Use these templates to learn how clients and referrers recognize a need, assess professional firms, and decide whom to contact. Turn their questions into useful information and a defined set of AI monitoring cases.
Connect the research to the approved scope in A.2 and proposed work in A.5. Replace brackets with actual entries. All sector examples are hypothetical. Preserve participants’ wording separately from researcher interpretations and illustrative prompts.
B.1 Research brief and participant plan
Complete this brief before recruiting. Choose a specific buying or referral decision that the research will inform.
Field | Entry to complete |
Research control | [Program ID; research round; owner; dates; approved scope reference] |
Decision | We need to decide [content, service explanation, or research priority] for [audience] considering [service]. |
Learning questions | [What triggers the search? Which uncertainties affect selection? What evidence is missing?] |
Boundaries | [Service; industry; market; jurisdiction; language; decision period to discuss] |
Participants | [Relevant roles and experience; recruitment routes; target interviews by group; selection rationale] |
Delivery | [Interview guide version; interviewer; note-taker; analysis owner; hours; findings review date] |
Information handling | [Participation and recording permissions; permitted uses; access; approved transcription tools; retention/deletion date; contact for withdrawal requests] |
Recruit across recent clients, prospects who chose another provider, people who deferred or handled the work internally, and relevant referrers. Include decision-makers and influential evaluators where their perspectives differ. Record inaccessible groups and recruitment through existing relationships.
Coverage register: [Group/role; target; invited; completed; recruitment route; missing perspective; next action/owner].
Treat the interview round as evidence about these participants. Do not infer market percentages or search volumes from interview counts. Plan another round when an important role or unresolved decision needs further investigation.
Explain participation, recording, intended use, access, and retention before obtaining agreement. Record what each participant permits. Use the firm’s approved research information and consent process.
Avoid collecting client identities, privileged matter details, confidential financial records, or proprietary designs when the decision process can be discussed without them. Keep contact details separate from coded research notes. Permission for an interview does not establish permission to publish an attributed quotation.
B.2 Client and prospective client interview guide
Allow approximately 30–40 minutes, adjusting to the participant. Use open questions about a specific experience, with follow-ups that clarify what happened. Pilot the guide and revise unclear wording. This approach follows GOV.UK’s guidance on in-depth interviews.
Interview record: [Participant ID; role/segment; date; interviewer; guide version; permissions record; decision discussed; notes location].
Opening script
“We are researching how people choose [service] so we can improve the information we provide. This is a research conversation. Please discuss the decision without identifying confidential clients or matters. You may skip questions or stop. As explained in [information sheet], we will use your responses for [purpose]. May we proceed?”
Confirm recording permission separately before recording. Introduce observers and any transcription service.
Core questions
- Context: Tell me about a recent occasion when you considered outside help with [need]. What was your role?
- Trigger: What happened that made the issue worth addressing? What outcome were you seeking?
- Alternatives: What options did you consider, including doing the work internally or postponing it?
- Discovery: Where did you look, or whom did you ask, first? What happened next?
- Language: What words did you use to describe the need? Which terms were unclear?
- Evaluation: How did you decide which options or providers deserved further attention? What information changed your view?
- Evidence: If you considered outside providers, what did you check about experience, proposed people, scope, cost, or delivery? What remained uncertain?
- Decision: Who else influenced the choice? What led to appointment, rejection, or delay?
- Reflection: What would have helped you make that decision more confidently? What have I missed?
Ask one question at a time. Useful probes include “Can you describe an example?” and “What did you do after that?” Capture the participant’s experience before showing proposed content or suggesting criteria.
AI use follow-up
After the unaided account, ask: “Did any AI assistant or AI-generated search answer play a part?”
If yes, ask which product or experience, when it was used, what they asked, what they checked, and whether it introduced a firm or helped assess one already known. Ask what action followed. If not, record that response and continue.
Invite voluntary sharing of a redacted prompt or answer only when permitted. Label wording as copied, recalled, or paraphrased. Keep first discovery, later evaluation, and reported influence distinct; the interview alone does not establish that AI caused an engagement.
Optional sector probes
Sector | Probe after the open account |
Consulting | Where were responsibilities for diagnosis, recommendations, and implementation unclear? What comparable experience mattered? |
Accounting | Which reporting, bookkeeping, tax, or assurance boundaries needed explanation? What did you need to know about onboarding? |
Law | What did you need to confirm about practice experience, jurisdictions, responsible lawyers, and initial contact? |
Engineering consulting | Which project phase and disciplines were involved? How did site constraints, proposed personnel, and responsibilities across the project team affect selection? |
B.3 Referrer interview guide
Use a recent referral or a decision to withhold one. Reuse B.2’s interview record, introduction, permissions, and information handling. Allow approximately 20–30 minutes.
