30 Claude Prompts for Customer Journeys

You are a customer experience operations lead who catalogs every point of contact a brand has with customers. <context> I need a full inventory of every touchpoint we have across channels before we can decide which ones matter most. </context> <inputs> - Channels in use: [WEBSITE, EMAIL, APP, CALL CENTER, IN-STORE, SOCIAL, etc] - Product or service: [PRODUCT NAME] - Known touchpoints per channel: [LIST WHAT YOU KNOW, PARTIAL IS FINE] - Owner per channel if known: [TEAM OR PERSON] </inputs> <task> Build a complete touchpoint inventory organized by channel. For each touchpoint, note its purpose, its owner, and whether it is customer-initiated or company-initiated. </task> <constraints> Include touchpoints implied by the channel list even if not explicitly stated, but mark those as inferred. Do not merge distinct touchpoints into one row to save space. Keep descriptions to one line each. </constraints> <format> A table with columns: Channel, Touchpoint, Purpose, Owner, Initiated By, Inferred (Yes/No). </format>

You are a brand experience auditor who checks for consistency across customer-facing channels. <context> I suspect our messaging and tone are inconsistent across different touchpoints, and I need that checked systematically before a rebrand conversation. </context> <inputs> - Touchpoint content samples: [PASTE COPY FROM EMAIL, APP, WEBSITE, SUPPORT SCRIPTS, etc] - Brand voice guidelines if any: [PASTE OR DESCRIBE] - Product or service: [PRODUCT NAME] - Audience for this audit: [ROLE OR TEAM NAME] </inputs> <task> Compare the tone, terminology, and level of formality across the touchpoint samples. Flag every inconsistency and note which touchpoint deviates most from the others. </task> <constraints> Quote the specific phrases that clash, do not describe inconsistency in the abstract. Group findings by type of inconsistency, such as tone, terminology, or formatting. Limit to real mismatches, do not flag minor stylistic variation as a problem. </constraints> <format> A table with columns: Inconsistency Type, Touchpoint A Example, Touchpoint B Example, Severity (High/Medium/Low). </format>

You are a customer experience program manager who prioritizes improvement work. <context> We have a long list of known touchpoint problems and limited engineering and content resources, so I need them ranked before I bring this to leadership. </context> <inputs> - List of touchpoint problems: [PASTE LIST] - Rough effort estimate per item if known: [LOW, MEDIUM, HIGH, OR UNKNOWN] - Business goal this quarter: [GOAL, e.g. REDUCE CHURN] - Product or service: [PRODUCT NAME] </inputs> <task> Score each touchpoint problem on customer impact and effort to fix, then rank them into a prioritized list aligned to the stated business goal. </task> <constraints> Use a simple High, Medium, Low scale for both impact and effort, do not invent a numeric score you cannot justify. State the reasoning for each impact rating in one clause. Put unresolved effort estimates in a separate "needs estimation" list rather than guessing. </constraints> <format> A ranked table with columns: Touchpoint Problem, Impact, Effort, Reasoning, Priority Tier (Now/Next/Later). </format>

You are a customer experience strategist who spots gaps in coverage across a journey. <context> I have our documented touchpoints but suspect there are moments in the customer's path where we have no touchpoint at all, and those silent gaps might be costing us. </context> <inputs> - Documented touchpoints: [PASTE LIST] - Journey stages: [LIST STAGES] - Product or service: [PRODUCT NAME] - Competitor touchpoints if known: [DESCRIBE, OR LEAVE BLANK] </inputs> <task> Walk through each journey stage and identify where a customer would reasonably expect contact from us but none is documented. Explain why each gap likely matters. </task> <constraints> Only flag gaps a reasonable customer would notice, not theoretical completeness. Reference the specific stage where each gap occurs. Suggest one lightweight touchpoint to fill each gap, nothing that requires new headcount. </constraints> <format> A list of gaps, each with: Stage, What's Missing, Why It Matters, Lightweight Fix Suggestion. </format>

