An AI Workforce Readiness Maturity Model for organisations preparing for AI adoption
By Chisom Obiudo
A two-hour AI training session can show an employee how to use an AI tool. It cannot tell staff what the organisation permits, provide an approved tool, fix disorganised information, identify suitable workflows, or equip managers to supervise AI-assisted work. Training alone is therefore insufficient to prepare a workforce for AI.
For a Namibian company considering or beginning to adopt AI, the better question is not how quickly licences can be rolled out. It is whether the organisation has created the conditions employees need to use AI safely and productively. Wanting employees to use AI is not the same as preparing them to use it well.
AI Workforce Readiness Maturity Model
An AI Workforce Readiness Maturity Model is a practical diagnostic tool for boards, executives, HR managers and governance professionals to use before and during AI adoption. It assesses six conditions that affect employees’ ability to use AI successfully and responsibly: policy and guardrails; access to appropriate, approved AI tools; management support; data readiness; workflow suitability; and training, time and opportunities to experiment.
These conditions should be assessed separately because strength in one area does not compensate for weakness in another. For example, providing employees with an approved AI tool will not resolve unclear policies, weak management support or unreliable information.
Four levels of workforce readiness
The model uses four maturity levels to indicate how prepared an organisation is to support responsible and productive AI use. An organisation may exhibit characteristics of more than one level, but the levels provide a useful starting point for assessing its overall position.
Level 1: Interested but Unprepared
Leadership wants to introduce AI, but the basic conditions are not yet in place. There may be no clear rules, approved tools, agreed use cases, or guidance for managers and employees. The priority is to establish these basics before encouraging wider use.
Level 2: Preparing
The organisation is moving from interest to deliberate preparation. Policies and guardrails are being developed, tools are being assessed, suitable use cases are being identified, and managers and employees are beginning to receive guidance and training. Initial AI use is limited to selected, controlled pilots with human review.
Level 3: Enabled
Employees have approved tools, clear rules, access to the information they need, and suitable workflows. Managers can guide AI use, and employees understand what they may do, what they must check, and when they should seek help. The organisation can now expand AI use carefully and measure results.
Level 4: Embedded
Responsible AI use has become part of everyday work. The organisation monitors use, measures whether AI is improving the work and learns from problems. Policies, controls, training and workflows are updated as needed, while uses that do not deliver value or create unacceptable risk are changed or stopped.
Six conditions that determine whether staff are using AI responsibly
The four levels show the organisation’s overall position. The following six dimensions show what is driving that position and where action is required:
- Policy and guardrails
Employees need clear rules on permitted uses, information handling, checking AI output, and reporting problems. If an employee has to ask, ‘Can I paste this client document into the chatbot?’, the rules are not clear enough. The policy should answer that question before the employee opens the tool. Leadership should provide clear rules with practical examples and a named contact for uncertain cases.
2. Access to appropriate and approved AI tools
Clear rules are not enough if employees lack an approved tool. Training staff and leaving them to use personal accounts or unapproved public tools encourages inconsistent practice and ‘shadow AI’, meaning AI use outside approved organisational channels. Before expanding AI use, leadership should decide which tools employees may use, for which tasks, who receives access, and where support will come from.
3. Management support
Approved tools alone are insufficient. Managers translate policy into day-to-day decisions. A manager who tells employees to ‘use AI to be more productive’ but cannot identify suitable tasks, required human checks, or situations where AI should not be used leaves employees to guess. Managers should be able to guide appropriate use, review AI-assisted work, and escalate concerns.
4. Data readiness
Data readiness encompasses the documents, records, procedures and other information employees need to do their work. If current procedures coexist with outdated versions or key documents are scattered across inboxes, employees may unknowingly provide AI with outdated or inconsistent information, leading to unreliable results. Leadership should remove obsolete material, clarify ownership and ensure authorised employees can locate the information required for priority use cases.
5. Workflow suitability
Having reliable information does not mean every workflow is suitable for AI. A common mistake is to start with the tool and look for somewhere to use it. A stronger approach begins with a recurring task, defines what AI will do and what a person must review, then sets a measure of time or quality. Leadership should test a few well-defined uses before expanding them.
6. Training, time and opportunity to experiment
Even a suitable workflow will stall if employees have no time to learn. AI literacy develops through guided practice, not a single presentation. Employees need role-specific practice on low-risk tasks, protected time to experiment, and a safe place to ask questions and receive feedback on their work. Leadership should then check whether employees can apply the rules while meeting normal quality standards.
A Namibian workplace scenario: when training is not the first fix
Imagine a Namibian organisation that wants employees to start using AI to improve productivity. Management plans to purchase AI licences and arrange a two-hour training session. Before doing so, it applies the maturity model.
The assessment finds no acceptable-use policy for AI, no approved tools and no agreed use cases. Employees are unsure whether confidential information may be entered into public AI tools, managers cannot identify suitable uses, and important documents are poorly organised. The organisation therefore largely sits at Level 1.
The immediate priority is not additional training. Management should first establish guardrails, approve an appropriate tool, identify a few suitable use cases, organise the information required for those tasks, and prepare managers to guide employees. Training can then focus on the tools, rules and workflows employees will use.
Questions for the board and executive team
The scenario illustrates why access to an AI tool should not be mistaken for workforce readiness. Boards and executives should therefore ask:
Are we preparing our people to use AI, or are we simply giving them access to an AI tool?
Which readiness gaps are preventing employees from using AI safely and effectively?
If we introduced an approved AI tool tomorrow, would employees know where it may be used, what information may be entered, which outputs must be checked, and when it should not be used?
Build readiness before you demand adoption
Low AI adoption does not necessarily indicate employee resistance or a need for more training. It may instead signal an organisational readiness gap. If rules are unclear, tools are unapproved, managers cannot guide AI use, information is unreliable, or suitable workflows have not been identified, asking employees to ‘use more AI’ will not address the underlying problem.
Leadership should first address the conditions that prevent employees from using AI safely and productively. Training becomes valuable when employees have clear rules, appropriate tools, management support, reliable information and suitable work on which to apply what they have learned.
For an organisation beginning its AI journey, the priority should not be how quickly it can get employees to use AI. The better starting point is to determine what must be put in place so employees can use AI effectively.
Chisom Obiudo is an admitted legal practitioner of the High Court of Namibia specialising in corporate governance and AI governance. She facilitates AI governance training for boards and delivers AI professional skills training. She also serves on the National Artificial Intelligence Technical Advisory Committee established by the National Commission on Research, Science and Technology (NCRST), with a focus on law and governance. She can be contacted at chisomokafor11@gmail.com
