As organisations move from AI pilots to scaled AI and autonomous agents, governance has become a business imperative, not just a compliance exercise. Given the speed of technological change and AI’s disruptive potential, there is growing recognition that AI must be implemented responsibly, ethically and with trust built in from day one.
This is according to Pranesh Kara, Head of Digital Risk Transformation and Trusted AI at KPMG, who says: “Organisations are expanding their AI use cases and taking a more strategic approach to governance. They need robust AI frameworks that accelerate innovation, manage risk and strengthen transparency. But, to scale AI and autonomous agents with confidence, organisations must also embed trust across every stage of the life cycle to turn responsible AI principles into practical action that unlocks measurable and sustainable business value.”
As AI adoption accelerates – so do the potential risks
Kara points to the KPMG Global tech report 2026,[1] which found that 74% of technology executives say their AI use cases are delivering business value, although only 24% report achieving ROI across multiple use cases.
At the same time, AI adoption continues to accelerate. KPMG’s global research[i] shows that 88% of organisations are already embedding AI agents into their workflows, products and value streams, while 69% of early adopters plan to scale AI widely by the end of 2026.
However, increased adoption is also exposing organisations to new operational and reputational risks. A recent KPMG AI trust research report,[2] conducted in collaboration with the University of Melbourne, found that nearly half of employees admit to using AI in ways that contravene company policies, including uploading sensitive company information into public AI tools, such as ChatGPT. The research also found that 66% of employees rely on AI-generated outputs without validating accuracy, while 56% say AI-related errors have affected their work.
According to Kara, these findings reinforce the importance of transparency, accountability and strong data foundations to support the accuracy and reliability of trusted AI outcomes.
“For many organisations, accelerating AI innovation responsibly has become a balancing act,” he says.
Preparing governance for the next frontier
Kara believes that organisations need AI-tailored risk strategies that are flexible enough to evolve alongside emerging technologies and the potential for future disruption. He explains: “While organisations are focused on AI and building trusted governance frameworks today, they also need to prepare for what comes next, including advances such as quantum computing."
The governance foundations being established now should be adaptable enough to support future waves of technological transformation. That thinking is also reflected in KPMG’s recently announced global alliance with Anthropic and the launch of KPMG Digital Gateway Powered by Claude, which brings Anthropic’s frontier AI directly into KPMG’s client delivery platform. The alliance combines Anthropic’s frontier AI, KPMG professionals’ domain experience and a shared commitment to security and trust in AI to help organisations redesign workflows, strengthen oversight, manage risk and operationalise AI responsibly at scale.
Trusted data remains the foundation of trusted AI
Kara emphasises the critical importance of data quality and integrity as central priorities in scaling up organisations’ AI capabilities, emphasising: “The reliability of data is foundational to building trusted AI models and use cases. Many organisations are now strengthening their data environments, retiring low-value legacy systems and modernising technology stacks to better support their AI use cases.”
To further strengthen KPMG’s trusted AI ecosystem, it collaborates closely with leading technology, data and services companies, and not-for-profit organisations to offer global reach and combined abilities to help solve your most pressing challenges, including sustainable growth, regulatory change and business protection.
KPMG works alongside its very own KPMG Lighthouse, a global network of data professionals, scientists and engineers focused on helping organisations unlock value from trusted data and responsible AI initiatives.
A trusted framework for operationalising AI trust
To help organisations manage AI responsibly at scale, KPMG has developed an AI Trust Framework designed to embed trust into every step of the AI and AI agent life cycle from day one.
From design and development through deployment, monitoring and ongoing evaluation, the framework is intended to help organisations accelerate AI adoption while maintaining ethical, secure and compliant practices. It can support CIOs, CROs and CAEs in identifying AI risks, establishing policies and implementing appropriate control procedures for AI development and deployment across 10 key focus areas: fairness, reliability, transparency, security, explainability, safety, accountability, privacy, data and sustainability.
“Our framework also aligns with major global standards and regulatory approaches, including the EU AI Act, ISO 42001 and NIST security frameworks,” Kara says. “We take a structured approach to identifying AI risks and designing controls based on each organisation’s AI inventory, maturity level and existing governance environment.”
He adds: “Organisations are increasingly focused not only on managing AI risk, but also on understanding whether their AI investments are delivering sustainable business value. While many organisations are developing AI use cases, they still need clearer ways to measure tangible outcomes and return on investment. We help them assess performance, identify opportunities for improvement and maximise the long-term value, efficiency and impact of their AI initiatives.”
To learn more about the KPMG AI Trust Framework and how it can support responsible AI adoption, contact Pranesh Kara, Head of Digital Risk Transformation and Trusted AI, KPMG. (Pranesh.kara@kpmg.co.za
[1] KPMG Global tech report 2026
[2] Gillespie, N., Lockey, S., Ward, T., Macdade, A., & Hassed, G. (2025). Trust, attitudes and use of artificial intelligence: A global study 2025. The University of Melbourne and KPMG. DOI 10.26188/28822919.
Share
