Terms we use, defined.

Short definitions for the words on this site and in our briefings.

Acoustic and prosodic features: Measurable properties of speech, such as pitch, loudness, pace and pauses.; The Voice module uses them alongside what is said.
Calibration: Adjusting an analysis to how a particular person normally speaks or answers.; It keeps an accent or a fast speaking style from being read as a signal.
Claims triage: Sorting incoming insurance claims by how much attention each needs.; Routine claims move fast, and investigators spend their time where an account doesn’t fit the file.
Closed environment: An IT environment the customer controls, separated from outside networks.; Veriteus can run on-premise in one.
Content consistency: Whether what a person says matches the rest of their statement and the documents on file.; The Content module marks where they diverge.
Controller and processor: Under GDPR, the controller decides why and how personal data is processed, and the processor handles it on the controller’s behalf.; Customers using Veriteus remain the controller.
Critical worker: Under Czech critical infrastructure rules, an employee whose role or access carries the most risk for an essential service.; Operators must identify these roles and verify the people in them.
Data Processing Agreement (DPA): A contract under Article 28 GDPR that sets out how a processor handles personal data for a controller.; We sign one with every customer.
Data protection impact assessment (DPIA): A review an organization carries out before processing that is likely to pose a high risk to people’s rights.; Many uses of statement analysis need one.
False positive: A flag raised where, on review, nothing was wrong.; A follow-up question is the quickest way to find one.
First notice of loss (FNOL): The first report a policyholder makes to an insurer after a loss.; It is often a phone call or an online form.
Flag: A passage Veriteus marks because it diverges from the rest of a statement or from the file.; A flag is a reason to ask again, never a decision.
Follow-up question: A question Veriteus suggests at a flagged passage, so the interviewer can check whether a concern holds.
Human in the loop: A setup in which a person reviews every output of an AI system and makes the decision.; Every Veriteus deployment works this way.
Internal reporting system: A channel for employees to report wrongdoing inside an organization.; In Czechia, employers with 50 or more employees must run one under Act No. 171/2023 Coll.
Known-outcome pilot: A test of Veriteus on past cases whose outcome is already known.; It shows how flags compare with what happened, before any live case.
On-premise deployment: Running software on the customer’s own infrastructure.; Recordings and statements then stay inside the customer’s environment.
Reliability check: The vetting Czech critical infrastructure operators must carry out for staff and applicants under Act No. 266/2025 Coll.; Its legal minimum is proof of identity and a criminal-record extract.
Speak-up channel: Another name for an internal reporting or whistleblowing channel.; Reports can be made by name or anonymously.
Statement analysis: Examining what a person says, and how, to see where an account needs a closer look.; Veriteus does it as people speak, against the documents on file.
Threshold: The setting that decides how strong a divergence must be before a flag appears.; Customers set it to fit their process.
Voice-stress analysis: An older family of tools that claimed to read deception from stress in the voice.; Field tests found them no better than chance.

Get in touch