- What is your relationship to the people who ask you for recommendations?
- Tell me about a recent request for help with [need]. How was the problem described?
- What made you decide to refer, withhold a referral, or suggest another course? How could you handle yourself?
- How did you identify possible firms or professionals? Which relationships or information sources mattered?
- What did you need to verify before introducing someone? What would have ruled a provider out?
- If you recommended someone, how did you explain the firm’s suitability and limitations?
- If an introduction proceeded, what information did the receiving firm need to assess fit? How did you arrange it and manage expectations?
- What happened afterward? What would make you refer again, choose someone else, or decline?
After the unaided account, use B.2’s AI follow-up where relevant. Ask what information the referrer checked personally and what they passed to the prospective client.
Sector follow-ups: For consulting, explore capability fit and implementation capacity; for accounting, service boundaries and existing adviser relationships; for law, jurisdiction and how the receiving firm handles initial checks; for engineering, discipline, project phase, delivery location, and interfaces with other specialists.
Referral record: [Participant ID; need/trigger; selection criteria; evidence checked; exclusions; information passed; introduction route; outcome; unresolved question].
Keep referral decisions identifiable in analysis. Do not assume that referrers use the same language, evidence, or selection process as buyers.
B.4 Question map
Create one record for each distinct decision question. Preserve differences in audience, market, or service when they change the answer needed. Link related records into a question family.
Question record
Field | Entry to complete |
Identity and context | [Question ID; client/referrer role; service; market; language; buying situation] |
Decision stage | [Recognize need / define scope / discover providers / evaluate / initiate contact / refer] |
Recorded wording | [Exact question or labeled paraphrase; original language; separately labeled translation] |
Evidence | [Participant/source IDs; dates; quote location; permission limits; copied/recalled/paraphrased] |
Support and uncertainty | [Distinct participants raising it; participants asked; prompted or spontaneous; conflicting accounts; gaps] |
Underlying need | The person needs to understand [information] before deciding [action]. |
Information response | [Existing asset or gap; evidence needed; professional reviewer; appropriate contact route] |
Action and monitoring | [Priority/reason; A.5 item; owner/date; proposed prompt family; monitoring status/version] |
Label internally proposed questions hypotheses until supported by research. Label CRM notes and colleagues’ reports by their actual source. One person repeating a concern remains one participant; several accounts of one engagement should remain linked.
Illustrative question map entries
These entries demonstrate the mapping method; they are hypotheses, not interview findings.
Audience and stage | Question | Information to prepare or verify |
Consulting buyer evaluating | Who will implement the recommendations? | Delivery responsibilities, handover, and available implementation support. |
Accounting buyer defining scope | Does monthly reporting include an audit opinion? | Clear engagement boundaries and explanation of separately scoped services. |
Law referrer checking fit | Can this team support a matter involving [jurisdictions]? | Responsible professionals, verified coverage, and an appropriate inquiry route. |
Engineering buyer defining scope | What happens after the feasibility study? | Study deliverables, limitations, decisions enabled, and separately commissioned later phases. |
Prioritize according to decision importance, business relevance, available evidence, and the consequences of misunderstanding. Repetition helps identify patterns; a consequential question can deserve action even when raised once.
B.5 Sample prompt families
Use the families below to draft cases appropriate to the approved audience and service. A family group questions serving the same decision. These examples are starting points; interview evidence should determine the final wording.
Replace brackets with nonconfidential details. Record whether each case preserves observed wording or is a researcher-designed diagnostic. Keep the natural question when studying buyer behavior. If you add requests for citations or uncertainty, save that version as a separate diagnostic case.
