By Matthes Derdack, CEO
Introduction
Over the past two years, artificial intelligence (AI) has experienced remarkable growth, significantly influencing various sectors and daily life. In 2023, the release of advanced large language models (LLMs), such as OpenAI’s GPT-4 and Google DeepMind’s Gemini, marked a pivotal shift by enabling AI systems to process and generate diverse data types, including text, images, and audio. This evolution continued into 2024 with the emergence of cost-effective yet powerful AI models like China’s DeepSeek. Concurrently, AI’s integration into practical applications has expanded, with tools like GPT-4 being utilized in chatbots, search engines, and creative industries, thereby streamlining tasks and boosting productivity.
Our approach AI
At SIGNL4, we have been closely monitoring AI’s evolution, practical applications, and limitations. With the release of ChatGPT in early 2023, I delved into the theory of LLMs and explored numerous AI-powered solutions from startups, spanning content creation, image and video generation, podcasting, legal AI, scientific research, and financial analysis.
By late 2023, our team at SIGNL4 concluded that AI had matured enough to provide real value in specific areas, particularly due to its strong capabilities in pattern detection and matching. However, we also identified limitations in reasoning and tasks requiring precision, logic, and complex calculations. Understanding the nature and foundation of LLMs helped us set realistic expectations. With this in mind, we developed a general AI strategy for SIGNL4, testing various AI models, prototyping solutions, and implementing our first ideas.
Implementation Timeline
Our first AI-powered feature, AI-based alert summaries, was launched in October 2024. This tool is particularly useful for interpreting alerts that contain extensive data, including multiple parameters and details. AI-generated summaries help cut through the noise by condensing alerts to the most relevant facts. This feature is easily accessible for each individual alert through the SIGNL4 mobile app.
In mid-February 2025, we introduced AI Insights, a feature that provides regular summaries of alert patterns within a configurable timeframe. This hybrid feature combines traditional calculations, such as average response times, with AI-driven pattern detection. By analyzing historical alert data within a specific team, AI Insights delivers valuable statistical insights and can identify patterns that may indicate potential issues or incidents.
Security and Data Privacy
We recognize that AI adoption comes with concerns about security and data privacy, and we take these issues seriously. To address them, we have implemented four key measures:
- AI usage control: Users can disable AI functionality within SIGNL4 if desired.
- Data privacy compliance: We use AI models hosted entirely within the EU, ensuring compliance with stringent data privacy laws.
- No data training on customer input: We do not use customer data to train AI models. Any information processed by our AI features is not stored or retained within the model, ensuring data security.
- Data minimization: We always submit the least amount of data necessary to fulfil the job.
By incorporating these safeguards, we believe our AI implementation is among the most transparent and privacy-conscious in the industry.
Cost and Pricing
AI usage often incurs significant costs due to token-based billing, where pricing is determined by the amount of data processed by an LLM. Many vendors charge separately for AI-powered add-ons, but we believe this approach can hinder adoption. Instead, we integrate AI into areas where it generates significant value while keeping it within our existing pricing model. This strategy accelerates adoption and ensures that AI is applied only where it delivers genuine benefits at a reasonable cost.
Future Roadmap and Applications
The development of AI Insights as a hybrid solution exemplifies why AI must be carefully applied rather than used indiscriminately. While exploring AI for analytics and reporting, we encountered common issues such as hallucinations and inaccurate data generation. These limitations stem from the fundamental nature of LLMs (you can read more about LLMs and how they work here) – just because AI is a powerful tool does not mean it is the right solution for every problem.
With this in mind, we are continuously testing new AI applications within SIGNL4 and have already started implementing and evaluating additional features from our development pipeline. As of now, we have at least two promising AI-driven features undergoing internal testing.
Potential future applications of AI in SIGNL4 include:
- Alert pattern detection to identify trends and anomalies
- Assessing alert relevance and severity to improve response prioritization
- Generating post-mortem reports based on alert details and response timelines
- Enhancing operational efficiency through AI-driven automation and alert workflows
While it is still too early to disclose upcoming features, rest assured that we are dedicated to delivering well-designed and practical AI-powered enhancements to SIGNL4.
Conclusion
I firmly believe that AI holds immense value in the field of critical alerting and incident management. Over time, AI models will become an integral part of our SaaS product, seamlessly blending into the user experience. However, I also anticipate that AI will become largely invisible—operating in the background, continuously improving processes, and delivering remarkable value, much like microprocessors in modern devices.
Stay tuned for exciting updates on our AI-powered feature pipeline and the upcoming release of cutting-edge AI applications in SIGNL4! And don’t forget, you can always try SIGNL4 for free for 30 days.
