, and what changed in the setup during the same week.
When not to use it
If your pipeline stages are used inconsistently, fix that before adding AI commentary, because the AI will explain noise with confidence. And if the company is small enough that the founder sees every deal personally, a weekly dashboard without the AI layer is usually enough. The CRM dashboard examples show simpler layouts.
Troubleshooting
The commentary repeats the same points each week. Use the follow-up field to ask only what changed since the previous week.
Figures jump for no visible reason. Usually an import, a bulk tag change or a duplicate merge touched many records at once. Check the audit history for bulk actions in that window.
Pipeline value looks too high. Old opportunities that are effectively dead remain open. Add a rule that flags any opportunity untouched for a set number of days, and close them in a monthly cleanup.
Leaders stop reading the email. It is too long. Cut it to five sentences and put the single most important change in the subject line.
Frequently asked questions
Is business analytics AI accurate enough to trust?
The figures come straight from your records, so they are as accurate as your data. The commentary should be treated as a first draft that a person checks. Accuracy improves sharply when the AI is told exactly which questions to answer.
Can it forecast revenue?
It can describe the current pipeline weighted by stage, which is a reasonable short-term view. It is not a statistical forecasting model. For longer forecasts, combine pipeline data with your finance system.
Can I add data from accounting or project tools?
You can bring data in through imports or webhooks, but keep the weekly review focused on the commercial funnel the platform holds. For the marketing side of the same review, read automating marketing reports.