The big idea: Make sure you are solving the right problem, then find out what is already known, with AI as a partner and your own judgment in charge, before you ask anyone anything.
In 60 seconds
- Check your challenge for traps: a good problem that is not the right one, symptoms instead of causes, and nothing you can act on.
- Use AI as a partner in curiosity, then open and judge every source with six questions.
- The gaps in what is known become your interview questions.
Do this: name the symptom and its possible causes, run the prompt pack, vet every source and turn the gaps into the interview questions you bring to Class 4.
How framing the challenge works
Class 2 starts the Inspiration phase. Before you interview anyone, make sure you are chasing the right problem, then learn what is already known about it. In BUAD 317 you will not spend weeks writing the design question from scratch, but you will tweak it as you learn. Getting it right on the first try is rare, and that is part of the process.
Refresh from IB 407 first
This page does not repeat what you learned in London. Review the Four Types of Problems and the three framing tests, the How Might We formula and steps and the empathy continuum. Then use the BUAD 317 additions below.
Ask better questions: Berger’s sequence
Warren Berger, author of A More Beautiful Question, argues that curiosity is a skill you can practice. Great questioners move through three kinds of questions in order, and each one opens the next.
Why?
Why does this problem exist? Why has it lasted? Why do people put up with it?
BUAD 317: your secondary research and interviews.
What if?
What if we looked at it differently, or borrowed an idea from another field?
BUAD 317: interpretation and brainstorming.
How?
How could we try one of those ideas quickly and learn from it?
BUAD 317: your rapid prototype and feedback.
Deception, detours and dead ends
The Class 2 slides warn about three traps that waste a team’s semester.
A good problem, not the right one
Some problems are interesting and important, but they are not the problem that matters most to the people you serve.
Watch out: a problem your team likes because it points to an idea you already love.
Symptoms, not causes
Chasing symptoms leads to more problems. Generative research digs for the etiology: the root cause behind what you see.
Watch out: “few people use the service” is a symptom. Ask why, several times.
The actionable test
Start with challenges someone can act on. If nobody on your team or the client side can change anything about it, it is not a problem you can solve.
Watch out: “the economy” or “human nature” as the cause. Those are gravity problems.
Problem statement (from the textbook example)
University students need to save as much money as possible on textbooks, because textbooks are a big yearly expense and students have little spare income.
Two HMW questions it could lead to
How might we help students connect with each other to buy and sell textbooks at lower prices? How might we work with local bookstores so students can buy and sell low-cost textbooks on and off campus?
One problem statement can open several HMW questions. Each points research in a different direction, so pick the one your evidence supports. The tool below checks your challenge for all three traps.
Choosing your focus
The focus area test
A good focus area passes four checks. Real: sources and stakeholders show the need exists. Narrow: one group of people in one situation, small enough to study this semester. Open: the literature leaves a gap your research can fill. Yours: your team cares about it enough to keep going into senior year.
Too broad
“Improve healthcare in East Tennessee.” No one could research that in a semester, and every source is somewhat relevant.
Focused
“How family caregivers manage new medications in the first 48 hours after a parent leaves the hospital.” Clear people, situation and moment.
How secondary research with AI works
Secondary research, also called a review of the related literature, tells you what is already known about your challenge. AI can speed it up, but only if your team stays in charge of the questions, the sources and the judgment.
One side of the debate
Some researchers argue that reading the literature first narrows your view. You start seeing what others saw, and you lose the fresh, bottom-up look that generative research is for.
The other side, and our approach
Others say an early review shows the assumptions underneath a problem and convinces clients you know the topic. In BUAD 317 you do both: read early to aim your interviews, then let the interviews surprise you.
The course framework: AI as a partner in curiosity
Your syllabus frames AI as a partner that broadens and deepens secondary research, not a shortcut. People’s needs, wants, values and contexts stay at the center. Five steps, each with the kind of output it produces:
Scope the research question
Use AI to reframe the broad challenge into focused queries about people’s needs, contexts and experiences.
Produces: research queries from several perspectives (academic, industry, community).
Search and gather sources
Search academic literature, industry reports, case studies and the open web. Look past the dominant, institutional voices.
Produces: a curated list of annotated sources.
Synthesize insights
Cluster, summarize and compare findings. Turn complex data into plain insights, empathy maps or proto-personas that show the human stories behind the numbers.
