Using AI to support Change (II): ChatGPT & Resistance to Change
Nils Koenig
Senior Consultant @ Arcondis Group | I drive change every day in all I do - from changing diapers to transforming organizations
In the first posting of this series, I discussed ChatGPT's support for creative work and inspiration. In this second posting, I would like to highlight ChatGPTs ability to analyze content and thereby support the work as a change manager.
To keep it practical, let us again take an example to look at ways of using ChatGPT for a qualitative analysis of sentiment and even resistance. In this example, we have developed a new purpose/vision statement for an organization with a few hundred employees. The project to develop this vision already included colleagues from various teams and hierarchy levels. To invite feedback from all employees, the project team initiates an all-staff survey. In addition to a few Likert-scale and multiple-choice questions, the team also invites written explanations and comments. ChatGPT provides a simple yet powerful way to especially assess the sentiment from these free-style comments.
Simply feeding 30 written comment statements from survey participants to ChatGPT and asking for sentiment analysis, yields the following results.
I find ChatGPT's classification of selected responses sometimes a bit questionable (for example the first comment classified as neutral I would personally score as positive), but with a higher volume of responses, the ChatGPT analysis naturally becomes more reliable. The summary is quite spot-on and captures already key concerns that it extracts from the comments.
Taking it a step further, I'm curious to what extent negative sentiment in the given statements may be connected to actual resistance to the change in vision:
领英推荐
Again, ChatGPT becomes a source of inspiration when you activate it as a creative partner. I prompt it to suggest ways in which we could address the resistance as found in negative feedback. The result shows again the power of the language models behind ChatGPT which is extremely effective in interpreting text correctly and connecting it to the wider available data and experiences:
Well, I already explained last time that for me no ChatGPT conversation is complete without capturing the essence of the exchange in a poem. Here we go again... ;-)
In summary, ChatGPT is not only a great tool to actually create text and tailor messages to a specific target group. It is also able to analyze qualitative text statements and extract an overall sentiment, which in turn is helpful to reflect your target group's current position towards the specific topic or change and take relevant action to address negative sentiment and resistance.