Scroll down to discover more
Scroll down to discover more
Back

Pronect: Productising SAM, a GenAI-Powered Security Advisor Agent

Pronect partnered with Faktion to develop SAM (Security Advisor Module), an AI-enabled security adviser within its platform. The GenAI copilot merges policy data, compliance standards, and real-time control metrics to detect gaps, suggest improvements, and enhance decision-making in physical security operations.

Pronect is a European technology company that helps large organisations take control of their physical security operations. Its AI-powered platform functions as a 360° cockpit, integrating data across risk, performance, and cost to support better decision-making. Pronect aims to address the challenges their client face in physical security operations:

  • Manual Insight Retrieval: Clients needed to manually interpret reports and documents to understand risk and performance.
  • Limited Automation: Tasks such as compliance reviews, reporting, and KPI tracking still required significant human effort.
  • Accessibility of Expertise: Security knowledge and documentation were not always readily accessible to all platform users, especially non-specialists.

With a clear vision for innovation, Pronect aimed to create an AI-enabled adviser module, capable of delivering contextual insights, guiding users in risk and compliance, and automating routine security tasks. This required a highly specialised application of generative AI within a critical, regulated domain.

AI offers great potential but needs significant effort for data security and quality, making specialists like Faktion essential.
Jeroen De Bock
CEO & Founder
of
Pronect

Need

An easily accessible solution that provides comprehensive, real-time insights, automates regulatory monitoring, and supports human experts in decision-making.

Challenge

Enhancing AI-Driven Security at Scale

The solution

Security Advisor Module Powered by Generative AI

To achieve this, Pronect partnered with Faktion to co-create SAM (Security Adviser Module). The goal was not to improve an existing tool, but to jointly develop a new capability that could elevate the entire platform; from structured and robust, to intelligent and adaptive.

The solution delivered superior capabilities in data consolidation, real-time analytics, and client interaction by:

AI Knowledge Assistant

Providing precise, contextualised answers by drawing on available source data to address fragmented data insights.

Domain-Specific Semantic Layer

Enabling relevant recommendations and scalability without compromising analytical depth through real-world security expertise.

Intuitive Conversational Interface

A Fully integrated chatbot experience with source attribution and personalised insights allowing clients to interact directly with complex security data and receive clear, actionable insights beyond mere data consumption.

Data Processing & Knowledge Base Structuring

Unifying diverse data sources, merging external standards (e.g., ISO, TAPA) with internal policies and dynamic security data for a comprehensive risk overview and analysis precision.

AI Knowledge Assistant

Providing precise, contextualised answers by drawing on available source data to address fragmented data insights.

Domain-Specific Semantic Layer

Enabling relevant recommendations and scalability without compromising analytical depth through real-world security expertise.

Intuitive Conversational Interface

A Fully integrated chatbot experience with source attribution and personalised insights allowing clients to interact directly with complex security data and receive clear, actionable insights beyond mere data consumption.

Data Processing & Knowledge Base Structuring

Unifying diverse data sources, merging external standards (e.g., ISO, TAPA) with internal policies and dynamic security data for a comprehensive risk overview and analysis precision.

Reusable Tools for Building Copilots

Faktion’s leveraged it's reusable tools for developing AI copilots facilitate streamlined integration and customisation, ensuring a cost-efficient implementation and a solution that is robust and adaptable.

Approach

Structured AI Integration

Faktion adopted a structured, phased integration methodology, grounded in best practices for LLMOps, Azure-based deployment, and advanced data governance. The structured approach included:

  • Data Preparation and Ingestion: Static PDF/text datasets provided by Pronect were systematically ingested and enriched using metadata schemas, facilitating optimised RAG functionalities and precision in output.
  • Specialised Generative AI: A foundational model (Mistral/GPT-4 class) was selected and fine-tuned specifically for security management, ensuring that every generated recommendation aligns with proven security practices, regulatory requirements, and domain relevance.
  • Deployment Support and Training: Hands-on deployment assistance was provided to Pronect’s Azure team, including reusable deployment scripts (Bicep) and knowledge transfer to ensure independent scalability post-project.
  • Continuous Validation and Iteration: An evaluation notebook for model assessment was implemented, measuring metrics such as answer correctness, context relevance, and recall. Iterative improvements were driven by continuous client feedback loops and performance assessments.

Outcome

Intelligent Platform, Scalable Value

The successful implementation of SAM represented a major milestone in Pronect’s product evolution, shifting from a structured, process-led platform to a truly AI-embedded solution. The benefits of this transformation are being experienced directly by Pronect’s clients.

Key value delivered to end users:

  • Faster Compliance and Reporting: Tasks like compliance verification, KPI reviews, and risk summaries can be completed through simple natural language queries, saving hours for end users.
  • Confident Decision-Making: Contextualised recommendations are backed by traceable sources, ensuring defensible, well-informed decisions across security operations.
  • Accessible Security Intelligence: Complex security data is transformed into clear, easily navigable insights, making critical information accessible to all users regardless of technical or domain expertise.
  • Standardised Best Practices: Security teams can standardise and automate procedures across sites, units, and geographies without adding manual workload.
  • Reduced Reporting Burden: Automated, real-time summaries and insights minimise manual reporting and keep leadership aligned with operational realities.

The SAM project illustrates what’s possible when domain expertise and AI engineering come together. Faktion and Pronect, together co-created a pioneering capability that redefines how physical security can be managed in a digital world.