Produces: a synthesis of themes, tensions and opportunities.
Communicate in a human-centered way
Turn the research into visuals, plain-language stories or other formats stakeholders can easily take in.
Produces: a short presentation or infographic.
Iterate and find the gaps
Ask what is contradictory, missing or unanswered. Use the gaps to design your interviews and observations.
Produces: a “next steps” plan linking secondary findings to primary research.
Transparency
Cite your sources and say when you used AI summaries.
Bias awareness
Ask whose voices the AI output amplifies and whose are missing.
Inclusion
Go looking for diverse and marginalized perspectives on purpose.
Collaboration
AI augments your judgment, empathy and creativity. It does not replace them.
The eight AI prompting moves (push it, give it a persona, add constraints, give examples and more) are on the IB 407 Frame the Challenge page. The GLS prompt guide below goes further.
A Perplexity workflow for human-centered secondary research
Perplexity answers with cited sources, which makes it useful for a first pass at the literature. The course guide uses four moves:
Frame human-centered queries
Ask about a specific group’s needs or problems, not just facts.
Discover the literature
Ask for reviews organized by theme, with frameworks and stakeholder concerns.
Synthesize and evaluate
Ask for conflicting findings, gaps and evidence behind each need.
Keep human judgment
Open the cited sources. AI tools can misread or invent findings.
| Research step | Example prompt |
|---|---|
| Literature review | “Give a literature review on inclusive design for public transit users.” |
| Gap identification | “What are the major unmet needs of older adults in digital banking studies?” |
| Methods synthesis | “Summarize user research methods most effective for remote populations.” |
| Stakeholder mapping | “Identify key stakeholder groups in school meal policy research.” |
Keep findings tied to real people’s needs and behaviors, ask for evidence instead of opinions, and combine stories and quotes with numbers. When the stakes are high or the topic is sensitive, back up AI results with your own database searches or the original reports.
Judge every source with six questions
Relevance
What problem did the study address? Does it match yours?
Author
Who performed the study, and what is their stake in the answer?
Source
Where was it published: a peer-reviewed journal, an agency, a vendor?
Method
How was it conducted? Data, interviews or opinion? How many people, and where?
Date
When was it conducted? Check closely for technology, prices and policies.
Context
What sources does it reference? A study that cites nothing is a red flag.
The nine-step literature reviewKeywords, databases, the library and the business librarian
- Identify keywords. The words you choose set your path. Add new ones as you learn the field’s vocabulary. Database subject headings in EBSCO, Business Source Complete, IBISWorld or APA PsycInfo suggest better terms.
- Identify sources. Decide which databases fit your challenge.
- Search the library. Learn the online catalog and government publications. Get to know the business librarian. If it takes you more than 15 or 20 minutes to find something, call in the pros.
- Use the library catalog. Find books and reports, not just articles.
- Browse the stacks. The books shelved next to a good one are often good too.
- Search computer databases. Use your keywords selectively, reformulate with the database’s headings and descriptors, follow links to related work, let the database build your citations, and create an account to save your work.
- Consult online resources. See the database list below.
- Prepare an annotated bibliography. Summarize and evaluate each source.
- Develop the report. Your synthesis feeds your Capstone presentation now and your technical report in senior year.
| Resource | Good for |
|---|---|
| Business Source Complete | Business journals and trade publications |
| IBISWorld | Industry research reports: market size, players, trends |
| Nexis Uni | News, company and legal sources |
| APA PsycInfo | Research on behavior, motivation and well-being |
| Dissertation abstracts | Recent graduate research, often on narrow topics |
| Google Scholar and Google Books | A wide first sweep, then find full text through the library |
| Consulting firms and research groups | Reports from firms such as McKinsey or Gallup |
| Professional and trade organizations | Practitioner data and standards in your field |
Your UT login opens these databases through UT Libraries (opens in a new tab). Find your subject librarian, including the business librarian (opens in a new tab).
From AI search to annotated bibliographyFive steps, plus how to annotate
An annotated bibliography lists your sources, each with a short description that summarizes it, judges how reliable it is and explains how it fits your challenge. The course guide turns AI-supported searching into one in five steps:
- Select and curate. Choose 10 to 15 strong sources: scholarly, industry, practitioner and community. Verify every AI-suggested source yourself: credibility, date and authorship.
