The potential of AI has sparked discussions about its role in shaping various aspects of employee engagement and management.
One area that often comes under scrutiny is the use of AI in crafting and conducting employee surveys. While technology undoubtedly offers new avenues for efficiency, it's important to carefully consider the implications of relying solely on AI for such a crucial task.
To put it to the test, our employee survey company used AI for a test employee survey. Here's what happened.
AI Written Employee Survey Example
Using any A.I. language model to write employee surveys for our clients at Drive Research is not something we would ever do, but with all the chatter about ChatGBT taking over the world, it had me thinking...
What kind of employee survey questions would AI design? Would they be comparable to the questions our employee survey company would recommend?
So, I typed the following query into ChatGPT-4:
Write an employee survey for our client who is a financial instituion on behalf of Drive Research
And, here’s what happened…
1. How satisfied are you with your current job role and responsibilities?
First off, this question is double-barrelled. This might sound picky, but survey writers should choose either “role” or “responsibilities” to avoid confusion/poor data quality.
Also, when starting an employee survey ask the broadest questions first and then get more specific as the survey progresses. Where to start will likely depend on the goals and objectives of the survey.
For example, consider starting with questions like employee net promoter score, overall satisfaction with the organization, or what matters most when determining job satisfaction.
Example of eNPS scores across different platforms with our Voice of Employee tool
2. Do you feel valued and appreciated by your colleagues and superiors?
After reading the issues with the first question, you will likely know the biggest issues here.
There are two areas where this question is double-barrelled. (Dare I say this question is quadruple-barrelled?)
The survey writer should choose either “valued” or “appreciated”, and “colleagues” or “superiors”. Leaving the question as-is will lead to employee confusion and poor-quality data.
3. How well do you believe the organization communicates its goals and objectives?
This question is at it again! The survey writer should choose the word “goals” or “objectives” to avoid the question being double-barrelled.
4. Are you provided with adequate resources and support to perform your job effectively?
Again, this question is double-barrelled. The survey writer should choose either “resources” or “support” to avoid confusion/poor data quality.
5. How satisfied are you with the work-life balance offered by the company?
This question is fine as-is, but when reading through the questions in this survey it’s clear several could be rephrased as agreement statements and asked in a grid format.
Reformatting the questions in a grid format allows for a better, more intuitive respondent experience knowing questions are phrased similarly (i.e., fewer scales to interpret).
6. Do you feel that your skills and talents are effectively utilized in your current position?
Again, this question is double-barreled. The writer should choose either “skills” or “talents” here to avoid miscommunication.
7. Are you given opportunities for growth and development within the organization?
I’m likely sounding like a broken record at this point. The survey writer should choose either the word “growth” or “development” to avoid the question being double-barrelled.
8. How would you rate the level of teamwork and collaboration within your department?
Again, this question is double-barrelled. Suggest choosing the word “teamwork” or “collaboration” to avoid misinterpreting the question.
9. Are you satisfied with the recognition and rewards system in place for employees?
Recognition and rewards are not the same thing! To avoid the question being double-barrelled the survey writer should either choose “recognition” or “rewards”.
Additionally, yes/no questions tend to be less insightful than ratings. Consider asking the level of satisfaction on a rating scale versus a simple yes/no. This provides similar results but with better insight!
10. Do you feel comfortable voicing your concerns and suggestions to your superiors?
You guessed it. The double-barrelled questions continue! The writer should choose the word “concerns” or “suggestions” here to avoid misinterpretation.
Also, this is another example of a yes/no question that is fine as-is but would gather more insight asked as a rating scale.
11. Are you satisfied with the level of feedback and performance evaluations you receive?
Again, “feedback” and “performance evaluations” are very different things. Feedback is more general and performance evaluations are a specific type of meeting. The survey writer should choose one term here.
Also, consider asking this as a rating scale for additional insight.
12. How well does the company promote a positive and inclusive work environment?
I love the idea behind this question, but again it’s double-barrelled. The writer needs to choose one term (i.e., positive or inclusive) to avoid misinterpretation/poor data quality.
13. Are you aware of the company's mission and how your role contributes to its success?
This is two completely different questions. It’s likely clear at this point that this current question phrased as a simple yes/no will lead to poor data quality.
Do respondents agree/disagree with both statements? Just one? We’d never know.
14. How satisfied are you with the overall leadership and management within the organization?
I suggest choosing just one term here too! Leadership is typically thought of as a level higher than just management. Avoid confusion and choose just one term.
15. Would you recommend the company as a good place to work to others?
This question is phrased similarly to employee net promoter score (eNPS), which is a commonly used employee engagement metric. Rephrasing this question to match that format allows the organization to compare its results to others.
In addition to asking eNPS, it’s really helpful to include a follow-up question asking why respondents choose their rating. Understanding what drives eNPS helps organizations understand how to improve their score.
More on employee net promoter scores in the video below!
Why You Shouldn't Use AI for Employee Surveys
Relying solely on AI to compose employee surveys may undermine the human connection and nuanced understanding that such surveys require. Employee engagement and satisfaction are deeply personal experiences, often necessitating empathetic and contextual questioning that AI might struggle to replicate accurately.
Additionally, AI-generated surveys could potentially lack the depth and specificity needed to address unique organizational dynamics, potentially leading to vague or irrelevant questions that fail to capture valuable insights.
Outside of survey design, AI cannot manage any other aspect of the employee survey process. It is important to utilize a professional third party for employee surveys because they...
- Bring an impartial perspective. This ensures that employee surveys are conducted without any internal biases. This promotes honest feedback and increases the likelihood of capturing genuine employee sentiments and concerns.
- Add anonymity and trust. Employees might hesitate to provide candid feedback if they believe their responses could be traced back to them. Third-party providers offer anonymity, fostering a trustful environment where employees feel comfortable sharing their thoughts without fear of reprisal.
- Provide expertise and best practices: Professional survey providers possess expertise in designing effective surveys and crafting questions that yield valuable insights. Their experience helps create surveys that are well-structured, focused, and aligned with industry best practices.
- Offer actionable recommendations for how to use the data: Third-party providers are skilled at collecting and analyzing data impartially, reducing the potential for data manipulation or misinterpretation. Their rigorous methodologies ensure that the collected data accurately represents employee sentiments.
In short, while AI can aid in survey creation, a wholly human touch is vital to crafting surveys that genuinely resonate with employees and yield meaningful results.
Mistakes to Avoid with AI Employee Surveys
Here are a few things to keep in mind when creating an employee survey, whether it be with AI tools or in-house:
- Several of these questions should be rephrased as statements to measure agreement. This will improve the experience for respondents (i.e., creates a better, more intuitive experience). This also allows the option to randomize the question order for the additional benefit of minimizing question order bias. Check out this example of an annual employee survey to learn more about ideal question flow and how the process works.
- Add more follow-up questions for additional context. In addition to considering agreement statements, the survey writer should also consider adding follow-up questions based on the goals and objectives of the survey. For example, if recognition is something the organization wants to improve on, ask a follow-up question asking respondents why they selected their rating or what improvements they would like to see made.
Drive Research is a full-service market research company, specializing in employee survey design, programming, fieldwork, and analysis. Our team has a combined 80+ years of experience in the market research space, learning and adapting to new techniques as we go.
Need help with your employee survey project? Contact our team today.
- Message us on our website
- Email us at [email protected]
- Call us at 888-725-DATA
- Text us at 315-303-2040
As a Research Manager, Emily is approaching a decade of experience in the market research industry and loves to challenge the status quo. She is a certified VoC professional with a passion for storytelling.
Learn more about Emily, here.