Seven priority AI risks identified in two weeks, with a clear plan to resolve them

Synetec helped an FCA-authorised asset manager identify critical vulnerabilities in a business-critical AI tool, delivering an independent assessment and prioritised remediation plan in just two weeks.

We needed to know the tool could be relied on without someone worrying about it every day. Synetec found things we did not know were there, told us plainly which ones mattered, and gave us a clear path to fix them.

Synetec helped an FCA-authorised asset manager identify critical vulnerabilities in a business-critical AI tool, delivering an independent assessment and prioritised remediation plan in just two weeks.

‍

Overview

‍

An FCA-authorised asset manager in London had developed an internal AI tool to support its investment team’s daily research, document retrieval and routine activities. Originally developed and maintained by a single member of the team, the tool had become increasingly important to everyday operations.

A file deletion incident prompted the firm to temporarily disable the system while investigating the cause. The immediate demand from employees to restore access highlighted how heavily the business had come to rely on it.

The firm engaged Synetec to independently assess the AI system, its security controls and long-term supportability. Within two weeks, Synetec identified seven priority risks, including two previously unknown critical vulnerabilities, and provided a prioritised remediation plan. The most serious issues were already being addressed before the assessment concluded.

‍

Challenges

Background Context

The asset manager had spent just over a year developing the AI tool, which had become embedded in daily investment workflows. Although it was delivering operational value, its development and maintenance depended on one individual.

The incident exposed the need for independent assurance and a clearer understanding of the risks associated with a system that had become operationally important.

‍

Specific Challenges

- Business-critical dependency: The AI tool was essential to daily research and routine activities.

- Key-person risk: Development, maintenance and technical knowledge were concentrated in one individual.

- Security vulnerabilities: The firm needed independent assurance of the application and its controls.

- Operational resilience: The incident raised questions about reliability and disruption management.

- Limited risk visibility: The firm needed a complete, prioritised picture of vulnerabilities.

‍

Solution

‍

Initial Planning

Synetec worked directly with the person who developed and maintained the tool to understand its architecture, functionality and design decisions. This collaborative approach retained first-hand technical knowledge while providing independent scrutiny.

‍

Proposed Solution

Synetec delivered an independent review of the AI tool, its controls and supportability in two weeks. The review covered collaborative technical assessment, code-level examination using enterprise-grade tooling, comprehensive risk identification and prioritised remediation planning.

‍

Implementation Process

Synetec kept the original developer closely involved throughout the review, ensuring the reasoning behind the design was understood and initial fixes could begin promptly.

Each identified issue was verified directly in the application code. Known and newly discovered risks were brought together into one prioritised list, separating urgent issues from those that could safely wait.

‍

Results

‍

‍Quantifiable Results

- Seven priority risks identified: two critical, three high and two medium.

- Two previously unknown critical vulnerabilities uncovered, both being addressed within days.

- Two weeks to complete the independent review.

- Under one month from the first conversation to an agreed plan.

- £16,500 fixed fee plus VAT.

‍

Qualitative Results

The review uncovered two particularly significant vulnerabilities. Credentials associated with the firm’s internal communications could be recovered from the application itself. Separately, a document read by the AI tool could instruct it to send firm data outside the business.

The assessment gave the firm its first complete independent view of a system it relied on every day, with evidence-based findings and a clear order of action.

‍

Long-Term Impact

The review established a clearer foundation for the AI tool’s security, reliability and supportability. It also helped separate key-person risk from the individual: the firm could retain the developer’s valuable knowledge while dismantling the single point of failure created by the existing arrangement.

The findings provided a route for Synetec to support the developer and strengthen the tool’s longer-term resilience.

‍

Client Perspective & Reflections

“We needed to know the tool could be relied on without someone worrying about it every day. Synetec found things we did not know were there, told us plainly which ones mattered, and gave us a clear path to fix them.”

Chief Operating Officer, FCA-authorised asset manager

Client details anonymised at the client’s request.

‍

Speak to an AI Solutions Specialist

If you're exploring how AI could work inside your business, book a 30-minute Clarity Call. We'll help you identify the right opportunities and the right approach. No obligation, No jargon.

Book Your Free 30 Min AI Call
• 30 minutes, no obligation
• Speak to a specialist, not a salesperson
• Leave with a clear next steps
Synetec Logo
Cookie Settings
By using this website, you agree to the storing of cookies on your device to enhance site navigation, analyse site usage, and assist in our marketing efforts. View our Privacy Policy for more information.