Which AI Search Prompts Should Higher-ed Marketers Track?

Summary

  • Choosing which prompts to track is one of the most common challenges higher education marketers face when monitoring AI search.
  • The Gander Prompt Pyramid applies Bloom’s Taxonomy to six types of cognitive tasks: Knowledge, Understand, Apply, Analyze, Evaluate, and Create.
  • Each level represents a different way prospective students use AI, from retrieving institutional information to evaluating programs and creating consideration sets.
  • These levels are not sequential. Students may ask questions from several levels throughout their research and decision-making.
  • Tracking prompts across the Pyramid helps institutions identify where they appear in AI answers, how competitors are represented, and whether information about their programs is accurate.
  • The framework provides a structured approach to developing representative prompt sets, identifying content gaps, and taking action to improve AI search visibility.

Introducing the Gander Prompt Pyramid

“What prompts do we need to track?” is the most common question we hear when onboarding a new institution.

Gander provides a comprehensive set of prompts covering how prospective students research programs, compare schools, evaluate their options, and make decisions. Understanding why these prompts matter, and what they reveal about student behaviour, can help you make better use of the data.

I first explored applying Bloom’s Taxonomy to AI search in a November 2025 article about B2B prompt tracking. I’ve since developed that approach into the Gander Prompt Pyramid, a framework for classifying the cognitive tasks behind the questions asked of AI search.

The Pyramid has six levels: Knowledge, Understand, Apply, Analyze, Evaluate, and Create. It can also inform top-of-funnel (ToFu), middle-of-funnel (MoFu), and bottom-of-funnel (BoFu) planning.

Students don’t necessarily progress through these levels sequentially. Someone comparing universities might ask a factual question about admission requirements, followed by a request for a recommendation. The levels describe what they’re asking AI to do, rather than their enrolment status.

We’ll use two prospective students as examples throughout this article:

  1. A high school senior from Michigan researching engineering programs, including co-op opportunities, admission requirements, and in-state versus out-of-state costs.
  2. An international student in Nairobi considering undergraduate business and finance programs in the US and Canada, based on student visa requirements, tuition, scholarships, internships, and career opportunities.

These are illustrative scenarios based on common research needs, rather than transcripts of actual student conversations.

Knowledge

Stage: Predominantly ToFu

Knowledge is about retrieving information. Students ask AI to identify institutions, programs, requirements, fees, dates, or other facts that are relevant to their research. They may already have a clear understanding of what they want. They’re using AI to find the information needed to investigate their options.

Example prompts

Michigander:

  • Which universities in Michigan and neighbouring states offer mechanical engineering with co-op opportunities?
  • What are the published admission requirements and application deadlines for engineering at Michigan Tech, Kettering, and Purdue?

International student in Nairobi:

  • Which universities in the U.S. and Canada offer undergraduate finance programs and merit scholarships for international applicants?
  • What are the international tuition fees and application deadlines for undergraduate business programs at Indiana University’s Kelley School of Business and UBC Sauder?

How to Apply to GEO

Monitor whether your institution appears in relevant lists, whether its program attributes are correct, and which sources AI cites for fees, admission requirements, and deadlines.

An institution may rank well for its own name while being absent from unbranded questions about the programs it offers. Your content should make important institutional information easy to find and verify, including program availability, admission requirements, scholarships, and tuition. For international students, consider publishing content around ancillary requirements.

2. Understand / Comprehension

Stage: ToFu to early MoFu

Understand is where students investigate how programs, policies, and opportunities work. They might know that a university offers co-op but want to understand how its schedule affects graduation. Similarly, they may have identified several scholarships but need to understand their renewal conditions.

Example prompts

Michigander:

  • How do mandatory and optional engineering co-op programs differ in work terms and graduation timelines?
  • What’s the difference between being admitted directly into an engineering major and entering a general first-year engineering program?

International student in Nairobi:

  • Explain the process for obtaining an F-1 student visa to study in the U.S. and a study permit to study in Canada.
  • How do renewable merit scholarships affect total tuition over four years compared with one-time awards?

How to Apply to GEO

Examine whether AI explains your program structures, financial commitments, and eligibility requirements accurately. This includes identifying where it confuses your policies with those of competitors or overlooks important exceptions. Institutions should publish explanations that address how programs operate and what their requirements mean for students. Marketing descriptions alone may not provide enough detail to answer these questions.

