- Home
- Skills
- Research & Analysis
- Parallax Customer Evidence
More screenshots
Works with the AI Agents you already use
Parallax Customer Evidence
Find recurring needs and buying objections in customer interviews, reviews and sales chats, with quotes to support each finding.
$7
Parallax Customer Evidence
Example session with this skill installed
Fictional raw research packet: BenchRelay
All people, organizations, products, quotations, and events below are fictional, created for this exercise.
User request
We are considering BenchRelay, a shared handover note for bicycle repair work. Our initial idea is a $20/month subscription that every small repair operator could use. We have not charged anyone for it. Please tell us what these materials support, which customer group and offer we should explore first if any, and what the next two interviews and one small pilot should learn. Include a traceable evidence table and counts I can explain to my co-founder. We would also like to know whether these notes justify a sales claim. Use English.
The three workshop businesses are the primary research group. We also spoke to a mobile operator and a volunteer club organizer because they repair bicycles. Participants were recruited through the founder's contacts; there was no random sampling. The dates below are in 2026. Internal analysis is permitted; nobody has given permission to become a public testimonial.
Participant registry
| Participant | Organization | Role | Context |
|---|---|---|---|
| P01, Dana | A01, Finch Wheelworks | Owner; purchases software | Workshop with four staff |
| P02, Ivo | A01, Finch Wheelworks | Lead mechanic; does not control software budget | Same workshop as P01 |
| P03, Lena | A02, Lantern Cycles | Owner and mechanic | Workshop with two staff |
| P04, Remy | A03, Corner Spoke | Workshop manager; owner approves subscriptions | Workshop with six staff |
| P05, Sal | A04, Rollside Mobile | Sole proprietor | Mobile repair visits; works alone |
| P06, Terry | A05, Northbank Repair Club | Volunteer organizer | Free monthly community event; no purchasing budget |
S01 — P01 interview, 4 October
Recruiter note: The interviewer described a shared handover app before asking about problems. Asked about recent missed handovers, current tools, and possible price. No observation of the workshop was conducted.
[L1] Interviewer: What happens when someone else picks up a repair?
[L2] P01: "We lose maybe five jobs a week because the next person doesn't know what was agreed. At least, that's what it feels like when the phone rings."
[L3] P01: "We use a whiteboard for status and a paper tag for the customer's instructions. I pay $45 a month for booking reminders, and I'd only want another tool if it replaced something."
[L4] Interviewer: Would $20 a month be reasonable for the handover app I showed you?
[L5] P01: "Twenty sounds fine if the mechanics actually use it and I can stop paying for reminders. I haven't tried your app yet."
S02 — P01 follow-up message, 8 October
This is the same P01 as S01. The founder asked Dana to check the examples and numbers from the interview.
[L1] P01: "I checked our book for the last four weeks. There were two late pickups after a handover, not five lost jobs a week. Neither customer asked for a refund. Please correct that."
[L2] P01: "One was waiting for a part. For the other, nobody wrote down that the customer needed the bike by lunchtime. I still want that note visible."
[L3] P01: "The reminder company can't be cancelled until January. I wouldn't buy another subscription this month."
S03 — CRM excerpt export, 9 October
Metadata: Participant P01, organization A01. Imported from the S01 interview transcript, not a new interview.
[Row 14, text] "We use a whiteboard for status and a paper tag for the customer's instructions. I pay $45 a month for booking reminders, and I'd only want another tool if it replaced something."
[Row 14, CRM sales status] Interested
S04 — P02 interview, 5 October
Asked about the last handover, current tools, and use of a shared note. Did not ask about price or purchasing authority beyond the role noted in the registry.
[L1] P02: "Yesterday I picked up a brake job after lunch. The paper tag said 'call first', so I called before fitting the part. It worked."
[L2] P02: "When the tag is missing, I walk over and ask whoever spoke to the customer. That happened twice last week. It took a minute or two each time."