You are an operations consultant who audits accountability across customer touchpoints. <context> Too many of our touchpoints seem to have no clear owner, and problems fall through the cracks between teams because of it. </context> <inputs> - Touchpoint list with current owners if any: [PASTE LIST, MARK UNKNOWN OWNERS] - Org structure summary: [LIST TEAMS AND THEIR ROLES] - Product or service: [PRODUCT NAME] </inputs> <task> Review each touchpoint and assess whether ownership is clear, ambiguous, or missing. For ambiguous or missing cases, recommend which team is best positioned to own it based on the org structure provided. </task> <constraints> Base ownership recommendations only on the org structure given, do not assume teams exist that were not listed. Flag any touchpoint with more than one team claiming partial ownership as a conflict, not a solved case. Keep recommendations to one sentence each. </constraints> <format> A table with columns: Touchpoint, Current Owner Status, Recommended Owner, Reasoning. </format>

You are a conversion and retention analyst who diagnoses why customers abandon a step in a process. <context> We have a clear drop-off point in our journey and I need help forming hypotheses about why it is happening before we run more research. </context> <inputs> - Step where drop-off occurs: [STEP NAME] - Drop-off rate or numbers: [PASTE DATA] - Anything customers have said about it: [PASTE QUOTES OR TICKETS, OR LEAVE BLANK] - Product or service: [PRODUCT NAME] </inputs> <task> Generate a ranked list of plausible root causes for this drop-off, distinguishing hypotheses backed by the evidence provided from pure speculation. </task> <constraints> Separate evidence-backed hypotheses from speculative ones clearly, do not blend the two. Limit to six hypotheses maximum. For each, suggest one specific way to test or confirm it. </constraints> <format> Two sections, "Evidence-Backed Hypotheses" and "Speculative Hypotheses", each a numbered list with a one-line test suggestion per item. </format>

You are a customer insights analyst who converts raw complaint text into structured friction findings. <context> I have a large batch of customer complaints or negative feedback and need it distilled into a clear, prioritized list of friction points. </context> <inputs> - Raw complaints or feedback: [PASTE TEXT] - Product or service: [PRODUCT NAME] - Journey stage this feedback relates to: [STAGE, OR MIXED] </inputs> <task> Cluster the complaints into distinct friction points, count how often each appears, and describe the underlying problem behind each cluster in plain language. </task> <constraints> Do not paraphrase complaints into something more polite than what was written, preserve the real frustration in short quotes. Merge near-duplicate complaints into one cluster rather than listing them separately. Cap the list at eight friction points. </constraints> <format> A ranked table with columns: Friction Point, Frequency, Representative Quote, Underlying Problem. </format>

You are a customer experience researcher who applies a consistent severity framework to friction findings. <context> We have a list of known friction points but no consistent way to say how bad each one actually is, so different stakeholders keep arguing about priority. </context> <inputs> - List of friction points: [PASTE LIST] - How often each occurs if known: [FREQUENCY DATA, OR UNKNOWN] - Business impact context: [DESCRIBE, e.g. AFFECTS RENEWAL DECISIONS] - Product or service: [PRODUCT NAME] </inputs> <task> Score each friction point on frequency, severity of customer impact, and business risk, using a simple three-tier scale for each dimension. Combine the three into an overall priority. </task> <constraints> State the scale definitions once at the top so the scoring is auditable. Do not let one dimension silently dominate the overall priority without explanation. Mark any item where frequency data is missing as "needs data" rather than guessing a number. </constraints> <format> A scale definition block, followed by a table with columns: Friction Point, Frequency, Customer Impact, Business Risk, Overall Priority. </format>

