Cybersecurity
Humanix Launches Complete Call Context to Detect Service Desk Voice Attacks

Key PointsHumanix correlates caller claims with identity and ticketing records to expose voice…
View more
- Humanix correlates caller claims with identity and ticketing records to expose voice social engineering.
- Logs can show an authorized MFA reset while hiding the manipulation that persuaded an employee to approve it.
- The release offers detection and alerting but does not claim to block resets or reverse account changes.
Humanix has launched Complete Call Context, a capability that connects service desk conversations with identity records, support tickets, and an organization’s account-recovery procedures. The aim is to expose voice-based social engineering that can disappear into otherwise routine password resets and multifactor authentication changes.
The update extends Humanix’s Human Threat Detection and Response platform. Its micro-agents examine sensitive requests individually, identify the evidence a company’s procedure requires, and compare what the caller says with records in identity provider and IT service management systems. Contradictions or missing evidence generate an alert with a severity rating and an explanation.
For security teams, the problem is the gap between an authorized action and the conversation that prompted it. An account log may accurately show that a help desk employee enrolled a new MFA factor. It may reveal little about whether someone persuaded that employee to bypass the checks intended to protect the account.
Following a Request From the Call to the Record
Complete Call Context creates a micro-agent to analyze each sensitive service request, such as an MFA reset. The organization’s own procedures determine which context matters. The agent then correlates the interaction with relevant identity and ticketing records, bringing those separate sources into the same investigation.
Humanix describes an example in which a caller claims to manage an employee and asks for an MFA reset on that person’s behalf. The caller supplies a ticket number and says the employee has already completed identity verification. Each statement creates something the system can check: the manager relationship, the ticket’s existence, and a record of the verification.
If the directory identifies a different manager, the ticket cannot be found, or the identity provider has no corresponding verification event, the alert connects those findings to the required procedure. These are illustrative discrepancies, rather than a claim that every missing record proves an attack.
The supplied product image shows that reasoning in a single view, with call evidence, account activity, and policy checks labeled as compliant, indeterminate, or violated. The value is in making the relationship between a suspicious request and the resulting access change visible to an analyst.
Conversational AI Adds Context to Security Logs
Humanix’s broader platform uses conversational AI to identify manipulation, impersonation, and attempts to bypass required procedures. Its published platform description explains that it connects to enterprise communication and service systems through APIs and evaluates the language, identity claims, behavior, and workflow context surrounding an interaction.
This differs from asking only whether a voice is synthetic. A caller using their own voice can still impersonate a manager or pressure a support worker to skip verification. A convincing voice does not establish that the requested action is authorized.
Complete Call Context adds corroboration to that conversational analysis. It tests a claim against operational evidence, rather than relying entirely on how persuasive or suspicious the conversation sounds. The release does not specify which underlying models power the micro-agents or publish accuracy, latency, or false-positive benchmarks for the new capability.
In an earlier Unite.AI interview, founder and CEO Keith Stewart described social engineering as an attack on business processes and trust. The new release makes that idea concrete by tying the conversation to the records and checks that govern account recovery.
Alerts Bring Together Evidence and Procedure
“AI-assisted attackers can now go from a single phone call to a full compromise in minutes,” Stewart said in the announcement. “Controls that only engage after the call can’t keep up.”
Humanix says its platform analyzes calls as they happen, while Complete Call Context extends that view into systems of record. An alert assembles the call, the account change, and supporting findings so security teams can assess what happened without first reconstructing the sequence across separate tools.
The announcement describes detection and alerting; it does not establish that the feature automatically blocks every reset or reverses an account change. Its usefulness depends on the evidence available from connected systems and the procedures against which each request is evaluated.
Availability and the Broader Security Workflow
Complete Call Context is available to Humanix customers. The company is backed by Acrew Capital, Boldstart Ventures, and Evolution Equity Partners, and was named a 2026 RSAC Innovation Sandbox Top 10 finalist.
The launch addresses a practical weakness in identity security: a sensitive action can be correctly logged while the deception behind it remains unseen. Connecting the caller’s claims, the organization’s verification requirements, and the eventual account activity gives defenders a clearer basis for deciding whether a familiar support workflow has become an intrusion.