[L3] P02: "I don't want another thing to type into with greasy gloves. If the front desk writes it once and I can see it beside the job, that's useful."
[L4] P02: "Dana says the phone interruptions are worse than I think. I'm usually on the repair stand, not answering calls."
S05 — P03 interview, 6 October
Asked about the last handover, consequences, current spending, and alternatives. No product demo was shown.
[L1] P03: "There are only two of us and we overlap all day. We say the job number aloud and put the tag on the handlebars."
[L2] P03: "The last missing note I remember was in August. I called the customer back. It was awkward, but the repair still went out that day."
[L3] P03: "Our problem is finding parts, not passing notes. I wouldn't add a subscription to solve this."
[L4] P03: "We pay the card processor and that's it for software. If we hired an evening shift, I might think differently."
S06 — P04 interview after a free trial, 10 October
Asked about actual trial use, current workaround, next use, and purchasing route. The founder supplied a free prototype on 7 October; it had no checkout or payment screen.
[L1] P04: "It looks lovely, and I clicked the thumbs-up on your launch post. I put eighteen open jobs into the trial on Tuesday."
[L2] P04: "By Thursday we were back to the group chat. I was entering the same update in the booking system and your tool. I couldn't keep doing both."
[L3] P04: "Every Saturday, the morning team leaves and a different team closes. Last Saturday I called someone at home to ask whether a wheel was ready. That's the moment I want to fix."
[L4] P04: "I could arrange a Saturday test, but the owner decides about subscriptions. She hasn't seen it. I can't tell you what she'd pay."
S07 — prototype event export, 7–10 October
Provided by the founder. Organization A03 and participant P04 are linked through the prototype account. The logger captures only the events listed; it does not record calls, job completion, or users of the shared screen.
| Row | Date | Participant | Event | Value |
|---|---|---|---|---|
| 1 | 7 October | P04 | account_created | free_trial |
| 2 | 7 October | P04 | import_completed | 18 jobs |
| 3 | 7 October | P04 | dashboard_opened | 1 session |
| 4 | 8 October | P04 | dashboard_opened | 1 session |
S08 — P05 interview, 7 October
Asked about recent work, customer updates, current spending, and price. Sal works alone and does not hand work to another technician.
[L1] P05: "I message customers when I'm on my way. Yesterday I forgot one update and the customer had already left for work. I drove there for nothing."
[L2] P05: "I use my calendar and ordinary messages. A reminder to send my arrival time could help. A shared job board would just be me talking to myself."
[L3] P05: "I haven't paid for a reminder tool. I would need to see it save me a wasted trip before I considered one."
S09 — P06 group message, 9 October
The founder posted a screenshot and asked for reactions. This was not a structured interview; nobody asked about a recent handover or financial consequence.
[L1] P06: "Nice idea! Our volunteers might enjoy the colored cards. Everything at the club is free, and I don't have a budget for software."
- Read your context and instructions
- Compiled the parallax customer evidence
- Generated the data export
Fictional demonstration from an actual local trial. This is not a real customer result. Local links and machine-specific paths are displayed as text.
Explore a shift-handover pilot, not a subscription for every repair operator
The strongest next opportunity is a narrow test at A03, where a different team closes on Saturdays and P04 can arrange access. First speak to its owner: P04 is a user and manager, not the subscription buyer. Test whether a handover note can stay beside the existing job record without duplicate entry. Do not sell a general $20/month proposition on the present evidence.
There are
zero recorded purchases of BenchRelay. The material does not establish how many people would pay $20, recurring retention, avoided lost jobs, revenue uplift, or market size. No participant has authorised a public testimonial.
