Case Studies

HUDRA, multi-agent AI B2B prospecting system: nearly, 1000 qualified meetings in four months.

The AI that finds your prospects, runs the conversations and books nearly 1,000 qualified meetings

Key takeaways In a few days, the system builds up to 100,000 qualified prospect records, the equivalent of several years of manual research. Over a four-month deployment (February to May 2026), it generated nearly 1,000 qualified meetings, with around 95% of conversations handled without human intervention. The cost per qualified meeting was cut by more than 50 times compared with human prospecting.

~95%of conversations handled without human intervention

See the case study

QubiLEX sorts judicial documents for a Geneva law firm: 98% handled without intervention, the lawyer reviews and sends.

98% of court documents sorted, filed and notified by AI, the lawyer reviews and sends

Key takeaways For six months, the system has been running in daily production in a firm of five lawyers in Geneva, processing around 30 court documents per day (summonses, briefs, judgments, orders, notices of appeal). It understands each document, extracts the case reference, the parties and the deadlines, identifies the client, files the document in its folder and drafts the notification letter, which the lawyer reviews before sending. In the end, 98% of documents are processed without intervention and 96% are filed automatically under the correct client. No letter is sent without a lawyer's approval, and only documents from approved addresses are accepted by the system. The firm thus saves about seven minutes per document, roughly 70 hours per month returned to legal

98%of court documents processed end to end, without intervention

See the case study

Member management dashboard for a Geneva association — from 160 to 20 hours of administration per year

From 160 to 20 hours a year: automated member management for a nonprofit

Key takeaways A Geneva-based nonprofit with 400 members managed its dues, statuses, invoices, reminders and bank reconciliations by hand, amounting to roughly 160 hours of volunteer work per year. Qubitech automated the entire cycle, from online registration to the receipt and the reconciliation of bank payments. Management time dropped from 160 to about twenty hours per year: member and dues administration from 70 to 10 hours, and payment tracking and collection from 90 to 10 hours. 94% of payments are matched automatically. The data can remain fully sovereign, hosted locally or with a Swiss provider such as Infomaniak or Proton.

-86%less time on member and dues administration

See the case study