Frequently Asked Questions

Concretely: which tables and records an agent can reach versus what anyone intended it to reach; whether it can be talked into ignoring its instructions through prompt injection; whether data can be pulled out through its tool calls; and what a sequence of ordinary-looking questions can reconstruct about records it will not return directly. Agents leak through what they are willing to answer, not only through what they store. You get findings you can hand to an auditor, not a score out of ten.

Because the platform shipping your agents cannot also be the party certifying them safe. AI Control Tower is a good product and we implement it — but a vendor assessing its own output is not an audit, and an auditor or a regulator will read it that way too. We are independent of ServiceNow, which is the entire point of buying this separately.

Obligations for high-risk systems under Annex III fall due on 2 December 2027, after the deferral from May 2026. The work is risk classification, technical documentation and demonstrable monitoring. Harmonised standards are running late, so the practical burden lands before the guidance does. We would rather sell you real security than compliance panic: the date has moved once already, and the underlying work is worth doing whether or not it moves again.

Yes. Scoped apps, IntegrationHub spokes, SAP and Workday connectors and multi-instance synchronisation are what put us inside these systems in the first place, and ITSM on BMC Remedy and asset management on IBM Maximo are still live work — including migrating off them when that is the honest recommendation.

The person who scopes it. We are small on purpose: there is no account layer between you and the engineer, which is why a question gets an answer the same day, and why we will tell you when the thing you asked for is not the thing you need. Offices in Prague and Eindhoven; ten years inside enterprise ServiceNow instances.