Consulting
Family | Illustrative prompt |
Recognizing need | “Our acquired businesses use different processes and accountability is unclear. What kinds of consulting support should we consider?” |
Defining scope | “What should an operating-model engagement deliver, and what responsibilities would remain with our leadership team?” |
Discovering providers | “Which consulting firms work with [company type] on post-acquisition operating models in [market]?” |
Evaluating providers | “What should we ask shortlisted consultants to establish who will do the work and how relevant their experience is?” |
Making a referral | “I advise a business seeking operating-model support. What should I clarify before introducing a consulting firm?” |
Checking a named firm | “What does [firm] offer for operating-model work in [market], and what is publicly documented about its implementation role?” |
Accounting
Family | Illustrative prompt |
Recognizing need | “Our finance team cannot produce timely monthly management reports. What types of external accounting support could we explore?” |
Defining scope | “What should we clarify when comparing outsourced bookkeeping, monthly management accounting, and a part-time finance director service?” |
Discovering providers | “Which accounting firms offer outsourced management accounting for [industry/company type] in [market]?” |
Evaluating providers | “What questions should we ask about the proposed team, systems, reporting timetable, onboarding, fees, and excluded services?” |
Making a referral | “A business owner needs monthly accounting support and already has a tax adviser. What should I clarify before making an introduction?” |
Checking a named firm | “What does [firm] publicly describe as included in its outsourced accounting service, and what would require direct confirmation?” |
Law
Family | Illustrative prompt |
Recognizing need | “Our business is planning an employment policy review involving [jurisdictions]. What specialist legal support should we look for?” |
Defining scope | “What should we clarify when commissioning an employment policy review, including covered jurisdictions, deliverables, and implementation support?” |
Discovering providers | “Which law firms describe an employment advisory practice serving [client type] in [jurisdiction]?” |
Evaluating providers | “What should we ask shortlisted employment lawyers about relevant experience, staffing, fees, and coordination across jurisdictions?” |
Making a referral | “I want to introduce a business seeking employment advice. What should I confirm about the receiving firm’s practice and initial contact process?” |
Checking a named firm | “What do public sources establish about [firm]’s employment practice and jurisdictional coverage, and which details need confirmation?” |
Engineering consulting
Family | Illustrative prompt |
Recognizing need | “We are planning changes to an occupied [facility type]. Which engineering disciplines should we discuss engaging at the outset?” |
Defining scope | “What should an owner clarify when commissioning a [discipline] feasibility study, including assumptions, deliverables, exclusions, and later design work?” |
Discovering providers | “Which engineering consultancies describe relevant [project type] experience and services in [location]?” |
Evaluating providers | “What evidence should we request about comparable projects, proposed engineers, relevant qualifications, and experience working around operating facilities?” |
Making a referral | “I am referring to an owner seeking [engineering service]. What nonconfidential project information would help a consultancy assess the assignment?” |
Checking a named firm | “What do public sources establish about [firm]’s [discipline] services, relevant projects, and delivery locations?” |
These are questions about selecting and understanding professional services. Confirm actual service capability, professional coverage, and project suitability with the responsible professionals.
Prompt case register
Record: [Case ID/version; question-map ID; family; audience; exact prompt; origin; language/location; AI product and mode; model if known; account/memory context; search setting; standalone or full conversation sequence; collection date; planned repetitions; reviewer; collection-specification and evaluation-rule references].
To extend a family, vary a meaningful feature such as buyer role, industry, location, or project phase. Give each variant its own ID and rationale. Multiple paraphrases do not establish additional client demand.
Resolve placeholders before collection. Start standalone cases in fresh conversations. Preserve the complete context for follow-up sequences. Keep named-firm checks separate from unbranded discovery cases; supplying the firm’s name changes what the observation can establish.
Retain prompts, answers, cited sources, timestamps, collection failures, and review decisions. Record material changes to wording or collection conditions and preserve the previous version. Report new question groups separately before combining them with established monitoring.
B.6 Findings and handoff
Use this form after reviewing interview notes and question maps.
Field | Entry to complete |
Finding | [Decision question; affected audience/service; what participants reported] |
Evidence and limits | [Source IDs; distinct participants and coverage; contrary accounts; recall limits; unresolved assumptions] |
Interpretation | [What the evidence suggests; plausible alternative explanations; what remains unknown] |
Proposed action | [Correct information / improve an asset / create material / investigate / defer; reason] |
Delivery | [A.5 item; owner; professional reviewer; supporting evidence; due date; completion criteria] |
Follow-up | [Monitoring cases; further interviews or other checks; owner; review date] |
For example, if participants describe confusion about engineering feasibility and detailed design, verify the firm’s actual service boundaries before revising its page. Keep the research finding, approved publication, and later AI observations as separately evidenced records.
Review findings with the practice and intake owners. Preserve disagreements rather than forcing a single journey. Route consequential factual errors promptly and place other proposed work in A.5.
Before handoff, confirm that every proposed prompt has an origin and audience, every finding has traceable evidence, and every selected action has an owner. Return to research when the service, market, referral route, or important buying questions change.