3. Apply / Application

Stage: Predominantly MoFu

Apply is where students use institutional information to assess their own circumstances. They introduce constraints such as academic performance, financial resources, housing preferences, and program availability. These details can substantially change the relevance of a particular school.

Example prompts

Michigander:

  • I have a 3.6 GPA. Which engineering programs have admission requirements I should review, and which math and science prerequisites should I check?
  • As a Michigan resident, how would my estimated costs compare between attending Michigan Tech and attending Purdue as an out-of-state student?

International student in Nairobi:

  • My family can contribute USD $40,000 annually. Based on published costs and available scholarships at Indiana University and UBC, how much additional funding would I need for each?
  • As a Kenyan citizen applying to universities in the U.S. and Canada, which visa or study permit requirements would apply to my circumstances?

How to Apply to GEO

Monitor whether AI applies the correct fees, eligibility rules, admission requirements, and financial information to the student’s circumstances. Pay particular attention to assumptions. Published admissions statistics are not guaranteed cut-offs, and AI-generated admission probabilities (i.e. you have an 80% chance of acceptance) should not be presented as reliable predictions.

For international students, questions about immigration and employment authorization require additional scrutiny. Requirements differ between countries and change over time. Information should be checked against official guidance from U.S. Citizenship and Immigration Services and Immigration, Refugees and Citizenship Canada.

4. Analyze

Stage: MoFu to early BoFu

Analyze is where students break information into components, investigate relationships, and examine differences between institutions. Your school is more likely to appear alongside named competitors, particularly when students are comparing academic programs, co-op structures, financial commitments, or student experiences.

Example prompts

Michigander:

  • Compare how Michigan Tech, Kettering, and Purdue structure engineering co-op and work-integrated learning opportunities.
  • How do their first-year engineering curricula differ, and when do students choose a specialization?

International student in Nairobi:

  • Compare the undergraduate finance curricula, internship opportunities, and career services at Indiana University’s Kelley School of Business and UBC Sauder.
  • How do the U.S. F-1 student visa and Canada’s study permit differ in their rules for internships and post-graduation employment?

How to Apply to GEO

Examine which differences AI identifies, what it overlooks, and whether its comparisons use equivalent measures. For example, graduate employment statistics may cover different student populations, reporting periods, or definitions of employment. Publishing detailed, comparable information about your institution helps address inaccuracies in how programs and outcomes are represented.

5. Evaluate / Evaluation

Stage: Late MoFu to BoFu

Evaluate is where students make judgments based on evidence and their own priorities. While Analyze examines differences, Evaluate uses those differences to determine which option is preferable. This is also where reputation, student experiences, independent reviews, and third-party assessments may influence AI recommendations.

Example prompts

Michigander:

  • Given my interest in co-op, affordability, and campus experience, which of Michigan Tech, Kettering, and Purdue is the strongest fit? What are the trade-offs?
  • How credible are the graduate employment claims these universities make, based on their published methodologies?

International student in Nairobi:

  • Is the higher cost of attending a U.S. business school justified by the available evidence about program quality, internship opportunities, and graduate outcomes compared with Canadian options?
  • Considering tuition, scholarships, and post-graduation work authorization, would Indiana University or UBC offer better value for an international student interested in finance?

How to Apply to GEO

Examine how AI evaluates your institution, the evidence it relies on, and how competitors are represented. An unfavourable answer isn’t necessarily inaccurate. It may reflect a legitimate concern about tuition, housing, academic outcomes, or student experience. Equally, positive claims need to be supported by credible evidence.

In Niche’s October 2026 survey of more than 1,200 college-bound U.S. high school seniors, only 18% said they trusted AI a lot. The credibility of the sources behind an AI answer matters, particularly when students are making consequential decisions.

Brand fidelity is particularly important when AI platforms evaluate institutions. A recommendation may be factually accurate while overlooking the attributes that distinguish your institution from its competitors. Examine whether AI understands your positioning, how it characterizes your strengths and weaknesses, and which sources influence those descriptions.

6. Create / Creation

Stage: Predominantly BoFu

Create is where students combine information to produce something useful. That artifact might be a consideration set, comparison table, budget, application timeline, or decision summary.

