Intelligence is the work of turning data into decisions. Until now by hand.

Every intelligence area has its sources. In product management, these are competitor websites, price lists and trade fair programmes. In sales, the commercial register, job advertisements and the prospect's company page. In wealth management, prices, fund data and the investment policy. Added to this are the internal systems against which everything has to be compared: CRM, ERP, filing systems.

By hand means: someone opens these sources when there is time, usually on Mondays, usually one person. An AI employee observes them continuously and remembers what has changed since the last time. Which sources these are is something you define during scoping. They remain your sources, it simply reads them more often than a person could.

The most invisible step, and the one that costs the most time. What sits in twenty windows has to be brought into one form: the same units, the same names, the same time periods. In the process, typing errors, duplicates and the famous question in the meeting arise about which version of the table is actually the current one.

An AI employee writes every entry, with its source and date, into the structure you specify. It recognises what is duplicated. It marks what is contradictory. The result sits in a knowledge base and can be called up from there in Excel, Power BI or SharePoint, in your structure, with your column names.

A competitor price is a number. A competitor price next to your own price list is a statement. The same applies to a portfolio against the investment policy, an enquiry against the target customer profile, a progress figure against the plan. In this step, companies differ most from one another, because each one calculates and weights differently.

This is why the AI employee adopts your calculation logic rather than one of its own. During scoping, you define what is compared against what and which thresholds apply. After that, the comparison runs the same way every time, with the calculation shown at each figure. What was previously held in an Excel formula that only one person knew can now be checked by anyone.

Of a hundred signals, three are relevant. Which ones is known only to someone who understands the market, the customer or the mandate. In most companies this step is carried out by the most experienced people, and it is crowded out by the steps before it: anyone who spends the morning collecting has half an hour left for assessing.

An AI employee proposes an assessment: what is new, what deviates, what needs a decision. It supports every proposal with the data behind it. The judgement stays with your team, which now has the morning for it. From the feedback it learns what counts as important in your company and what does not.

The board reads one page, the specialist department ten, sales a Teams message before the meeting. Each recipient needs a different form, and each form is produced by hand today: building slides, adding sources, maintaining distribution lists. And the next day comes the query about where the figure on page four came from.

An AI employee produces the report in the corporate design, the newsletter and the Teams message, each recipient in the agreed form, with a source and date on every figure. It sends actively and gets in touch when something important happens, instead of waiting for the next reporting date. The query about page four is answered by the reference that stands next to the figure.

Market sizes, prices, raw materials and regulation, competitors and target customers, your own portfolio and the risks in the observed field: four perspectives that otherwise sit in four departments come together as one information basis. Every entry carries a source and date.

Sales spends its time in conversations rather than on preparation. Enquiries arrive with a company profile, a trigger and a priority, researched from the commercial register, the website, job adverts and the CRM. Guided conversations come back to the team evaluated, measured against the metrics you defined beforehand.

Every night an AI employee checks mandates against the investment policy and brings together metrics, a peer comparison and a risk picture for each fund, with data from verified sources. In the morning every deviation is set out with a rationale, together with a draft of the advisory documentation. Assessment and decision remain with your advisers and your institution.

An AI employee requests interim updates from the teams, pulls metrics from the connected systems and prepares the quarterly reporting, with the source on every figure. Leadership and the specialist departments see the same status at the same time. Your leadership round discusses deviations, not slides.