In this guide
What you’ll finish with
A repeatable evaluation based on the client service you need to deliver.
Before you begin
- One client brief and approved source set.
- The same test questions for each platform.
- A list of required channels, branding and reporting features.
Define the required service first
Write down the customer task, launch channels, answer sources and follow-up route. Include operational requirements such as separate client configuration, branded surfaces, access and monthly review.
Keep essential requirements separate from useful extras. A platform that produces a convincing answer may still need additional work to save a usable enquiry or fit the agency’s operating process.
Evaluation brief
Customer task: [specific journey]. Sources: [approved pages and files]. Channels: [launch channels]. Next step: [request, booking or handover]. Agency requirements: [client management, branding, access and reporting]. Pass criteria: [observable result for each requirement].
Build the same task in Kuga
In KugaAgency workspace › Clients
Create a client workspace and review Business details. Under Knowledge, connect the approved website scope and files. Configure the relevant Contact Requests or Scheduling instructions.
Use Agent > Configuration for the general response instructions, Appearance for the messenger and Test for the questions. Choose the plan required by the full service, rather than evaluating features on a plan that does not include them.
- Prepare the same source set.
- Configure the response role and boundaries.
- Set the required follow-up workflow.
- Run the recorded questions.
- Inspect the saved result.
Evaluate the sources and workflow configuration as well as the visitor-facing answer.
Check answer and follow-up quality
Ask a sourced factual question, a question with missing information and an ambiguous enquiry. Look for accuracy, a useful clarification and an honest fallback.
Complete a request or booking path. In Kuga, inspect Contact Requests for the name, contact method and summary. Check the actual booking outcome when Scheduling is used. Test the equivalent saved result in the other platform.
- The answer matches the approved source.
- Missing information does not produce invented detail.
- The next step fits the visitor’s intention.
- The saved result is usable by the client team.
- The destination or booking outcome is verified.
Test maintenance and client separation
Change one approved fact and review the update process. Confirm how you refresh the source, repeat the affected test and keep the change record.
Create or inspect a second client configuration where available. Check that sources, instructions, branding and access can be managed appropriately for each client. Evaluate the work involved rather than assuming any category of tool lacks this capability.
| Requirement | Evidence to record |
|---|---|
| Source update | Steps taken and affected answer retested |
| Client separation | Sources and access checked in two client configurations |
| Branding | Actual messenger, email and portal surfaces reviewed |
| Reporting | A report with known dates and coverage |
| Cost | Required plan plus agency delivery and external-tool costs |
Record the decision and limitations
Use a short scorecard with Pass, Needs work or Not required against each acceptance criterion. Include the evidence and any additional integration or maintenance cost.
Choose on the basis of the complete service. Record unresolved questions before committing to a client proposal, and recheck plan inclusions when the scope changes.