Source and sample map
I read the full participant registry and S01–S09. The nine supplied source collections include interviews, a follow-up, a duplicate CRM excerpt, a prototype event export, and a group reaction. The event export has four rows; it does not add four people. evidence-ledger.csv retains the exact participant excerpts with their source locators; duplicate material is identified below.
| Source | Date, 2026 | Person/account | Type and collection context |
|---|---|---|---|
| S01 | 4 Oct | P01/A01 | Interview after interviewer described the proposed app; price prompted at $20 |
| S02 | 8 Oct | P01/A01 | Same person's requested correction/follow-up |
| S03 | 9 Oct | P01/A01 | CRM export of S01 L3; duplicate quotation, no new interview |
| S04 | 5 Oct | P02/A01 | Mechanic interview; price not asked |
| S05 | 6 Oct | P03/A02 | Owner/mechanic interview; no demo shown |
| S06 | 10 Oct | P04/A03 | Manager interview after free trial; buyer not interviewed |
| S07 | 7–10 Oct | P04/A03 | Four logger rows: one account, import of 18 jobs, two dashboard openings |
| S08 | 7 Oct | P05/A04 | Solo mobile operator interview, different workflow |
| S09 | 9 Oct | P06/A05 | Volunteer screenshot reaction; not a structured problem interview |
The registry establishes
six distinct participants across five organisations. The primary workshop sample is
four people across three buying accounts: P01 and P02 are two voices at A01; P03 is A02; P04 is A03. The mobile operator is one person/account; the volunteer club is one person/account with no software budget. Do not pool those segments into a six-person workshop denominator. Recruitment through founder contacts and leading product/price questions at S01 limit generalisation.
Traceable evidence table
Quote IDs below are stable locators; the companion ledger preserves complete source quotations. “Reported” means reported by the participant, not independently observed.
| ID | Source | Exact excerpt or explicitly labelled record | Kind | Interpretation and limit |
|---|---|---|---|---|
| Q01 | S01 L2, P01/A01 | “We lose maybe five jobs a week because the next person doesn't know what was agreed. At least, that's what it feels like when the phone rings.” | Initial reported impression | Superseded as a quantity by Q04; never use as measured loss |
| Q02 | S01 L3, P01/A01 | “We use a whiteboard for status and a paper tag for the customer's instructions. I pay $45 a month for booking reminders, and I'd only want another tool if it replaced something.” | Reported workaround/spending; conditional preference | Existing spend is for reminders, not BenchRelay; S03 repeats this exact quote |
| Q03 | S01 L5, P01/A01 | “Twenty sounds fine if the mechanics actually use it and I can stop paying for reminders. I haven't tried your app yet.” | Prompted future intention | Conditional on use and replacement; no purchase |
| Q04 | S02 L1, P01/A01 | “I checked our book for the last four weeks. There were two late pickups after a handover, not five lost jobs a week. Neither customer asked for a refund. Please correct that.” | Corrected reported past events | Two late pickups in four weeks, not lost jobs; book itself not supplied |
| Q05 | S02 L2, P01/A01 | “One was waiting for a part. For the other, nobody wrote down that the customer needed the bike by lunchtime. I still want that note visible.” | Reported causes and current preference | Only one of the two described cases directly concerns missing customer timing information |
| Q06 | S02 L3, P01/A01 | “The reminder company can't be cancelled until January. I wouldn't buy another subscription this month.” | Current constraint and buying refusal | Revises the practical meaning of Q03; no purchase this month |
| Q07 | S04 L1, P02/A01 | “Yesterday I picked up a brake job after lunch. The paper tag said 'call first', so I called before fitting the part. It worked.” | Reported successful workaround | Paper already succeeds in at least this case |
| Q08 | S04 L2, P02/A01 | “When the tag is missing, I walk over and ask whoever spoke to the customer. That happened twice last week. It took a minute or two each time.” | Reported past interruption | Limited stated time consequence, not a lost job |
| Q09 | S04 L3, P02/A01 | “I don't want another thing to type into with greasy gloves. If the front desk writes it once and I can see it beside the job, that's useful.” | Conditional usability preference | Front-desk entry plus visibility is more credible than mechanic re-entry |