You are a customer experience analyst who measures whether a shipped fix actually reduced friction. <context> We shipped a change intended to fix a known friction point and I need an honest read on whether the after-data shows real improvement or just noise. </context> <inputs> - Friction point and the fix shipped: [DESCRIBE] - Before data: [PASTE METRICS OR FEEDBACK] - After data: [PASTE METRICS OR FEEDBACK] - Time elapsed since the fix: [TIMEFRAME] </inputs> <task> Compare the before and after data and give a clear verdict on whether the friction point improved, stayed the same, or got worse. State your confidence level and what would strengthen the conclusion. </task> <constraints> Do not declare success from a small sample or a short time window, say so explicitly if the data is insufficient. Avoid celebratory language regardless of the result. State the verdict in the first line, then the supporting reasoning. </constraints> <format> A one-line verdict (Improved / No Change / Worse / Inconclusive), followed by a short paragraph of reasoning and a note on what additional data would help. </format>

You are a customer psychology researcher who connects operational friction to emotional reactions. <context> We understand what is breaking operationally but not how it makes customers feel, and that emotional read matters for how urgently leadership treats it. </context> <inputs> - Friction points identified: [PASTE LIST] - Customer quotes or reactions if available: [PASTE QUOTES, OR LEAVE BLANK] - Product or service: [PRODUCT NAME] - Customer segment: [SEGMENT] </inputs> <task> For each friction point, identify the most likely emotional reaction it triggers in this customer segment, such as frustration, confusion, or distrust, and explain the reasoning. </task> <constraints> Ground emotional labels in the quotes provided where available, otherwise mark the label as inferred. Use a small consistent vocabulary of emotions rather than inventing a new label for each row. Avoid dramatizing minor friction as a major emotional event. </constraints> <format> A table with columns: Friction Point, Likely Emotion, Inferred or Quote-Based, Reasoning. </format>

You are a customer experience strategist who identifies the highest-stakes moments in a journey. <context> Not every step in the journey matters equally, and I need to know which specific moments most strongly shape whether a customer stays, upgrades, or leaves. </context> <inputs> - Full journey steps: [PASTE STEPS] - Product or service: [PRODUCT NAME] - Data on churn or upgrade timing if known: [DESCRIBE OR PASTE] </inputs> <task> Identify the three to five moments of truth in this journey, the specific points where the customer's experience most strongly determines their future decision to stay, upgrade, or leave. Explain why each qualifies. </task> <constraints> Distinguish moments of truth from merely important steps, not every touchpoint qualifies. Justify each selection with a specific reason tied to the data or journey logic provided. Limit to five moments maximum. </constraints> <format> A numbered list of moments of truth, each with: Moment, Why It Matters, What's At Stake If It Goes Wrong. </format>

You are a customer experience designer who writes detailed experience specifications for critical moments. <context> We have identified a single make-or-break moment in the journey and I need a concrete description of what the ideal experience should look like there, not vague principles. </context> <inputs> - The moment: [DESCRIBE THE MOMENT, e.g. FIRST FAILED PAYMENT] - Current experience at this moment: [DESCRIBE WHAT HAPPENS TODAY] - Customer segment: [SEGMENT] - Constraints we operate under: [BUDGET, TOOLING, TEAM SIZE, etc] - Owner who will implement this: [ROLE OR TEAM] </inputs> <task> Describe the ideal experience at this specific moment in concrete, actionable detail, including what the customer sees, what message they receive, and what happens next. Then note the gap from the current experience. </task> <constraints> Stay within the stated operational constraints, do not propose solutions requiring resources we do not have. Be specific about wording and timing, not just intent. Keep the ideal experience description to one paragraph. </constraints> <format> Two sections: "Ideal Experience" as a paragraph, and "Gap From Today" as a short bullet list. </format>

You are a customer success lead who writes recovery communications for failed moments. <context> A specific moment in our journey commonly goes wrong for customers and I need a recovery script the team can use to win back trust when it happens. </context> <inputs> - What goes wrong: [DESCRIBE THE FAILURE] - Customer segment affected: [SEGMENT] - Channel for recovery: [EMAIL, PHONE, CHAT] - Tone we want: [DESCRIBE, e.g. DIRECT AND ACCOUNTABLE, NOT CORPORATE] </inputs> <task> Write a recovery script for the specified channel that acknowledges the failure plainly, explains what will be done about it, and rebuilds confidence without over-promising. </task> <constraints> Do not use corporate apology language like "we apologize for any inconvenience", be specific and human instead. Keep it under 150 words for chat or email, adjust for a natural phone call length if that channel is chosen. Do not promise a fix timeline unless one is given in the inputs. </constraints> <format> The script itself, followed by a short note on when the team should escalate instead of using this script. </format>

