AI as a defensive tool: how three CISOs describe AI detection and threat operations

Three CISOs, three kinds of document

This explainer draws on three company publications. Each is tied to a chief information security officer. Synechron published an undated article on vulnerability detection. Its CISO, Aaron Momin, co-wrote it with senior architect Bharadwaj D Jagannath.

Kai, a security platform company, put out a press release dated March 19, 2026, naming its new CISO. The release says the company announced the appointment of Alfredo Hickman. iCOUNTER put out its own press release on June 23, 2026, naming Ali Waezzadah as CISO.

The three sources differ in form. Synechron's piece explains how a tool is meant to work. Kai's release gives a short quote from its new CISO alongside the company's product description. iCOUNTER's release gives a quote and a mission statement. The views on AI below should be read with that in mind: each comes from a company describing its own work.

How AI is described in vulnerability detection

The Synechron article frames the stakes in its first line. It says a hacker needs access to only one small weakness for a company to lose millions.

The authors then describe how vulnerability detection works without AI. In their account, the work combines automated scanning, expert review and checks against public databases such as CVE and the NVD. They add that teams usually rank findings by severity scores from the Common Vulnerability Scoring System (CVSS).

Against that baseline, the authors write that AI tools process data and code faster than older methods and produce fewer false positives. They also claim AI can look ahead, predicting likely weaknesses from past data and new threat patterns.

In this account, AI does not replace the scanning step. It changes how fast findings are processed, how accurate they are, and how they are put in context.

What the Synechron tool is said to do

The article uses Synechron's own accelerator, AppSec.AI, as its example. The authors say the tool was due for release within a month of writing.

The first function listed is pulling results from several scanner types into one view. The article also describes ranking findings on more than a severity score. That ranking adds whether an exploit exists and how critical the affected application is.

The other listed functions are of the same kind:

• removing duplicate findings
• automating the steps from starting a scan to producing a report
• calculating a financial score for findings left unfixed
• offering a chat assistant that analysts can use to query the data

Read together, the list puts most of the AI work in sorting, merging, ranking and explaining findings that existing scanners already produce.

How AI is described in security operations

Kai's release moves from finding vulnerabilities to running security work. Hickman's quote opens with the problem he sees on the defender's side. He says many teams are stuck with manual work, scattered tools and too many alerts.

The company describes its answer as AI agents that carry out security tasks themselves, not tools that only flag them.

The release also sets out how Hickman will use the platform. He will run it inside Kai's own environment as an internal test of the product. Under this arrangement, the CISO's job covers securing the company and also acting as an internal user and tester of the AI system the company sells.

Proactive threat operations at iCOUNTER

The iCOUNTER release does not use the term AI. It frames the company's work as moving customers from reacting to attacks toward acting before them.

Waezzadah's quote names the factors he thinks matter most now: the speed of decisions, intelligence about risk, and execution. The company says its platform spots adversary intent early and drives fixes across a customer's outside partners.

The release also says his role faces outward as well as inward. In this document, "threat operations" covers intelligence, third-party risk and fixing problems. It is described without reference to any specific AI method.

Where the limits are drawn

The documents give limits less space than claims, but they do name some.

Synechron's article has a short section on challenges. It warns that AI brings its own weaknesses into the systems it is meant to protect. The same section says security, privacy and compliance measures are needed to protect sensitive data and keep trust.

The article also frames several gains as things to come rather than things already delivered. Among the advances it says to expect is a sharp drop in false positives and duplicate findings. That sits next to the article's earlier claim that AI already reduces false positives. On this account, the problem is not yet solved.

Hickman puts his limit as a condition. In his view, AI changes security work only when it fits how practitioners actually work. The Kai release also uses human performance as its benchmark: it describes the platform as working at machine speed "with human expert accuracy".

The iCOUNTER release does not discuss limits of AI, because it does not discuss AI directly.

The career paths behind the views

The three CISOs reached their posts by different routes. Synechron's author note credits Momin with close to thirty years across risk, security and continuity work. The same note says he held roles at PwC, Ernst & Young and Accenture before becoming a CISO.

Kai's release says Hickman's previous CISO job involved building a security function at Obsidian Security from scratch. It also lists earlier work at Rackspace Managed Security, CSIS and the U.S. Department of Defense, and service in the Marine Corps.

iCOUNTER's release points to a large-enterprise background. It says Waezzadah came from the CISO role at Frontier Communications, which ended with that company's sale to Verizon.

All three now work at companies that sell security technology or services. In each case, the CISO's view of AI is set out in documents that also describe what the employer offers.