Kyndryl, a global IT services provider, launched the Agentic AI Framework on July 18, 2025, in Singapore. The framework is designed to improve business operations by deploying self-learning AI agents.

It helps enterprises scale AI solutions across on-premises, cloud, or hybrid environments.

It features advanced algorithms and a secure design to support complex IT workflows while ensuring human oversight.

The framework has garnered interest from various sectors.

A national government is considering it to enhance public services and compliance. A financial institution is exploring its use for automating compliance and optimizing IT operations.

🔗 Source: Kyndryl

🧠 Food for thought

1️⃣ Enterprise AI adoption faces a significant trust gap

Kyndryl’s announcement directly addresses the skepticism in AI implementation that continues to limit business impact.

According to Kyndryl’s own research, only 4 in 10 leaders currently use AI-powered insights for decision-making, highlighting the gap between AI’s potential and actual business utilization.

This trust deficit stems from several factors, including AI’s “black box” nature, which makes it difficult for users to understand decision-making processes, as well as concerns about data quality affecting outcomes.

Organizations specifically struggle with integration challenges when implementing AI with existing systems, creating technical barriers to adoption even when leadership is supportive.

Kyndryl’s emphasis on “humans in the loop,” traceability, and “secure-by-design” principles directly addresses these documented trust barriers that have limited enterprise AI deployment.

2️⃣ The unofficial-official AI usage gap creates governance challenges

Research indicates a significant disconnect between sanctioned and unsanctioned AI use in enterprises that Kyndryl aims to resolve.

While only 36% of professionals report using corporate AI capabilities, a much larger 66% admit to using free AI tools like ChatGPT, creating substantial “shadow AI” risks for enterprises.