- Summarize. AI may draft a summary, but rewrite it in your own words. Cover the purpose, key findings and relevance to your challenge.
- Add a human-centered reflection of 2 to 3 sentences: whose voices are represented or missing, how the source helps you empathize with stakeholders, and what design opportunities or tensions it raises.
- Format in APA style: full citation, summary, reflection.
- Synthesize across sources. Write a 1 to 2 page introduction on the patterns, contradictions and gaps across all your sources. AI can help cluster themes, but the interpretation is yours.
How to annotate as you read: read the abstract to judge relevance, skim for the main points, write the full reference, code it by theme, summarize the essentials, note your own reactions and any quotes worth keeping, then add it to your literature matrix.
Existing records, triangulation and the literature matrixMake your evidence trustworthy
Examining existing records. The Field Guide counts records as research too: archival documents, journals, maps, video and audio recordings, photos and film, transcripts and other artifacts. Your client’s own reports and data are often the richest source you will find.
Triangulation. A finding is stronger when three different kinds of evidence point the same way. A triangulation matrix lists your research questions down the side and at least three data sources across the top, so you can see which questions have enough support. Geoffrey Mills’s optional supplemental book shows the format.
| Research question | Source 1 | Source 2 | Source 3 |
|---|---|---|---|
| Why do students who need the pantry not visit? | Published studies on stigma | Pantry sign-in data | Student interviews |
| When could students actually get there? | Class and work schedule data | Observation at the pantry door | Advisor interviews |
Annotated bibliography vs. literature matrix. A bibliography is organized by source. A literature matrix is organized by theme: sources down the side, themes across the top. One glance shows which themes are well supported, which rest on one source and which are empty. Do both.
Validity, reliability, generalizability and ethicsThe supplemental slides, and where to learn them
The Class 2 supplemental slides also introduce how to judge the quality of field data: validity (are you capturing what you think you are?), reliability (would the same techniques give consistent data?), generalizability (inside the community you studied, and beyond it) and ethics (informed consent, freedom from harm, the IRB). You use these most once you start interviewing, so they are covered in depth on Feet on the Street Interviews: Consent, Ethics and Trustworthy Data.
Try it: Frame and Search Builder
Tool · BUAD 317
Frame and Search Builder
Check that you are chasing the right problem, get a ready-to-use AI research prompt pack, and vet every source before you cite it.
- 1Frame
- 2Prompt
- 3Vet
Your design challenge
All checks
Your AI research prompt pack
Put it to work: from secondary research to your interview questions
Bring a laptop to Class 2. Your secondary research does two jobs in BUAD 317. Early on, its gaps become the AI-generated interview questions you bring to Class 4 and aim your 10+ street interviews. At the end, it is the relevant and credible evidence behind your Capstone presentation.
Check the challenge for traps
Run your class design challenge through the builder above. Name the symptom, list possible root causes and confirm someone can act on it. Tweak the wording if it hides a solution.
Your deliverable: a tuned challenge and root-cause list at the top of your team doc, and your voice in the Class 2 discussion (Deliverable #1, class engagement).
Prompt, then open every source
Run the prompt pack in Perplexity or another AI tool. Open each cited source and run the six questions. Drop anything you cannot find. Add library database searches and ask the business librarian when you get stuck.
Your deliverable: a list of verified sources, each marked with its type, so you can see whether your evidence is triangulated.
Annotate and synthesize
Write each entry in your own words with a human-centered reflection, then plot sources by theme in a literature matrix. Note patterns, contradictions and gaps.
Your deliverable: annotated entries and a short synthesis of what is known and what is missing.
Turn the gaps into interview questions
Use the last prompt in the pack to draft questions with AI, then rewrite them yourself. Keep only open questions that ask people about their own lives. Keep the prompts you used and an AI acknowledgment.
Your deliverable: Deliverable #2, your own set of AI-generated interview questions, brought to Class 4 to discuss. They are judged on question clarity. See AI-Generated Interview Questions.
Carry the evidence forward
Your gaps tell your interview team where to look in Deliverable #3, Part A. Later, the strongest sources become part of your research overview, along with your plan to find more evidence.
Your deliverable: the research overview in Deliverable #4, your 15 to 20 minute Capstone presentation in Class 14.
Example: one annotated entryA made-up entry for a city permit office challenge
This entry is invented to show the format. It is not a real source.
