Global Advisors approached AI adoption as a change in how people work, learn and take responsibility. The firm gave consultants access to a shared conversational AI environment, began deploying a consultant-aligned agent for each person, connected those tools to permissioned internal knowledge, introduced a formal digital learning platform and revised its competency model to include AI-era expectations.

These elements were designed as one people system. Tools create opportunity, but working habits, judgement, learning pathways, role expectations and management practices determine whether that opportunity becomes useful capability.

Best pratice sharing in regular team updates are fundamental to raising common practice.

Maturity: everyday AI access, permissioned knowledge retrieval and consultant-aligned agent patterns are in active internal use. The formal learning platform and competency-management foundation are implemented, while unified learning pathways, live cross-system integration and sustained measurement of competency development remain incomplete.

Giving employees a general-purpose AI account is a quick way to begin experimenting. It does little on its own to change an organisation.

Consultants need to know which tools are appropriate, what information they may use, how to test generated work and when to rely on their own judgement. New joiners need a route into methods that were previously learned through repeated explanation and apprenticeship. Managers need to distinguish productive augmentation from polished but weak output. The firm must also decide whether AI fluency is optional enthusiasm or part of the expected standard of professional performance.

Global Advisors faced these questions while rebuilding much of its internal technology and knowledge environment. The technical programme created new possibilities, but it also made a human constraint visible: capability would remain uneven unless access, learning, role design and accountability changed together.

We developed a layered people model around the work itself.

The first layer gives every consultant practical access to AI. A shared conversational workbench provides a consistent entry point to approved models and specialist configurations. Permissioned retrieval lets a user ask questions of internal documents while preserving the access boundary that applies to the underlying material. Citations take the user back to the source rather than asking them to trust generated prose.

The second layer moves from generic assistance to personal working support. The firm began deploying consultant-aligned agents with narrower user scope, persistent context, reusable skills and controlled access to tools. Each agent is designed to support its consultant's work without becoming an independent holder of authority or an unrestricted copy of the firm's knowledge.

The third layer addresses development. A formal learning platform provides a home for structured courses, onboarding, assessment and competency-linked learning. A separate microlearning direction explores shorter, repeated practice for material that benefits from reinforcement. AI can help create and adapt learning content, subject to source, review and publication controls.

The fourth layer changes what the organisation expects of its people. AI fluency is incorporated into the competency discussion alongside structured thinking, numeracy, technical curiosity, writing quality, ownership and client judgement. Consultants are expected to use AI to improve the work while remaining able to explain, test and own the answer.

Structured learning and reusable onboarding

The learning platform was introduced to reduce dependence on repeated manual teaching. It provides a basis for standardised onboarding, course delivery, assessment, learning records and competency-linked development.

The wider design recognises that formal courses and daily practice are different needs. The formal platform remains the anchor for governed learning and progression. A developing microlearning surface is intended to support shorter, more frequent review. Early product work includes source ingestion, AI-assisted card generation and spaced repetition, but it is not yet a production-wide learning experience.

Revised competencies for AI-native work

The competency-management foundation now supports defined models, role levels, competency families, expectations and target ratings. Global Advisors has used that structure to revisit what good consulting looks like when AI is part of daily work.

The revised expectation is broader than prompt technique. People need to frame problems, interrogate evidence, work with data and models, understand the tools well enough to use them responsibly, communicate clearly and take ownership of the final recommendation. AI makes these existing qualities more visible and introduces new requirements around verification, context, privacy and tool selection.

This also changes progression. Junior staff need an intentionally designed apprenticeship because some of the tasks through which earlier generations learned may now be heavily assisted. More experienced consultants need to demonstrate that their judgement improves the generated work, not simply that they can produce it faster.

Human accountability in the workflow

The firm's emerging AI-enabled processes retain explicit review points. A consultant remains responsible for whether evidence is appropriate, an analysis is sound, a communication is fit for its audience and a recommendation can be defended.

That responsibility cannot be reduced to a warning beneath a chat box. It has to appear in competency expectations, workflow states, review practices, source links and the design of the agent's permissions.

Generated learning content needs its own controls

AI can accelerate course and exercise creation, but incorrect material can scale just as quickly. Learning content therefore needs an approved source chain, citations where appropriate, editorial review and a controlled publication step. The model may help draft or adapt material; it does not make that material authoritative.

AI changes the meaning of performance

Output volume becomes less informative when everyone can generate drafts quickly. Managers need to look more closely at problem framing, source discipline, challenge, judgement, learning speed and ownership. This is a harder management task than counting artefacts, but it is closer to the value clients expect from a consultant.

Uneven fluency can widen performance gaps

Strong performers often use AI to extend existing judgement, while weaker performers may use it to conceal gaps. Shared tools alone can therefore increase variation. Learning pathways, examples, review and explicit competencies are needed to help people develop rather than merely giving them access.

Leaders pursuing AI adoption should ask:

• Does every employee have a practical, approved route to use AI in everyday work?
• Can people reach institutional knowledge without bypassing the permissions on the source material?
• Which context belongs to an individual employee, and which methods and knowledge should be shared?
• How has the competency model changed to reflect verification, technical fluency and human accountability?
• What will junior employees learn when AI performs many traditional apprenticeship tasks?
• Are learning, performance, coaching and workflow systems reinforcing the same behaviours?
• How will the organisation distinguish increased output from improved judgement and client value?
• Which adoption claims are supported by observed behaviour rather than tool availability?

Becoming AI-native is a people transformation supported by technology. The operating model changes only when access, learning, standards, incentives and responsibility move together.