Customer surveys can be hit or miss. When you give someone a blank textbox to fill in, they might take a while to get to the point or never actually outright state how they feel about a product or service. If you need to tally up the responses from vague answers, you can use Power Automate AI survey synthesisto bring them in line with expected sentiments so they can be properly analyzed later. Here’s how.
Part 1: Creating a Spreadsheet that Holds Customer Survey Responses
For this part, there are a few different routes you can take depending on what form your customer surveys come in. You will usually need the following:
- Survey response ID: A unique identifier for each submission.
- Open-Text Survey Answers: One or more fields where respondents can freely enter text.
While not strictly necessary for the AI processing, adding other columns, such as the date of submission or more clarifying and identifying information would help you in your analysis (and sorting through the responses).
For demonstration purposes, we’ll use the following columns:
- Feedback ID
- Feedback (open text field)
- Sentiment (for result storage)
The Excel spreadsheet can be created manually for testing, or you can sync Microsoft Forms to Excel to make one, then just add a column to store the sentiment. Here’s an example of how your form might look like, but it can differ wildly depending on what you’re trying to do.
Part 2: Using Power Automate AI Survey Synthesis for Sentiment Generation
For the most basic part, you can use AI to condense long-form responses, such as the “Feedback” from the sample, into a single-word answer that’s much easier to incorporate into charts and reports.
Step 1. Go to Power Automate and select “Create.”
Step 2. Choose “Instant cloud flow” from the top row of options.
Step 3. Give the flow a name, check “Manually trigger a flow” and select “Create.”
Step 4. Click on “New step.”
Step 5. Search for “Get file content” and select “Get file content” from OneDrive for Business.
Step 6. You may need to log in with your Microsoft Business account, then find the file on your OneDrive.
Step 7. Click on “New Step,” then find “List rows present in a table.”
Step 8. Enter “OneDrive for Business” as your location, “OneDrive” as the drive, then your sheet file path. The “Table” should self-populate with the sheets in the folder, so choose the one option you have.
Step 9. Go to “New step” again, search for “Apply to each” and select the first option.
Step 10. Click on the “Select an output” textbox, then in “Dynamic content,” find and select “value.”
Step 11. Within the loop, add an action and find “Analyze positive or negative sentiment in a text” from the AI builder tools.
Step 12. Select the language of the context (usually English) and then the column that the text is stored in (“Feedback” in the sample).
Step 13. Add another action, select “Update a row” then select the same Excel file again. For the “Key Column” and key value, pick “Feedback ID” (as it’s the one that should hold unique IDs). Click on the “Sentiment” column text box and select “Sentences sentiment” from the Dynamic content selector.
Step 14. Hit “Save” and then click on “Test” in the top-right corner, which should automatically apply the flow to the sample page.
If you’ve properly linked the Excel file, you should get a complete Power Automate AI customer survey synthesis with a populated column with “positive,” “negative” or “neutral” sentiments. This might take a bit of time.
Additional Options
If you want to go deeper into analysis, you can combine custom GPT prompts with Power Automate and build a more complicated flow that takes into account multiple survey responses. This allows you to distill large amounts of text to single-word responses, but you’ll need to ensure that the AI actually outputs a legible single-word response by noting so in the prompt.
However, you may need to get acquainted with Power Automate more to get to this point.
Disclaimer: Some pages on this site may include an affiliate link. This does not effect our editorial in any way.
