Skip to main content

Notifications

Power Platform Community / Blogs / Power Automate Community Blog / 𝐔𝐬𝐞𝐂𝐚𝐬𝐞 𝟏: PAD + GP...

𝐔𝐬𝐞𝐂𝐚𝐬𝐞 𝟏: PAD + GPT: 𝗧𝗶𝘁𝗹𝗲: 𝖤𝗆𝗈𝗍𝗂𝗈𝗇 𝖺𝗇𝖽 𝖲𝖾𝗇𝗍𝗂𝗆𝖾𝗇𝗍 𝖠𝗇𝖺𝗅𝗒𝗌𝗂𝗌 𝗎𝗌𝗂𝗇𝗀 𝖦𝖯𝖳

VJR Profile Picture VJR 7,635

𝐔𝐬𝐞𝐂𝐚𝐬𝐞 𝟏: 𝐏𝐨𝐰𝐞𝐫 𝐀𝐮𝐭𝐨𝐦𝐚𝐭𝐞 𝐃𝐞𝐬𝐤𝐭𝐨𝐩 + 𝐆𝐏𝐓

𝗧𝗶𝘁𝗹𝗲: 𝖤𝗆𝗈𝗍𝗂𝗈𝗇 𝖺𝗇𝖽 𝖲𝖾𝗇𝗍𝗂𝗆𝖾𝗇𝗍 𝖠𝗇𝖺𝗅𝗒𝗌𝗂𝗌 𝗎𝗌𝗂𝗇𝗀 𝖦𝖯𝖳

 

VJR_1-1702547513591.png

 

 


🔹Use it with the "Create text with GPT" action of Power Automate Desktop

Or

🔹Call GPT OpenAI API from Power Automate Desktop

https://powerusers.microsoft.com/t5/Power-Automate-Cookbook/Integrating-ChatGPT-with-Power-Automate-Desktop/td-p/1980676

 


𝗪𝗼𝗿𝗸𝗳𝗹𝗼𝘄 𝗼𝗳 𝗥𝗣𝗔 𝗯𝗼𝘁: Example 1

𝟭. 𝗦𝗼𝘂𝗿𝗰𝗲: Incoming email body / Ticket description / Any field on any application readable by RPA

𝟮. 𝗥𝗣𝗔 𝗯𝗼𝘁: Power Automate Desktop (PAD) reads the above sources using API or via screen scraping.

𝟯. 𝗠𝗲𝗻𝘁𝗶𝗼𝗻𝗲𝗱 𝗯𝘆 𝗮 𝗵𝘂𝗺𝗮𝗻 𝗶𝗻 𝗮𝗻𝘆 𝗼𝗳 𝘁𝗵𝗲 𝗮𝗯𝗼𝘃𝗲 𝘀𝗼𝘂𝗿𝗰𝗲𝘀:
"𝘛𝘩𝘦 𝘱𝘳𝘰𝘥𝘶𝘤𝘵 𝘐 𝘳𝘦𝘤𝘦𝘪𝘷𝘦𝘥 𝘸𝘢𝘴 𝘥𝘦𝘧𝘦𝘤𝘵𝘪𝘷𝘦 𝘢𝘯𝘥 𝘵𝘩𝘦 𝘤𝘶𝘴𝘵𝘰𝘮𝘦𝘳 𝘴𝘦𝘳𝘷𝘪𝘤𝘦 𝘵𝘦𝘢𝘮 𝘸𝘢𝘴 𝘶𝘯𝘳𝘦𝘴𝘱𝘰𝘯𝘴𝘪𝘷𝘦 𝘢𝘯𝘥 𝘶𝘯𝘩𝘦𝘭𝘱𝘧𝘶𝘭 𝘸𝘩𝘦𝘯 𝘐 𝘳𝘦𝘢𝘤𝘩𝘦𝘥 𝘰𝘶𝘵 𝘧𝘰𝘳 𝘢𝘴𝘴𝘪𝘴𝘵𝘢𝘯𝘤𝘦. 𝘐 𝘢𝘮 𝘦𝘹𝘵𝘳𝘦𝘮𝘦𝘭𝘺 𝘥𝘪𝘴𝘴𝘢𝘵𝘪𝘴𝘧𝘪𝘦𝘥 𝘸𝘪𝘵𝘩 𝘮𝘺 𝘦𝘹𝘱𝘦𝘳𝘪𝘦��𝘤𝘦 𝘢𝘯𝘥 𝘸𝘪𝘭𝘭 𝘯𝘰𝘵 𝘣𝘦 𝘱𝘶𝘳𝘤𝘩𝘢𝘴𝘪𝘯𝘨 𝘧𝘳𝘰𝘮 𝘵𝘩𝘪𝘴 𝘤𝘰𝘮𝘱𝘢𝘯𝘺 𝘢𝘨𝘢𝘪𝘯."

𝟰. 𝗥𝗣𝗔 𝗯𝗼𝘁:
Sends the above text to ChatGPT with a question suffixed as:
[Can you give me the emotion and sentiment analysis of the above text each in one word?]

