Autonomous Agent to perform QA

(0) Share

Report

Posted on by NP-02012003-0

Hi All,

I am exploring creation of a Copilot Studio Agent that can perform set of Quality assurance scripts on a fixed frequency basis. Please find below a brief summary of the context.

As a part of our QA process to validate adherence to internal policy we perform a set of scripts on a monthly basis. They fall into two major categories
1. Validating certain fields are accurately populated in our inventory, checking if an user is still active in the company etc.
2. Validate if the in scope teams have documented control procedures as required by policy
The major sources of information for this would be
1. Inventory Report
2. Policy documentation
3. Individual documentation by each of the teams
We have a set of scripts to verify adherence to each of the policy standard

This is what i could find out based on my readings about Copilot Agent.

Can access Sharepoint (this is where all the documentation will be stored)
Can be triggered based on event such as Recurrence
If required can access send out an email for human in loop scenarios.

I am looking for some guidance on the following

Is it possible to have the Agent perform set of tasks (basically execute a set of scripts i.e. prompts) and store the output of this whole exercise in a predefined excel sheet
How can agent access external systems for e.g. checking if a certain employee is still active, posting the results of the QA in an external system ( I think this will require MCP).

Thanks

Categories:

Autonomous agents

I have the same question (0)

All responses (2)

Answers (0)

Sort by

Suggested answer

Michael E. Gernaey 53,978 Moderator on at

Like (0)

Report
Copy link

Link copied!
Hello @NP-02012003-0

See inline

This is what i could find out based on my readings about Copilot Agent.

Can access Sharepoint (this is where all the documentation will be stored)

Can be triggered based on event such as Recurrence

If required can access send out an email for human in loop scenarios.

I am looking for some guidance on the following

Is it possible to have the Agent perform set of tasks (basically execute a set of scripts i.e. prompts) and store the output of this whole exercise in a predefined excel sheet
[Michael] Yes it is

How can agent access external systems for e.g. checking if a certain employee is still active, posting the results of the QA in an external system ( I think this will require MCP).
[Michael] No this does not require MCP, it requires connectors. It is possible that the external system HAS an MCP built for it, but if not, you need to use a Custom Connector or HTTP Connector to call API's it exposes.

If these suggestions help resolve your issue, Please consider Marking the answer as such and also maybe a like.

Thank you!
Sincerely, Michael Gernaey

1 people found this reply helpful.

Was this reply helpful? Yes No
NP-02012003-0 2 on at

Like (0)

Report
Copy link

Link copied!
Thanks @Michael E. Gernaey !

Are there any materials/guide that I can read up for both of these.

In my usecase the QA analyst today manually checks a data dump and documents to verify if they are in line with internal policy.

I am thinking this as a multi agent solution

1) Agent to procure the data

2) Agent to clean up the data

3) Agent to execute the checks

For Point 2 below

I thought MCP is required for an LLM to interact with system.

Is it possible to have the Agent perform set of tasks (basically execute a set of scripts i.e. prompts) and store the output of this whole exercise in a predefined excel sheet
[Michael] Yes it is

How can agent access external systems for e.g. checking if a certain employee is still active, posting the results of the QA in an external system ( I think this will require MCP).
[Michael] No this does not require MCP, it requires connectors. It is possible that the external system HAS an MCP built for it, but if not, you need to use a Custom Connector or HTTP Connector to call API's it exposes.

Was this reply helpful? Yes No