Why you should probably (not) let ChatGPT write your next strategic plan

As a higher education strategic planning facilitator, there are a few party tricks I like to use to kick off the brainstorming process. A common one is to present the planning committee with five different, real mission statements and challenge them to match each one with the institution it represents. Most groups fail to even recognize their own university’s mission statement among the indistinguishable verbiage frequently used by colleges and universities.

My new trick involves bringing a prepared strategic plan for their university to the first planning committee meeting. Most folks think it looks pretty good. But the catch is that I asked ChatGPT to write the plan. So, if the ChatGPT version is so close to what committees are going to end up with, why spend dozens of hours in meetings arguing over word choice and strategic pillars? Why not just start with this, make a few tweaks, and call it a day?

But, of course, it’s not that simple. The reason you shouldn’t use ChatGPT to write your plan lies in the difference between strategy and a strategic plan. Developing a distinctive strategy is not the same as writing a document your stakeholders will find reassuring.

“All Things to All People” is not a strategy, but it is the theme of almost every strategic plan

One of the primary goals of every higher education strategic planning process is to build consensus across a diverse community of stakeholders. All constituents – faculty from every discipline, staff from every function, students of all backgrounds, alumni of every generation, board members of every political background, community and corporate partners – should be able to see themselves in the final glossy plan. This miracle of consensus in a polarized world is achieved by stripping out any specific arguments that might seem to prioritize one group over another and by using increasingly abstract nouns (like “excellence,” “student-centeredness,” and “community engagement”) with which no one will disagree. These floating signifiers are unobjectionable because they don’t actually refer to anything. They may sound good, but they don’t help your organization decide which course of action to take or not to take.

How LLM’s are like town hall meetings

The process by which Large Language Models (LLMs) like ChatGPT create blocks of text is ironically quite similar to a university strategic planning committee. They analyze a large corpus of existing content and then predict the phrases that are most likely to come next. They are trained to please the reader by identifying the most common response with just enough randomization to make it slightly surprising. Plans written by ChatGPT (or Gemini or Claude or other LLMs) are “good” in the sense that they are unobjectionable, and an unobjectionable plan is almost always the most important criteria for most planning processes (implicitly if not explicitly).

A strategy is not the same thing as a “strategic plan”

A higher education strategy is not simply the statement of a goal or ambition. No one writes in their plan:

• “Invest in student amenities and country club atmosphere to appeal to wealthy applicants”
• “Recruit full pay international students to subsidize local students as public funding is reduced”
• “Increase undergraduate class sizes and the share of graduate student and adjunct instructors in order to free up more time for faculty research to improve our research reputation”

A strategy is not simply the statement of a goal or ambition. It is the description of a viable path to achieve that goal. The bulleted statements above are real strategies in the sense that they indicate an actual path from A to B. And looking at historical data you can see that many universities have indeed pursued these strategies even as their public plans talked about excellence, access, and engagement. However, these institutional strategies (like all strategies) require difficult tradeoffs and compromises. They sacrifice certain goals (like smaller classes and a greater share of tenure track instructors) for other goals (like greater research output).

Planning processes that engage large numbers of stakeholders are great at surfacing aspirations, but they are terrible at making tradeoffs. Like any representative democratic process, they frame policy decisions as power struggles between different constituencies. In the tenuous communities of today’s fractured campuses, it is simply not acceptable for any single group to win or lose. Every group must see the plan as benefiting their interests.

So, what’s the answer?

One lesson I have learned in working with universities is that the published plan might sound similar to many others to an outsider, but for the people who participated in creating it, it is deeply meaningful. Just like a student’s education, the written output is not the goal, it is the process of writing the plan, arguing about word choice, and negotiating tradeoffs to build a shared understanding. So with strategic planning, using AI as a substitute for thinking may get you a polished piece of prose, but it will not build the shared understanding necessary to keep your community together as you face challenging decisions.

There are five pieces of advice I would give to those struggling with their strategic planning process:

1. Recognize this tension from the beginning of your process. Discuss it openly. Explore how explicit you can be in prioritizing without losing key stakeholder groups.
2. Educate all of your stakeholders on the constraints and tradeoffs that you face. Make it clear that hard choices need to be made.
3. Focus on what is unique and distinctive about your institution. Request an interactive workshop on “Achieving Principled Differentiation” by contacting your Strategic Leader.
4. Engage the community in helping you to find potential paths through these constraints (rather than simply asking them about their aspirations).
5. Use the process to build a shared understanding of otherwise generic terms (e.g., “Excellence for our university means instructors bringing high impact practices into the classroom. It does not mean faculty researchers who publish in high impact journals but do very little teaching.”)

Remember, the published strategic plan document does a different job than an actual strategy. It is designed to align the interests of diverse stakeholders around a shared vision. These documents require significant effort to create, but make sure you don’t forget to build a real strategy at the same time. And ChatGPT is the last place that I would go to develop a truly differentiated university strategy.

For many higher education leaders, balancing the strategic planning process with actual institutional strategy can feel overwhelming. Our Strategic Advisory Services can help you navigate these situations with confidence. For more information on creating a differentiated strategy for your institution, visit our Dynamic Strategy Resource Center.

This blog post was NOT written by ChatGPT.