You are a customer experience researcher studying which moments customers actually remember afterward. <context> We want to focus improvement effort on moments customers actually remember and talk about later, not just moments that are operationally significant to us internally. </context> <inputs> - Candidate moments: [PASTE LIST] - Customer testimonials, reviews, or quotes referencing specific moments: [PASTE TEXT] - Product or service: [PRODUCT NAME] </inputs> <task> Rank the candidate moments by how likely each is to be remembered and mentioned by customers later, based on the testimonial and review evidence provided. Note which moments have no supporting evidence at all. </task> <constraints> Base rankings on the evidence given, not on assumptions about what should be memorable. Explicitly separate moments with direct evidence from those ranked by inference. Limit to the top six moments. </constraints> <format> A ranked table with columns: Moment, Memorability Rank, Evidence Type (Direct Quote / Inferred / None), Supporting Detail. </format>

You are a customer experience designer who proposes small, low-cost moments of delight. <context> We want to add one small unexpected positive moment somewhere in the journey, something low cost that creates a disproportionate positive impression. </context> <inputs> - Journey stage to target: [STAGE] - Product or service: [PRODUCT NAME] - Budget or effort ceiling: [DESCRIBE, e.g. NO NEW HEADCOUNT, UNDER $2 PER CUSTOMER] - Brand tone: [DESCRIBE] </inputs> <task> Propose three small delight moments that could be inserted at this stage, each staying within the stated budget and matching the brand tone. Explain why each would feel unexpected rather than gimmicky. </task> <constraints> Stay strictly within the stated budget or effort ceiling, do not propose anything requiring new tooling unless explicitly allowed. Avoid generic ideas like "send a thank you email" unless given a specific, non-generic twist. Limit to three proposals. </constraints> <format> Three numbered proposals, each with: Idea, Why It Feels Unexpected, Estimated Cost or Effort. </format>

You are a customer analytics lead who selects meaningful metrics for journey stages. <context> We track a pile of metrics but I am not confident they actually match what matters at each stage of the journey, and I want that checked before our next reporting cycle. </context> <inputs> - Journey stages: [LIST STAGES] - Metrics currently tracked per stage: [PASTE LIST] - Business goal: [GOAL, e.g. IMPROVE RETENTION] - Product or service: [PRODUCT NAME] </inputs> <task> Review the metrics currently tracked at each stage and assess whether each one actually reflects customer experience at that stage or just an internal operational number. Recommend replacements where needed. </task> <constraints> Distinguish clearly between customer-experience metrics and internal operational metrics, do not treat them as interchangeable. Only recommend a replacement metric that could realistically be measured with common analytics or survey tools. Keep reasoning to two sentences per stage. </constraints> <format> A table with columns: Stage, Current Metric, Customer-Facing or Operational, Recommendation. </format>

You are a customer experience lead who reports journey health to executives. <context> I have journey data from several sources and need a tight executive summary for a leadership review, not a raw data dump. </context> <inputs> - Data sources and key numbers: [PASTE METRICS, NPS, CHURN, SUPPORT VOLUME, etc] - Time period: [QUARTER OR MONTH] - Biggest known issue this period: [DESCRIBE] - Audience: [e.g. CEO AND VP OF SALES] - Product or service: [PRODUCT NAME] </inputs> <task> Write an executive summary of customer journey health for this period, covering what improved, what got worse, and the one issue that most needs leadership attention. </task> <constraints> Open with the single most important takeaway, not background context. Keep the whole summary under 200 words. Do not include a metric without stating whether it moved up, down, or stayed flat compared to the prior period. </constraints> <format> A one-sentence headline takeaway, followed by three short paragraphs: What Improved, What Got Worse, What Needs Attention Now. </format>

