Power Platform Community Forum Thread Details

Hi everyone,

I'm facing an issue with response time when using AI Builder. Specifically, the response takes too long—sometimes over 30 seconds—which is not ideal for my use case.

Here’s my setup:

Prompt Type: I’m using the "Summarize Text" prompt in AI Hub

Input Source: Dataverse table (contains 213 records)
Output Format: JSON
Model Used: GPT-4o (standard); I also tested with basic GPT-4o, but the responses were not accurate.
Temperature: Set to 0
Records to Retrieve: Set to 250 (to ensure all 213 records are retrieved)
Columns: No filtering applied, as I want all columns included
Instructions: Custom instructions provided to fetch data from the Dataverse table and return only the required columns in a specified JSON format

Despite this configuration, the response time is consistently slow.

I’m wondering:

Is this due to resource allocation?
Can anything be optimized in terms of the prompt or configuration to speed up the response?
Has anyone else experienced this with GPT-4o in AI Builder?

Any insights or suggestions to improve the performance would be appreciated!