| Q10 | S05 L3, P03/A02 | “Our problem is finding parts, not passing notes. I wouldn't add a subscription to solve this.” | Current priority/refusal | Strong contrary case for a universal workshop subscription |
| Q11 | S05 L4, P03/A02 | “We pay the card processor and that's it for software. If we hired an evening shift, I might think differently.” | Reported spending; conditional future intention | A different shift pattern could change need, but is not present demand |
| Q12 | S06 L1, P04/A03 | “It looks lovely, and I clicked the thumbs-up on your launch post. I put eighteen open jobs into the trial on Tuesday.” | Attitude and reported trial use | Compliment/like not payment; import corroborated by S07, not all chronology |
| Q13 | S06 L2, P04/A03 | “By Thursday we were back to the group chat. I was entering the same update in the booking system and your tool. I couldn't keep doing both.” | Reported abandonment and burden | Material negative evidence despite initial use |
| Q14 | S06 L3, P04/A03 | “Every Saturday, the morning team leaves and a different team closes. Last Saturday I called someone at home to ask whether a wheel was ready. That's the moment I want to fix.” | Reported recurring situation and recent event | Concrete shift transition; frequency of disruptive calls remains unknown |
| Q15 | S06 L4, P04/A03 | “I could arrange a Saturday test, but the owner decides about subscriptions. She hasn't seen it. I can't tell you what she'd pay.” | Access intention and buying authority | Test access possible; payer's view missing |
| E01 | S07 rows 1–4, P04/A03 | Record: free account; 18-job import; dashboard open on 7 and 8 Oct | Supplied event record | No payment, call, completion, or shared-screen-user data |
| Q16 | S08 L1, P05/A04 | “I message customers when I'm on my way. Yesterday I forgot one update and the customer had already left for work. I drove there for nothing.” | Reported past event | A customer-arrival communication problem, not technician handover |
| Q17 | S08 L2, P05/A04 | “I use my calendar and ordinary messages. A reminder to send my arrival time could help. A shared job board would just be me talking to myself.” | Reported workaround and preference | Separate solo-operator use case |
| Q18 | S08 L3, P05/A04 | “I haven't paid for a reminder tool. I would need to see it save me a wasted trip before I considered one.” | No reported spend; conditional future intention | No present buying commitment |
| Q19 | S09 L1, P06/A05 | “Nice idea! Our volunteers might enjoy the colored cards. Everything at the club is free, and I don't have a budget for software.” | Friendly attitude; budget constraint | No tested handover problem or payer |
P02 also reports a perspective difference with P01 at S04 L4, not independent evidence of P01's interruption count. S05 L1–L2 describe an overlapping two-person team and an August missing note resolved by a same-day callback; those details reinforce Q10 without proving the problem never occurs.
Themes and counterexamples
| Theme | Primary-workshop support and denominator | Counterexample or unknown | Decision implication |
|---|---|---|---|
| Some recent handover information goes missing | 3/4 workshop participants: P01 Q05, P02 Q08, P04 Q14; across 2/3 accounts A01/A03 | P03/A02 describes low current need; P02 Q07 shows a working paper tag | Explore the triggering transition rather than all repair businesses |
| Adding another tool/entry step can defeat adoption | 3/4 participants: P01 Q02/Q06, P02 Q09, P04 Q13; 2/3 accounts | P03 rejects the need for a subscription for a different reason; no tested integration solution | Avoid another duplicate entry surface; do not claim an integration exists |
| Shift change is a concrete test situation | P04 is the 1/4 reporting a current different-team Saturday handover; 1/3 accounts | P03 only says a future evening shift might change their view; P01/P02's shift structure not established | A03 is the first test location, not a proven market segment |
| $20/month purchasing support | 0 confirmed BenchRelay buyers across 3 workshop accounts | P01's earlier conditional response is followed by a current refusal; P03 refuses; A03's owner unasked; P02 not asked and lacks authority | No demand count or sales extrapolation |
These are overlapping themes, not additive customer totals. “Not asked” and “not mentioned” are not negative answers. All rates refer only to this small recruited sample.