It doesn’t need to be complicated. A student asking ChatGPT to build a shortlist based on specific requirements is using AI to synthesize information into something they can act on.

Example prompts

Michigander:

  • Give me the five best engineering schools set based on my 3.6 GPA, co-op preferences, and budget. Include in-state and out-of-state options and link to their official program pages.
  • Create a comparison matrix for Michigan Tech, Kettering, and Purdue showing admission requirements, co-op opportunities, annual costs, and application deadlines.

International student in Nairobi:

  • Build a comparison matrix for Indiana University and UBC showing total costs over four years, scholarships, student immigration requirements, and post-graduation employment options. Include links to official sources.

How to Apply to GEO

Monitor whether your institution appears in AI-generated consideration sets and planning materials, which attributes are included, and what gets overlooked. An institution might appear in general recommendations yet disappear when a student asks AI to create a shortlist with specific financial, academic, or geographic constraints.

These prompts can reveal gaps in competitive positioning that aren’t apparent when monitoring general recommendations alone.

Using the Gander Prompt Pyramid for AI Search Monitoring

The Gander Prompt Pyramid provides a structured approach to developing a comprehensive set of prompts. By considering all six cognitive levels, your team can better understand how prospective students use AI throughout their research and decision-making. The emphasis placed on each level will depend on your programs, audiences, and recruitment priorities.

A practical approach:

  1. Identify your audiences. Undergraduate, postgraduate, domestic, international, and transfer students have different priorities. College and university applicants may also have distinct requirements.
  2. Research their questions. Use prospective student interviews, recruitment staff, admissions enquiries, site-search data, and competitive research to identify meaningful prompts.
  3. Classify prompts using the Pyramid. Include factual retrieval, explanations, practical applications, comparisons, evaluations, and requests to create artifacts.
  4. Include branded and unbranded prompts. Monitor questions naming your institution, comparing competitors, and describing requirements without mentioning a particular school.
  5. Examine the responses. Track institutional visibility, competitor mentions, cited sources, factual accuracy, and brand fidelity. Compare how AI platforms describe your institution against its stated positioning. Are its distinguishing attributes represented accurately, overlooked, or attributed to competitors?
  6. Take action and monitor changes. Correct inaccurate information, address content gaps, and examine the sources influencing AI responses. Continue monitoring to determine whether your institution’s representation changes.

Final Thoughts

For over two decades, digital marketers have relied on keyword research to understand search volume, phrasing, and intent. AI search introduces a different challenge: students can ask highly specific questions that combine their grades, budgets, preferences, and personal circumstances.

Some vendors offer estimates based on limited clickstream samples, traditional keyword data, and statistical modelling. These figures should be treated with caution, particularly when the underlying data and potential sampling biases aren’t disclosed. A precise-looking number doesn’t necessarily reflect actual demand. At the time of writing, there are no reliable sources of prompt volume data.

The Gander Prompt Pyramid provides a structured way to understand the cognitive work students are asking AI to perform.

For higher-ed marketers, that means developing prompts that examine whether their institution appears in relevant recommendations, how accurately its programs are explained, how it compares with competitors, and whether it’s included in the consideration sets students create. Those findings can inform content priorities, identify inaccuracies, and help institutions take action to improve their ranking and representation in AI search.

About the Author: Adam Malamis

Adam Malamis is Head of Product at Gander, where he leads development of the company's AI search intelligence platform for tracking brand visibility across generative engines, like ChatGPT, Gemini, and Perplexity.

With over 20 years designing digital products for regulated industries including healthcare and finance, he brings a focus on information accuracy and user-centered design to the emerging field of Generative Engine Optimization (GEO). Adam holds certifications in accessibility (CPACC) and UX management from Nielsen Norman Group. When he's not analyzing AI search patterns, he's usually experimenting in the kitchen, in the garden, or exploring landscapes with his camera.


Show leadership where your institution stands in AI search

Track the accuracy, visibility, and brand fidelity of what the most popular AI platforms say about your institution. Identify factual inaccuracies, see how closely AI answers reflect your institution's positioning, and understand how you compare with competing schools. Get the evidence you need to correct misinformation and take action to improve ranking.

SCHEDULE A CONSULT