Expect AI to surface authenticity gaps between stated and actual values
An AI system that takes an organisation's stated values seriously will quickly surface where stated and actual behaviour diverge; leadership should expect and plan for these findings before commissioning the work, because surfacing them without being prepared to respond is worse than not surfacing them at all.
An AI system that takes an organisation’s stated purpose, vision and values seriously will quickly highlight where those stated values and actual behaviour diverge. The heuristic names the predictable consequence: any attempt to use AI as an interpreter of PVV — see AI as an operational interpreter of purpose, vision and values — will produce uncomfortable findings before it produces useful ones.
The heuristic is not “do not attempt the work”. It is “expect the uncomfortable findings, and plan for them before commissioning the work”. An organisation that surfaces authenticity gaps and responds to them well ends in a stronger place than one that avoided surfacing them. An organisation that surfaces them and cannot respond — because leadership is not prepared, because the stated values were not really meant, because addressing the gap would cost more than the organisation will spend — is in a weaker place than if the AI had never been pointed at the values in the first place.
Why the gap is predictable
Most organisations’ stated values are aspirational. They describe what the organisation would like to be rather than what it reliably is. The gap is not a secret; staff know it, clients sometimes notice it, and leadership usually knows it too, though it tends to be unnamed. What keeps the aspiration functional is precisely that the gap stays unnamed — everyone can invoke the value in support of their position, no one is forced to acknowledge where it is not met, and the stated value does useful cultural work even without consistent operational backing.
An AI that takes the values literally disrupts that equilibrium. It cannot distinguish the aspirational reading from the operational reading. It holds the organisation to the stated values and notices when specific decisions or practices do not match. The noticing is not an AI bug; it is the mechanism working as advertised. But the organisation has to be prepared for what it produces.
How to prepare
Three things help.
Leadership commitment to acting on what surfaces. Not “we will consider whatever comes up” but specifically “we have identified the values we most want to hold, and we are prepared to change practice where the gap is material.” The commitment has to exist before the work starts, because once the gaps are named the question of whether to address them becomes political.
Choice about which values to expose first. Not every stated value needs to be held to account by the AI at once. Starting with the values the organisation is most confident in its ability to live out, and extending as confidence grows, is a defensible sequencing. Starting with the values that are mostly aspirational is a recipe for a finding list no one wants to receive.
Framing the findings as intelligence, not indictment. An authenticity gap is a signal of where the organisation can improve, not a verdict on its legitimacy. Leadership that receives the findings in the first framing can act on them; leadership that receives them in the second will defend against them, and the initiative will stall.
The heuristic sits alongside AI as an operational interpreter of purpose, vision and values: the two describe the same mechanism from different angles. The pattern describes the possibility; the heuristic describes what the possibility reliably produces along the way.