Offer hypothesis and next two interviews
Hypothesis: a workshop with different people opening and closing may value a handover note entered once and visible beside the existing job. The user is the team doing handovers; at A03, P04 can coordinate the test but the owner decides and pays. A bounded offer candidate is a two-Saturday handover setup/test using their existing workflow, not an extra monthly tool sold on unbuilt integration promises. The paying reason—fewer disruptive follow-up calls or less uncertainty—is still a hypothesis, not a saving already measured.
Interview 1: A03's owner, with P04's context available. Ask: “Walk me through the last Saturday handover you personally dealt with”; “What did the team do when information was missing?”; “What did that cost in time or other consequences, if anything?”; “Which systems must remain in place?”; “Who authorises a workflow or software purchase, and what did the last such purchase involve?” Do not lead with $20 or assume the owner wants the prototype.
Interview 2: a contrasting workshop decision-maker whose opening and closing teams differ but who currently manages handovers without a new tool; recruitment is a proposed criterion, not an existing contact. Ask for the last actual shift transition, how the note was recorded and found, a recent exception, alternatives tried, actual spend, and why the present method might be preferable. This tests whether A03's problem is a wider transition need or a local duplicate-entry/workflow problem. If access is unavailable, deepen A02's successful overlapping-team case rather than invent a new interview.
One small pilot
Ask A03's owner and P04 to agree a
two-Saturday pilot for the existing set of handovers, using only a note arrangement the team can actually maintain without promised integrations. Scope: agree the needed job fields, rehearse the handover, observe the two transitions with permission, and review a simple log afterward. Keep each note linked to the existing job and capture who entered it and whether a second entry was required.
Proposed learning measures: number of eligible handovers, notes available when needed, clarification contacts, duplicate entries, entry time, missing fields, and whether the team returns to group chat. Establish denominators before the first Saturday. Capture a prior comparable Saturday if records exist, but do not claim a causal improvement from an uncontrolled before/after comparison.
Proposed decision rule: continue exploring only if both Saturdays are completed, the team chooses to keep the approach afterward, and the log shows no routine second update needed for the same fact. Any information gap, adoption failure, or workaround return is material evidence to revise the approach. These are selected pilot criteria, not industry benchmarks. A small free pilot tests workability, not payment. After reviewing actual use, show the owner a clearly scoped continuation at the proposed $20/month with explicit cancellation and data-handling terms; only an actual payment would establish a purchase at those terms. Do not treat a yes, a click, a trial, or a deposit as recurring retention.
No outreach, pilot, charge, or publication was executed. The present packet justifies investigating a narrow handover situation; it does
not justify a public claim of proven demand, avoided lost jobs, or results. Testimonials require separate permission. The CRM “Interested” label and launch-post reaction add no new buyer evidence.
evidence-ledger.csv
CSV · data export
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
Find out what customers keep asking for and what stops them from buying. Paste your interview notes, reviews or sales conversations.
What you get
- A summary of recurring needs, objections and alternatives customers use.
- Supporting quotes linked to the original notes.
- Differences between customer groups, plus follow-up questions and a small next test.
Repeated records are checked so the same feedback is not counted twice. Small samples stay small: a few positive comments are not treated as proof of demand.
Load this reusable skill in your Agensi-connected AI assistant. It analyzes the material you provide; it does not collect interviews or scrape reviews. AI service fees are separate.
How to use
Skills on Agensi run inside your AI agent. Copy the prompt, paste it into your agent, and it connects to Agensi and installs this skill for you. If a payment is needed, your agent shows it to you first.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill passes our trust checks below.
Security scanned
Passed our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scanned
Verified clean 3 days ago
- Passed all security checks, Safe to install