1. Operational need
The assessment should begin with workflows, delays, errors, knowledge gaps, service requirements and decisions. A vague goal such as “use AI” is not a use case.
A useful opportunity statement identifies who performs the work, what information is required, where friction occurs and what a better outcome would look like.
2. Information and data
The organization needs to know what information exists, where it is stored, who owns it and whether it is current enough to support the proposed system.
For knowledge applications, poor documents can create poor answers. For analytics or automation, incomplete records can distort decisions.
3. Technology environment
Readiness includes identity systems, networks, hardware, cloud services, integration points, access patterns and internal support capacity.
The assessment should not assume that a new AI platform can be added without affecting the rest of the environment.
4. Security, privacy and risk
The assessment should identify sensitive information, allowed uses, data-location expectations, access roles, logging, retention, recovery and human-review requirements.
Higher-risk uses require stronger controls and may not be appropriate for an early pilot.
5. People and work practices
Employees need to understand the purpose of the system, how their work will change and where their responsibility remains. Leadership needs to assign ownership for decisions, content, approvals and support.
AI literacy and change readiness are implementation requirements, not optional extras.
6. Governance
A useful assessment clarifies who can approve use cases, which uses are prohibited, how output is reviewed, who owns knowledge bases and how incidents or unexpected behaviour will be escalated.
Governance should be proportionate to risk. It should not be so heavy that small, low-risk experiments become impossible.
7. Pilot selection
The first pilot should have a defined user group, manageable information, measurable value and a clear stop or review point.
A good pilot is not always the most ambitious idea. It is the use case that can produce credible learning with controlled risk.
8. Roadmap and commercial separation
The final output should be a vendor-neutral roadmap that prioritizes opportunities, gaps and next steps. If the assessor also offers implementation services, the implementation proposal should be separate and optional.
This protects the credibility of the assessment while giving the organization a practical path forward.