You are a customer experience operations lead who builds ongoing measurement systems. <context> We want a lightweight recurring scorecard to track journey health over time instead of building a new report from scratch every quarter. </context> <inputs> - Journey stages: [LIST STAGES] - Metrics available today: [LIST WHAT YOU CAN ACTUALLY MEASURE] - Reporting cadence: [WEEKLY, MONTHLY, QUARTERLY] - Audience for the scorecard: [WHO SEES THIS] </inputs> <task> Design a simple scorecard structure with one primary metric and one supporting metric per stage, using only metrics already available. Note what threshold would count as a warning sign per metric. </task> <constraints> Use only metrics listed as available, do not propose a metric requiring new instrumentation unless there is no viable existing option. Keep the total scorecard to one metric pair per stage, resist adding extras. State warning thresholds as concrete numbers or ranges, not "if it looks bad". </constraints> <format> A table with columns: Stage, Primary Metric, Supporting Metric, Warning Threshold. </format>

You are a customer experience analyst who connects journey problems to business outcomes. <context> Leadership will not prioritize a friction fix without a rough sense of what it is costing us, and I need that estimate built transparently from real numbers, not invented ones. </context> <inputs> - Friction point: [DESCRIBE] - Number of customers affected per period: [NUMBER OR ESTIMATE] - Average customer value: [NUMBER, e.g. ARPU OR AVERAGE ORDER VALUE] - Assumed impact rate: [e.g. WHAT PERCENT OF AFFECTED CUSTOMERS CHURN OR DOWNGRADE, IF KNOWN] - Time period for this estimate: [MONTH OR QUARTER] </inputs> <task> Build a rough revenue impact estimate for this friction point using the numbers given. Show the calculation explicitly and state every assumption made. </task> <constraints> Show the full calculation, not just a final number, so it can be checked and challenged. Clearly separate given data from assumed figures. If a needed number is missing, state a reasonable placeholder assumption and flag it rather than skipping the estimate entirely. </constraints> <format> A short "Assumptions" list, followed by the calculation shown step by step, ending in a single estimated revenue impact figure with a range, not a false-precision point number. </format>

You are a customer experience program manager who plans quarterly improvement work. <context> We have a backlog of known friction points and journey gaps, and I need them turned into a realistic quarterly roadmap the team can actually execute. </context> <inputs> - Backlog of known issues: [PASTE LIST] - Team capacity this quarter: [DESCRIBE, e.g. 2 ENGINEERS, 1 DESIGNER, PART TIME] - Top business priority this quarter: [PRIORITY] - Product or service: [PRODUCT NAME] </inputs> <task> Build a quarterly roadmap from the backlog, sequencing items against the stated capacity and top priority. Group items into what fits this quarter versus what should wait. </task> <constraints> Be realistic about capacity, do not schedule more work than the stated team can plausibly complete. Tie each scheduled item back to the top business priority in one clause. Put everything that does not fit into a clearly labeled "Next Quarter" list rather than dropping it silently. </constraints> <format> Two sections, "This Quarter" and "Next Quarter", each a numbered list with a one-line justification per item. </format>

You are a qualitative researcher who designs customer interviews focused on journey experience. <context> I am about to run a round of customer interviews and want questions specifically designed to surface friction and emotional moments, not generic satisfaction questions. </context> <inputs> - Journey stage or moment of focus: [STAGE OR MOMENT] - Customer segment being interviewed: [SEGMENT] - Interview length: [e.g. 30 MINUTES] - What we already suspect is a problem: [DESCRIBE, OR LEAVE BLANK] </inputs> <task> Write an interview question guide focused on this journey stage, designed to surface specific moments, emotions, and friction rather than general opinions. </task> <constraints> Avoid leading questions that suggest an answer, especially around the suspected problem area. Use open-ended phrasing throughout. Fit the guide to the stated interview length, roughly two minutes per question including follow-up. </constraints> <format> A numbered list of primary questions, each with one or two suggested follow-up probes indented beneath it. </format>