𝟱. 𝗥𝗲𝘀𝗽𝗼𝗻𝘀𝗲 𝗳𝗿𝗼𝗺 𝗖𝗵𝗮𝘁𝗚𝗣𝗧:
Emotion: Frustration
Sentiment: Negative

𝗛𝘂𝗺𝗮𝗻 𝘁𝗮𝗸𝗲𝘀 𝗽𝗼𝘀𝘀𝗶𝗯𝗹𝗲 𝗙𝗼𝗹𝗹𝗼𝘄-𝘂𝗽 𝗮𝗰𝘁𝗶𝗼𝗻𝘀:
🔹Give a call to the sender
🔹Elevate the customer service
🔹‘Up’ the priority of the ticket (𝗥𝗣𝗔 can do this)
🔹Expedite the resolution of the sender’s issue (𝗥𝗣𝗔 can assist on this)


𝗞𝗲𝘆 𝘁𝗶𝗽:
When you pass this from RPA to ChatGPT, it is 𝘃𝗶𝘁𝗮𝗹 to mention as “𝘦𝘢𝘤𝘩 𝘪𝘯 𝘰𝘯𝘦 𝘸𝘰𝘳𝘥”, else it returns a long answer.
[Can you give me the emotion and sentiment analysis of the above text 𝗲𝗮𝗰𝗵 𝗶𝗻 𝗼𝗻𝗲 𝘄𝗼𝗿𝗱?]


Example 2:
𝗜𝗻𝗽𝘂𝘁 𝘀𝗲𝗻𝘁 𝗯𝘆 𝗥𝗣𝗔:
"𝘐 𝘢𝘮 𝘦𝘹𝘵𝘳𝘦𝘮𝘦𝘭𝘺 𝘴𝘢𝘵𝘪𝘴𝘧𝘪𝘦𝘥 𝘸𝘪𝘵𝘩 𝘵𝘩𝘦 𝘴𝘦𝘳𝘷𝘪𝘤𝘦 𝘐 𝘳𝘦𝘤𝘦𝘪𝘷𝘦𝘥. 𝘛𝘩𝘦 𝘤𝘶𝘴𝘵𝘰𝘮𝘦𝘳 𝘴𝘦𝘳𝘷𝘪𝘤𝘦 𝘳𝘦𝘱𝘳𝘦𝘴𝘦𝘯𝘵𝘢𝘵𝘪��𝘦 𝘸𝘢𝘴 𝘷𝘦𝘳𝘺 𝘩𝘦𝘭𝘱𝘧𝘶𝘭 𝘢𝘯𝘥 𝘳𝘦𝘴𝘰𝘭𝘷𝘦𝘥 𝘮𝘺 𝘪𝘴𝘴𝘶𝘦 𝘦𝘧𝘧𝘪𝘤𝘪𝘦𝘯𝘵𝘭𝘺. 𝘐 𝘸𝘪𝘭𝘭 𝘥𝘦𝘧𝘪𝘯𝘪𝘵𝘦𝘭𝘺 𝘣𝘦 𝘳𝘦𝘵𝘶𝘳𝘯𝘪𝘯𝘨 𝘢𝘴 𝘢 𝘤𝘶𝘴𝘵𝘰𝘮𝘦𝘳."

𝗢𝘂𝘁𝗽𝘂𝘁 𝗳𝗿𝗼𝗺 𝗖𝗵𝗮𝘁𝗚𝗣𝗧:
Emotion: Pleased
Sentiment: Positive


Example 3:
𝗜𝗻𝗽𝘂𝘁 𝘀𝗲𝗻𝘁 𝗯𝘆 𝗥𝗣𝗔:
"𝘛𝘩𝘦 𝘱𝘳𝘰𝘥𝘶𝘤𝘵 𝘮𝘦𝘵 𝘮𝘺 𝘦𝘹𝘱𝘦𝘤𝘵𝘢𝘵𝘪𝘰𝘯𝘴 𝘢𝘯𝘥 𝘵𝘩𝘦 𝘤𝘶𝘴𝘵𝘰𝘮𝘦𝘳 𝘴𝘦𝘳𝘷𝘪𝘤𝘦 𝘵𝘦𝘢𝘮 𝘸𝘢𝘴 𝘳𝘦𝘴𝘱𝘰𝘯𝘴𝘪𝘷𝘦 𝘸𝘩𝘦𝘯 𝘐 𝘳𝘦𝘢𝘤𝘩𝘦𝘥 𝘰𝘶𝘵 𝘧𝘰𝘳 𝘢𝘴𝘴𝘪𝘴𝘵𝘢𝘯𝘤𝘦. 𝘐 𝘢𝘮 𝘺𝘦𝘵 𝘵𝘰 𝘥𝘦𝘵𝘦𝘳𝘮𝘪𝘯𝘦 𝘢𝘯𝘥 𝘯𝘰𝘵 𝘺𝘦𝘵 𝘴𝘶𝘳𝘦 𝘢𝘣𝘰𝘶𝘵 𝘮𝘺 𝘦𝘹𝘱𝘦𝘳𝘪𝘦𝘯𝘤𝘦"

𝗢𝘂𝘁𝗽𝘂𝘁 𝗳𝗿𝗼𝗺 𝗖𝗵𝗮𝘁𝗚𝗣𝗧:
Emotion: Indifferent
Sentiment: Neutral


Final Words:

🔹These are just examples but the real idea is to use it in any industry where you pass a text to evaluate the emotions and sentiments behind it and take the necessary course of action.

Comments

*This post is locked for comments