Investment and alignment are different states

Kyndryl research finds that while 95% of senior executives report investing in AI, only 14% feel they have successfully aligned their workforce, technology and business goals around that investment. This is not a failure of intent. Every organisation in the 95% intends for AI to work. The gap is structural: investment decisions and alignment decisions are different types of work, and most organisations make the first without designing the second. Investment puts the technology in place. Alignment ensures the organisation is structured to use it. These are sequential requirements, and the sequencing is not optional.

What alignment actually requires

Alignment has three components that must be designed rather than assumed. Strategic alignment: the AI strategy specifies what the organisation will and will not do, and that specification has been communicated to the people who make day-to-day decisions, not just to the leadership team. Workflow alignment: the processes in which AI is deployed have been redesigned so that AI output is usable within existing accountability structures rather than conflicting with them. Capability alignment: the people expected to operate AI workflows have the judgment, checkpoint criteria and escalation knowledge required for their specific role, not general AI literacy but role-specific competence in the workflow they operate. Investment funds the tools. These three conditions determine whether the tools produce aligned outcomes.

Test alignment before it is assumed

Assess alignment against three specific questions. Can frontline practitioners articulate two or three things the AI strategy explicitly rules out, not just what it is for but what it has decided against? Are there workflows in production where the AI output is reviewed against documented criteria rather than individual judgment? Has each practitioner operating an AI workflow received training specific to their checkpoint responsibilities, not general AI training? A no to any of these is a specific alignment gap with a specific design remedy. The organisations in the 14% answered yes to all three before they counted themselves aligned.

·
Ready to build institutional AI capability?
Let's talk about your operating model.

setmode.io is a 14-module programme that closes the gap between AI adoption and organisational transformation. Every engagement produces named deliverables that form an AI-enabled operating model.

Start a conversation