A short eligibility check can make an advertised service easier to understand, but only if its questions actually change the next step. A quiz that always recommends the same package is a sales presentation disguised as an assessment. A useful check makes its limits visible.
Consider a fictional data-migration service promoted through ChatGPT ads. The provider supports a defined set of source systems and a particular destination. Someone arriving from the ad needs to know whether their migration can be considered before describing the whole project. This is an interaction problem: how to ask, explain and route. The detailed reasons a service is unsuitable belong in the limitations guide.
Start with a decision tree on paper
List the few conditions that genuinely determine the route. For the migration example, the first might be the source system. A second might be whether the customer can supply an export in a supported format. A third could establish whether they want the destination the provider actually supports.
Remove questions that merely enrich a lead record. Company size may help later, but it should not appear as an eligibility criterion unless it truly changes the service offered. Likewise, do not ask about a budget to produce a technical compatibility result. Those are different decisions.
Draw every result before designing the form. A supported source with an available export can lead to a project enquiry. An unsupported source can lead to a clear explanation. An uncertain source can lead to a small review request. If a branch has no meaningful consequence, reconsider the question.
Let the visitor say they do not know
A customer may know the name of the software but not its version or export format. Requiring a confident answer can generate guesses that look like reliable qualification. Include an uncertainty option when uncertainty is plausible.
For example, “I can export a CSV file”, “I can export another format” and “I have not checked yet” tell the team different things. The third answer should not automatically mean the service is impossible. It can reveal the next useful instruction: where to find the export information or what to ask an administrator.
Keep explanations beside the question. W3C’s form instructions guidance covers required inputs, formats and instructions for individual controls, including making instructions available to assistive technology. In the migration example, explain “export format” as the type of file the programme can save, with an example beside the choices. Leave that explanation visible after the customer answers so they can consult it again.
Also distinguish “I do not know” from an unanswered question. Someone deliberately choosing uncertainty has supplied useful information. They should not receive the same error as someone who accidentally skipped a required question. If the check requires a choice for every question, explain that without treating the uncertainty option as an invalid answer.
Describe the result as precisely as the check allows
“Your project is approved” is too strong if the check has only established that the source system appears on a list. Prefer a result that names what is known: “This source can be considered for the migration service. We still need to review the export and requested scope.”
A result page can contain three compact elements:
- The answers that led to this route, with a way to change them.
- What those answers establish, expressed without a guarantee.
- The next piece of information or action needed.
If the result is unsupported, explain the relevant boundary. Do not display an unrelated offer as though it were the original service. A genuine alternative can be useful when clearly labelled. If there is no suitable alternative, a clear stop is more respectful of the customer’s time than another form.
Avoid an overall score when individual conditions determine whether the service can be delivered. Two positive answers do not compensate for an unsupported destination. In the fictional migration example, a known source and an unknown export can lead to “Check your export”, while a destination outside the service’s scope needs a different result even when the export is ready. Name the input that determines the next step. This makes the assessment understandable and gives the visitor a way to correct it.
What remains to assess the migration?
- Source, export and destination supported
Proceed to review of the file and project scope.
- Export is unknown
Show how to find the export information.
- Destination unsupported
Explain why this particular service does not fit.
Connect the check to the ChatGPT ad
The ad should not imply universal compatibility if the first destination question excludes many systems. Inspect the actual offer and the check together. A qualification process cannot repair an overbroad promise simply by revealing the restriction after the click.
The destination should introduce the service before launching the questions. A visitor should know what they are checking eligibility for, approximately what information they need and what the result will mean. Do not assume they remember every word of the ad. Do not ask them to paste a private ChatGPT conversation to obtain a service-fit answer. The inputs should concern their project.
If a supported result leads to an enquiry form, carry forward the answers in a visible, editable summary where appropriate. Asking the same questions again creates an opportunity for contradictory answers. The multistep form guide covers retaining context across a longer interaction.
Test the awkward cases first
Try a familiar system with an unknown version, a supported format with an unsupported destination and a visitor who changes an earlier answer after seeing the result. Check that the recommendation updates and that stale answers do not continue to influence it invisibly.
Changing the source system may also require a new answer about its export. Show which answers remain valid and which need checking again. Avoid clearing everything without explanation, but do not carry an export format from the previous system into the new assessment as though it had been confirmed. The visible summary and the information reaching the delivery team should describe the same project.
Ask the service team to review each outcome. They should be able to explain why the branch exists and what they would do with the resulting enquiry. Assign an owner to update the supported list when the service changes. A simple maintained check is preferable to an elaborate one whose rules have drifted away from delivery.
A suitable project should proceed with clearer information, an unsuitable one receive a truthful explanation, and an uncertain one reach a person or instruction capable of resolving the uncertainty. If the uncertainty concerns where the work will take place, use the separate service-area check. Choose the checks the advertised service actually needs and place them within the wider landing-page structure.
Sources and scope
The interaction that helps a ChatGPT ad visitor check service fit before enquiring, including uncertain answers and review routes.
Working methods and examples are editorial suggestions. Check current platform requirements and available features before implementation.