Citation: Author, A., and Author, B. (Year). Title of the article. Journal Name, volume(issue), pages.
Summary: The authors compare cities that moved permit applications online. Processing times dropped, but first-time applicants still called the office often, mostly to ask which documents applied to them.
Human-centered reflection: The voices here are city staff and experienced contractors. First-time small business owners are missing, and they are our focus. The source suggests the real pain may be knowing what applies, not filling out the form, so we will ask owners to walk us through their first application.
Saying how you used AIWhat the syllabus asks you to disclose
When AI meaningfully contributes to your work, the syllabus asks you to:
- Name the tool you used and describe how it helped.
- Share the detailed prompts you used.
- Explain how you refined the output through your own research and analysis.
- Never enter confidential client, company or personal information into a public AI tool.
You are fully responsible for the accuracy of anything you submit. The syllabus lists fabricating sources as an inappropriate use of AI, and improper AI use is treated the same as plagiarism.
Do this, not thatHabits that make secondary research useful
Do this
- Name the symptom, then ask why until you reach possible causes.
- Open every source yourself before you cite it.
- Search the fields of every major on your team, plus one analogous setting.
- Find each key finding in three kinds of source.
- Call the business librarian after 15 or 20 minutes stuck.
- Treat a gap as an interview question, not a failure.
Not that
- Researching a solution you already picked.
- Pasting an AI literature summary as your review.
- Citing a source you only saw in a chatbot answer.
- Ten vendor blogs and no research, or ten studies and no real voices.
- Blaming a cause nobody can change.
- Hunting only for sources that agree with you.
Before you submitCheck your framing, sources and questions
- [ ] Our challenge is a question with no solution baked in.
- [ ] We named the symptom and at least two possible root causes.
- [ ] Someone on our team or the client side can act on the problem.
- [ ] We opened every source and ran the six questions on it.
- [ ] Our key findings appear in at least three kinds of source.
- [ ] Each annotation is in our own words, with a human-centered reflection.
- [ ] We saved our prompts and wrote how we used AI.
- [ ] Every gap became an interview question or something to observe.
Go deeper
Class 2 readings and tools
Related topics
Back to all BUAD 317 topics · Next: Field Research Methods: Do, Say, Make
In-class team activity · 25 min
Root Cause Relay
By the end: A root-cause chain for your class challenge, a set of AI-cited sources your team actually opened and judged, and two interview questions built from the biggest gap, checked by another team. Your instructor will give you the password in class.
Part of BUAD 317, taught by Lane Morris, PhD and Jules Morris, PhD, in the Smith GLS human-centered leadership program, toward the Human Centered Consulting Certificate.
Frequently asked questions
How do you know you are solving the right problem?
Check for three traps. First, a really good problem that is not the right problem for the people you serve. Second, chasing symptoms instead of root causes, so name what you noticed and ask why until you reach possible causes. Third, a problem nobody can act on. If no one on your team or the client side controls anything about it, it is not a problem you can solve.
How can students use AI for secondary research responsibly?
Treat AI as a partner in curiosity, not a shortcut. Use it to scope questions, search, synthesize, communicate findings and find gaps. Follow four guidelines: transparency about AI use, awareness of whose voices are missing, inclusion of diverse perspectives, and collaboration where AI augments your judgment. Open every source it cites, because AI tools can misread findings or invent sources that do not exist.
How do you evaluate a source for a literature review?
Ask six questions. Relevance: what problem did the study address? Author: who performed it? Source: where was it published? Method: how was it conducted? Date: when was it conducted? Context: what sources does it reference? Then ask whose voices are represented or missing. A source that only appears in an AI answer does not count until you find and open the original.
What is a triangulation matrix in research?
A triangulation matrix lists your research questions down the side and at least three data sources across the top, such as published studies, existing records and interviews. It shows which questions have support from several kinds of evidence and which rest on one source. When three different kinds of sources point the same way, you can trust the finding more.
How do you turn AI secondary research into an annotated bibliography?
Pick 10 to 15 strong sources and verify each one yourself. Summarize each in your own words: purpose, key findings and relevance. Add a two to three sentence human-centered reflection on whose voices are present or missing and what design opportunities appear. Format entries in APA style, then write a one to two page synthesis of patterns, contradictions and gaps across all sources.