Licences deployed is not the same as AI adopted.

The problem

Your dashboards are telling you the wrong story.

Usage data shows two false positives and hides one real asset. People who log in to be seen logging in look like adopters. People quietly using AI outside sanctioned tools look like resisters. And the largest group, capable professionals who were trained, licensed and willing, appear simply as low usage with no explanation attached.

Most organisations respond with more training. Training closes a skills gap. It does not account for the professional who was trained, licensed and willing, and whose usage fell away anyway. Leaders consistently describe this group as the one they cannot explain. HumanOS looks at what usage data and training records leave out: how people read the arrival of AI against the work their professional standing rests on.

What we measure

The part readiness indices are not designed to measure.

HumanOS assesses perceived AI role encroachment: the degree to which people experience AI as absorbing the work that carries their professional identity. Around it we measure the identity needs that determine whether someone integrates a new tool or works around it, and the wider organisational conditions that amplify or contain the effect.

The instrument is built on Identity Process Theory, published research in social psychology that has been applied to health, migration and environmental behaviour, and is now being applied to AI at work. The diagnostic is fielded in Arabic or English.

What leaders receive

Distinct adoption patterns, not a single score.

The diagnostic sorts responses into distinct adoption patterns rather than a single score. Each pattern has a different cause and a different intervention, and treating them the same is why enablement budgets underperform.

These patterns are theoretical prototypes rather than empirically established categories. Current research is examining whether these profiles, or different patterns, emerge empirically.

Reports show distribution by team, function and seniority, and they report spread as well as averages. Two teams with identical average scores can need opposite interventions: one has a shared belief problem, the other is split down the middle. Averages hide that, so we never report averages alone.

The territory

AI changes more than tasks. It changes the organisation.

We focus on the organisational and psychological changes required when AI moves from a tool into the way work actually gets done. As AI takes on more cognitive work, adoption is only the first question. Behind it sit larger ones:

HumanOS studies and addresses the organisational consequences of AI, beginning where they first surface: in whether the work actually changes.

How we work

Founder-led, from scoping to findings.

Every engagement runs directly with the founder, not a delivery team. A scoping conversation agrees the population, language and reporting cuts. The diagnostic is fielded in Arabic or English. Findings are presented to your leadership in person: how adoption patterns distribute across your teams, the specific breakpoints, and an enablement plan sequenced against the 4E tracks, delivered by us or by your own L&D function. Individual responses never reach the employer, no team-level result is reported below ten respondents, and no subgroup below five.

HumanOS was founded by Iya Kobakhidze, an AI transformation consultant based in Dubai with a background spanning EY, INTERPOL and the EU institutions. She is completing graduate research in business psychology at Heriot-Watt University; HumanOS is her separate, independent commercial venture. The instrument draws on the same body of published identity research: grounded in theory rather than assembled from consulting heuristics.

Start the conversation

If your AI usage numbers have plateaued and nobody can tell you why, that is the conversation to have.