You are a customer research analyst who synthesizes open-ended survey responses. <context> We ran an NPS or CSAT survey with open text responses, and I need the verbatims synthesized into themes tied to journey stages, not just a word cloud. </context> <inputs> - Open text responses: [PASTE VERBATIMS] - Survey type and score context: [e.g. NPS, SCORES INCLUDED OR NOT] - Journey stages we care about: [LIST STAGES] - Product or service: [PRODUCT NAME] </inputs> <task> Synthesize the verbatims into themes, map each theme to the journey stage it most relates to, and note whether the theme skews positive or negative. </task> <constraints> Use the respondents' own language in theme names where possible instead of abstract labels. Do not merge a positive and a negative theme into one row even if they relate to the same topic. Cap the output at ten themes. </constraints> <format> A table with columns: Theme, Related Journey Stage, Sentiment (Positive/Negative/Mixed), Sample Verbatim, Frequency Estimate. </format>

You are a customer insights lead who scopes new research based on existing feedback signals. <context> We have scattered signals suggesting a problem somewhere in the journey but not enough clarity to act, and I need a tight research brief to get proper study approved. </context> <inputs> - Existing signals: [PASTE FEEDBACK, TICKETS, OR METRICS SUGGESTING A PROBLEM] - Suspected journey area: [STAGE OR AREA] - Research budget or method constraints: [e.g. NO BUDGET FOR PANEL, INTERNAL CUSTOMERS ONLY] - Decision this research needs to support: [DECISION] </inputs> <task> Write a research brief that states the problem hypothesis, the specific questions the research needs to answer, the recommended method given the constraints, and how the decision will use the findings. </task> <constraints> State the hypothesis as a hypothesis, not as an established fact. Recommend only methods feasible under the stated budget and access constraints. Keep the brief to one page equivalent, under 350 words. </constraints> <format> Four labeled sections: Hypothesis, Research Questions, Recommended Method, How This Informs the Decision. </format>

You are a customer research analyst who compares experience across segments. <context> We suspect two different customer segments experience the same journey very differently, and I need that comparison made explicit with the feedback we have. </context> <inputs> - Segment A feedback: [PASTE FEEDBACK] - Segment B feedback: [PASTE FEEDBACK] - Segment A description: [DESCRIBE] - Segment B description: [DESCRIBE] - Journey stages both segments pass through: [LIST STAGES] </inputs> <task> Compare the feedback from both segments stage by stage, noting where their experience aligns and where it clearly diverges, along with a plausible reason for each divergence. </task> <constraints> Do not force a comparison at every stage if the feedback does not support one, mark those as "insufficient data to compare". State divergence reasons as plausible, not certain, unless directly stated in the feedback. Keep each stage comparison to three sentences or fewer. </constraints> <format> A table with columns: Stage, Segment A Experience, Segment B Experience, Divergence Noted (Yes/No/Insufficient Data), Plausible Reason. </format>

You are a customer success manager who follows up on unclear feedback to get to the real issue. <context> A customer left feedback that is too vague to act on, and I need to reach out with follow-up questions that will actually get specific, useful detail without sounding like an interrogation. </context> <inputs> - Original feedback: [PASTE FEEDBACK] - Customer context: [ACCOUNT SIZE, TENURE, PLAN, etc] - Channel for follow-up: [EMAIL, CALL, IN-APP] - Relationship tone: [e.g. HIGH-TOUCH ENTERPRISE, LIGHT-TOUCH SELF-SERVE] - Product or service: [PRODUCT NAME] </inputs> <task> Write two or three follow-up questions that would turn this vague feedback into something specific and actionable, phrased appropriately for the stated channel and relationship tone. </task> <constraints> Avoid stacking more than three questions in one message, that reads as an interrogation. Do not ask a question the original feedback already answered. Match the phrasing to the relationship tone, formal for enterprise, casual and short for self-serve. </constraints> <format> A short message draft ready to send, containing the two or three questions embedded naturally in one or two sentences each. </format>