According to Anekanta, the reality is that too many AI systems never cross from pilot into production. What holds them there is seldom capability – it is trust – which if designed in with intent, allows the value to convert into commercial return. This is an excerpt of a longer piece. Find the links below.
In McKinsey’s 2026 AI Trust Maturity Survey, close to two-thirds of organisations named security and risk concerns as their main barrier to scaling agentic AI, ahead of both regulatory uncertainty and technical limitations. The binding constraint is not the capability of the technology but the confidence to deploy it.
The same research finds trust increasingly regarded as a business enabler rather than a compliance exercise, with adoption now driven by value and performance rather than by regulation alone. Trust does two things at once: it allows an organisation to realise value from its AI by supporting sustained adoption, and it allows it to manage a widening landscape of risk as systems take on greater autonomy.
Trust is the precondition for return, not a reward for it
Trust is not a by-product of commercial success, acquired once a product is already selling. It is the condition that must be met before the value in an AI system can be released at all. The capability is usually present in the pilot, what is missing is the assurance that lets a buyer purchase it, a board approve it into production, a regulator permit it and an insurer underwrite it. Each of those is a commercial gate, and trust is what opens it. Withhold it, and the value stays stranded, however capable the model.
The relationship is measurable. McKinsey reports that organisations making substantial investment in responsible-AI practices achieve both higher trust maturity and a materially greater likelihood of realising financial benefit. This includes earnings impact above five percent, and frames that investment as a driver of value rather than a constraint on innovation. Trust is therefore necessary, though not on its own sufficient: a well-governed AI product that no one wants will still fail.
For the large majority of organisations whose AI works and yet earns little, the missing factor is not a better model. It is the demonstrable trustworthiness that turns a capable system into an adopted one.
The dimensions of trust, and the instruments that build it
Trust of this kind is not achieved by a single act of compliance. It has several dimensions, whether an AI system is safe and reliable, whether it is fair and free from unlawful bias, whether its outputs can be explained, whether it is accountable to identifiable owners, and whether all of this can be evidenced to someone outside the organisation.
No single instrument addresses them all, and regulation is only one of the available routes. The EU AI Act makes certain of these obligations mandatory for the European market. Beyond it sits a range of instruments an organisation can adopt voluntarily, and for commercial reasons: the ISO/IEC 42001 standard, which provides a certifiable AI management system; the NIST AI Risk Management Framework, an internationally recognised structure for identifying and managing AI risk, and principle-based frameworks, notably the OECD AI Principles and the 12 Principles of AI governance that Anekanta® developed into a commercial context. These translate high-level values into decisions a board can actually take. Used together, these are how an organisation evidences, across every dimension a customer, regulator, insurer or investor cares about, that its AI can be trusted. This article considers the regulatory position first, because its timetable has just changed, before looking at how the wider set of instruments can be put to work.
How Anekanta® can help
Anekanta® helps boards, CEOs and senior leaders turn AI ambition into trustworthy, commercially viable adoption, working across AI strategy and use-case evaluation, EU AI Act risk assessment and classification, AI literacy for senior leaders and across enterprises, and preparation for ISO/IEC 42001. Explore our services further.
To build board-level competence and confidence, our AI Literacy and Governance Workshops move from principles to practice. The next open session is hosted by the ISACA London Chapter on 23 September 2026.
To assess where your organisation’s AI portfolio stands, or to prepare for ISO/IEC 42001, begin a specialist advisory enquiry with our team.
Read the full piece here
For more Anekanta news, click here
