Alprime

Alprime analyses financial figures and earnings — precise but time-consuming work. Together we automated the evaluation and preparation. Result: over 600 hours a year that no longer go into routine but into analysis.

Starting point

Alprime works with what everything in financial analysis hangs on: earnings figures and financial ratios. This data arrives in cycles — quarter after quarter, across a whole set of companies under observation. Every single report has to be gathered, checked, brought into a standard form and evaluated.

It is demanding work, but a large part of it is routine: pulling figures from sources, transferring them into the same structure, calculating ratios, preparing them cleanly. These steps repeat with every report and every cycle — and they tie up exactly the hours that should go into the analysis and interpretation itself. The more companies covered, the more this manual effort scales up with them.

The question was not whether time is lost here, but how much — and which part a machine can do more cleanly than a person by hand.

Solution

We made the earnings reports the first lever. Built specifically:

• AI-assisted earnings reports — the evaluation and preparation of financial ratios runs automatically. Data is read in, brought into a standard structure and condensed into a finished report. The manual part shrinks from «gather everything yourself» to «check and interpret».
• More AI projects — from additional data analysis to process automation. This track is extended continuously rather than used up in one big project.

The approach is deliberately sober: not the whole analysis is «automated away» — professional judgement stays with people. What gets automated is the repeatable groundwork that runs the same way for every report. That is where the time is gained, without the quality of the evaluation suffering.

Result

The hard number, confirmed by Alprime: over 600 hours saved per year. These are hours that used to go into manually gathering and preparing earnings data — and are now free for analysis, depth and wider coverage.

What 600 hours mean can be roughly put into perspective: at a conservative assumption of CHF 80 per hour (internal full-cost rate), that is a value equivalent of around CHF 48,000 per year — a rough model calculation, not a guaranteed value. In financial analysis, real hourly rates for qualified staff tend to be higher, which is why this figure is deliberately set at the lower end.

Just as important as the hours: the earnings reports were the entry point, not the end point. Further AI projects run on the same base and are added step by step — data analysis, process automation. The lever grows, instead of fizzling out after the first project.

Transparency on the numbers: the «600+ hours per year» is a real figure confirmed by Alprime and taken over unchanged. The value equivalent of around CHF 48,000/year is our own rough estimate, not a guaranteed number: 600 h × CHF 80/h (a conservative internal full-cost rate for a Swiss SME). Real hourly rates in financial analysis tend to be higher — we deliberately calculate at the lower end, preferring to undershoot rather than overshoot. The assumed volume behind the 600 hours (e.g. recurring earnings cycles across several covered companies) serves only as a plausibility check and should be read as an assumption, not as a reported operating figure of Alprime.

Are the 600 hours a real number or an estimate?

The over 600 hours saved per year are real and confirmed by Alprime — we took them over unchanged. Only the franc equivalent derived from them (around CHF 48,000/year) is estimated, deliberately conservatively at CHF 80 per hour.

Does this mean an AI does the whole financial analysis?

No. What gets automated is the recurring groundwork — reading in data, standardising it, preparing ratios, condensing the report. The professional evaluation and interpretation stay with people. That split is the point: gain time without giving up the quality of the analysis.

Can something like this be built for other finance or data-heavy SMEs?

Yes — wherever the same evaluation repeats in cycles (quarterly figures, reporting, recurring data preparation), the lever is similar. We typically start with the scoping and planning, which works out the concrete hours and a cautious saving figure before anything is built.