AI Generation Problems
If an AI-generated ROI Model doesn't complete as expected, use the checks below to identify the most likely cause and decide what to do next.
Generation doesn't complete
When you create an ROI Model with AI, Ekphrastic displays a progress experience while the model is being generated.
Generation can take time to complete.
If generation is still in progress, keep the process open and allow Ekphrastic time to finish creating the model.
The progress stages indicate that generation is underway. They should not be interpreted as separate completed technical operations.
Check your AI usage
AI generation uses your account's available AI allowance.
If you don't have sufficient AI usage available, Ekphrastic can prevent additional AI generation.
Check the usage information displayed in your workspace to see whether your AI allowance has been reached.
Reaching your AI limit does not delete Models you've already created.
For more information, see Understanding Usage & Limits.
Check your request
If generation fails or the result isn't what you expected, review the description you provided.
A useful request should clearly explain:
The solution or initiative you're modelling
The business problem being addressed
Important volumes or operational measures
Current costs or performance where known
Expected improvements or benefits
The outcomes you want the model to calculate
You don't need to design every calculation yourself, but the request should provide enough business context for Ekphrastic to understand what you're trying to model.
Include known values
If important values are known, include them directly in your request.
For example:
Annual or monthly volumes
Employee numbers
Labour costs
Current processing times
Current performance rates
Expected improvements
Implementation or technology costs
Providing these values reduces the amount of business context Ekphrastic needs to establish from a more general description.
Generation completed, but the model needs changes
A successfully generated model may still need refinement.
This doesn't necessarily mean generation failed.
AI generation is designed to create the initial foundation of the business case. Review the generated:
Inputs
Assumptions
Calculations
Outputs
KPIs
Costs
Terminology
Then refine the model for the specific customer or opportunity.
Use ROI Studio when you need to change the model itself.
The generated assumptions aren't right
AI-generated assumptions are starting values.
If you have better customer-specific or business-specific information, use that information when refining the model.
Pay particular attention to assumptions that materially affect financial outputs such as:
Volumes
Costs
Adoption
Performance improvement
Time savings
Implementation costs
A model becomes more useful as generic assumptions are replaced with relevant business information.
The model is already published
Models created successfully with AI are automatically published when generation completes.
This is expected behaviour.
Publishing does not mean that the model has been sent to anyone, and it does not automatically create a share link.
If you don't want the model in its published state, use Unpublish.
For more information, see Publishing Content.
I can't find the generated model
After successful generation, the new model should appear in My ROI Models & Calculators.
If you can't immediately find it:
Return to My ROI Models & Calculators.
Clear any search term that may be restricting the results.
Check any active folder or tag filters.
Review the current sorting.
Search for the model again.
For more information about finding content, see Folders, Tags, Search & Sort.
Should I retry?
If generation clearly fails and no completed model is created, you can try the request again.
Before retrying:
Check your available AI usage.
Review the business-case description.
Add any important information that was missing.
Make the intended outcomes clear.
Avoid repeatedly submitting the same request while an earlier generation may still be processing.
The result is different from another generation
AI-generated models can vary between generations, even when the underlying business case is similar.
The generated model should therefore be treated as a starting foundation rather than a fixed template that must always produce exactly the same structure.
Review each generated model based on whether its assumptions, calculations and outputs appropriately represent the business case you're trying to build.
When to create manually instead
You don't have to use AI for every model.
Consider Creating a Model Manually when:
You already know the exact structure you want.
You're building from a specific modelling methodology.
You don't need AI to create the initial foundation.
Manual creation starts with a blank model and does not use AI to generate its initial structure.
Quick checks
If AI generation isn't working as expected, check:
Generation won't start
Available AI usage
Generation is still running
Allow the current process time to complete
Generation fails
Review the request and retry
Result lacks important context
Add known business values and requirements
Assumptions aren't appropriate
Refine them in ROI Studio
Model can't be found
Search, folders, tags and sorting
Model is unexpectedly published
AI-generated models publish automatically after successful generation
Still having problems?
If you've checked your AI allowance, reviewed your request and retried generation but the problem continues, contact Ekphrastic support with as much relevant information as possible.
Include:
What you were trying to create
What happened
Any message displayed by Ekphrastic
Whether the problem occurs again when you retry
Do not include confidential customer information unless it is necessary and appropriate to share for support purposes.
Last updated