Evaluate AR glasses against one real workplace task before considering a wider rollout. The useful question is whether the device helps the intended person access information or complete an action under realistic conditions. A compelling demonstration is not enough to answer that question.
Begin with a controlled, low-risk trial. Identify the task, the people who will use the equipment and the information they need. Then assess the device, application and operating arrangements together. Hardware specifications matter, but they do not describe the entire working experience.
Define what the glasses must actually do
Terms such as smart glasses, AR glasses and wearable display can describe different capabilities. Ask what the proposed device displays, how that display relates to the surroundings and what input methods it supports. Do not infer spatial tracking from the shape of the frame.
A task may require only a visible reference panel. Another may require information to remain associated with a specific object or location. Those are different requirements and should not be treated as interchangeable.
Describe the required behaviour without using a product name. For example, the user needs to consult a diagram while examining a safe tabletop model. This gives the evaluation a purpose and makes it possible to compare alternatives without being led entirely by a supplier's prepared demonstration.
Ask how the physical world is presented
Clarify whether the user sees the surroundings directly through a transparent optical system or through a camera-mediated display. Microsoft's mixed reality overview distinguishes see-through and opaque display approaches. Its historical device examples are not a current purchasing shortlist.
Do not treat either approach as an automatic guarantee of suitability. Examine the particular device and application. Ask how digital content behaves when the user moves, looks away or changes the viewing distance.
Where a task depends on accurate alignment, define the requirement and ask how it is validated. A display that appears approximately associated with an object should not be described as precise enough for every operational purpose. Keep the trial within the limits agreed by the responsible specialists.
Build a representative task, not an obstacle course
Choose a normal activity that exposes the important requirements. For a tabletop review, the participant might locate a component, consult its description and compare it with a physical reference. Avoid adding unrelated interactions simply to test every available feature.
Prepare the information as it would appear in use. A large demonstration label is not representative if the actual application will display denser instructions. The content should be realistic enough to reveal readability and navigation problems.
Record what a successful task looks like without inventing performance targets. The person should be able to find the right information and explain the intended result. If time or accuracy matters to the organisation, agree an appropriate measurement method before the trial rather than selecting a favourable measure afterwards.
Use a consistent observation sheet for each session. Record the device configuration, the task attempted, assistance provided and any interruption. Give participants space to describe a problem in their own words before assigning it to a category. This helps distinguish a hardware limitation from an unclear instruction or an unsuitable application layout, which may call for a different response.
Check readability in the intended setting
Use the lighting and working positions expected in the real environment where it is safe to do so. Observe whether the person can read the relevant information without repeatedly adjusting their posture or losing the object they are examining.
Review text, diagrams and indicators separately. A device may present a simple notification adequately while struggling to make a detailed drawing useful. The evaluation should reflect the information the task genuinely requires.
Do not rely solely on a recording or external monitor view of the display. Ask participants about their own experience and investigate differences. What a reviewer sees in a captured image may not answer how comfortably someone can read the material through the device.
Make input part of the trial
Determine how the user advances, selects, dismisses and corrects an action. Voice, touch, gesture and connected controls have different implications for the setting. Test the method the proposed application will actually use.
A voice demonstration in a quiet office does not establish suitability in a noisy shared room. A gesture sequence may be awkward when hands are occupied. Treat these observations as requirements to resolve rather than problems the user should simply overcome.
Provide an obvious way to stop or return to a known state. People should not be trapped in an unfamiliar interaction while a facilitator searches for the correct command. Recovery is part of usability, not an optional technical extra.
Include the people who will wear the equipment
Invite representative users and make participation voluntary within the organisation's appropriate process. Ask about fit, interaction and information clarity. Provide a conventional alternative so useful work does not depend on wearing the device.
The W3C XR Accessibility User Requirements identifies needs involving different forms of input and output. Use it to inform questions about access rather than claiming that one trial establishes accessibility for everyone.
Follow the manufacturer's usage and fitting guidance. Stop a session when a participant wants to stop or encounters a problem. Do not turn this editorial evaluation approach into a medical assessment or an instruction to tolerate discomfort for the sake of completing a test.
Evaluate the supporting equipment and services
Check whether the device depends on a phone, computer, network or particular account. Include those dependencies in the trial. A demonstration that quietly relies on a specialist workstation may not represent the arrangement available to each employee.
Ask who prepares the equipment, manages software updates and handles faults. Confirm what support and repair arrangements are actually available for the organisation's location. Do not infer local availability from an international product announcement.
Document power and storage requirements without improvising device care. Follow the relevant manufacturer guidance and organisational procedures. If equipment must be shared, agree how access, settings and appropriate care will be managed between users.
Compare against a practical alternative
Repeat the task using the existing method, such as a tablet, printed reference or ordinary computer display. The comparison should be fair: use equivalent information and avoid giving one method better instructions.
Observe what the glasses add and what they make harder. Keeping information in view may help a particular task, while setup or interaction may add friction. The outcome may justify a narrow use case rather than an organisation-wide programme.
Separate novelty from sustained usefulness. Positive first impressions are valuable feedback, but ask whether participants would choose the method again for the same work. Identify the conditions under which the benefit is clear and the conditions under which another approach remains preferable.
Write a decision that includes unresolved questions
Summarise the tested device and software version, participants, task, setting and observations. List blockers separately from improvements. Record what has not been evaluated, especially if the eventual use would involve different environments or more consequential work.
Possible decisions include proceeding to a broader trial, revising the application, trying a different device or retaining the existing method. A useful evaluation does not have to end in a purchase.
If the task requires a tailored application, discuss custom AR development with the requirements and trial findings. The contact page provides a route for that conversation. Start with the work and the people, then determine whether a wearable display is the right way to support them.