Aquila’s Wecan choice: testing Swiss data control
Aquila’s Wecan selection and a Swiss government software study raise a useful buying question: what can an institution demonstrate about control of its data?
Aquila has selected Wecan for client lifecycle management following a competitive procurement process, according to an announcement published by Wecan on 8 September. The supplier says the project covers onboarding, periodic reviews and changes in circumstances, with implementation underway. Read Wecan’s Aquila announcement.
The announcement describes data remaining under Swiss control and AI outcomes that can be explained, reproduced and audited. These are supplier claims about the selected approach. The release is not an independent assessment of the deployed system. Wecan sets out those claims here.
For another institution choosing software, the useful question is what evidence would turn such descriptions into a decision it can defend.
Control has to survive ordinary work
Our analysis starts with a simple distinction. The stated location of data is one procurement fact. An institution also needs to understand who can access it, how instructions change, how an outcome can be checked and what happens when the service is unavailable. A hosting label does not answer all those questions.
A separate Swiss project illustrates why practical testing matters. The Federal Chancellery’s BOSS office-software study involved 172 participants in its browser-based user acceptance test. Its published summary reports suitability in principle for key standard processes, alongside technical, operational and organisational barriers to widespread production use. Results were released on 2 September. Read the BOSS feasibility study summary.
BOSS is a public-administration office-software project, not an evaluation of Wecan. We draw a methodological lesson from it: report working capabilities and unresolved limitations together.
Follow one fictional file from arrival to exit
For procurement teams, we propose a demonstration built around one fictional client relationship. Include an intermediary, a supporting document, a request for clarification and an approval. Ask each shortlisted supplier to handle the same sequence.
- Locate the processing. Identify where the original document, extracted information, temporary copies and backups are handled. Ask which organisations operate each part. Distinguish a demonstrated fact from a contractual promise.
- Show access in practice. Use different staff roles. Attempt to view, amend, approve and export the file. Include the supplier’s support access in the discussion and establish how that access is authorised and recorded.
- Explain an outcome. Select one extracted value or proposed classification. Trace it to the input, the processing version and any human correction. Ask the supplier to show what can actually be reproduced.
- Change an instruction. Replace a document or withdraw an approval. Check whether the earlier state and the reason for the change remain distinguishable from the current state.
- Leave or lose the service. Export the material and inspect whether another authorised person can read the documents, messages and decision history. Discuss how essential work would continue during an outage.
These are proposed questions, not allegations about either project. They turn a broad claim of control into a set of observable tasks.
Keep a record of what the demonstration did not prove
For each step, distinguish “demonstrated”, “documented but not tested” and “unresolved”. Name the person responsible for resolving the gap and the evidence required. A well-run demonstration should be allowed to produce an incomplete answer.
For example, an export may contain every attachment but omit the link between an approval and the document version it concerned. Counting files would miss that gap. Asking a second person to reconstruct the exchange would reveal it.
This is the formal communication question within procurement: can the institution preserve an intelligible connection between a participant, an instruction, its authority and the retained record? The answer should remain useful after a member of staff leaves or the supplier changes.
The immediate takeaway is to attach a completed file walkthrough to the buying decision. It gives decision-makers a record of what they saw, what they relied on and what remains to be checked. Our AI pilot scorecard applies the same discipline to task quality and review time.
We compared Wecan’s announcement with the Federal Chancellery’s published BOSS summary, accessed on 29 September 2026. The projects have different purposes and are not compared as competing products. We did not inspect Aquila’s implementation, contracts or technical controls. The fictional-file walkthrough is our proposed procurement exercise, not a Swiss legal or certification standard.