Expert Interview Synthesis
Plan neutral interviews and synthesize supplied evidence into traceable findings, conflicting views and verification actions.
$5
Works with the AI Agents you already use
Claude Code
Cursor
Codex CLI
Muse
Dots+22 moreClaude CodeCursorCodex CLIMuseDots
OpenClaw
Manus
Grok
Grok Bot+18 agentsExpert Interview Synthesis
Example session with this skill installed
Fictional training scenario; no real client data.
Use $expert-interview-synthesis to decide whether our eight-week triage pilot supports rolling out to all sites. Prepare next-round questions and synthesize the supplied notes without claiming representative market demand.
All four source records refer to the same pilot period. Recording/publication consent is not supplied; use anonymized IDs and do not initiate outreach.
- I1, site operations lead, transcript paragraph 4: “The routing queue was shorter on most mornings, but we still checked the sensitive tickets manually.” Firsthand experience; no queue measurements supplied.
- I1, paragraph 7: “The vendor slide says each site saves twenty hours a week.” Repetition of a vendor claim, not a measured result.
- I2, finance analyst, transcript paragraph 3: “Paid overtime was unchanged in the pilot. We have not separated triage time from other support work.” Firsthand access to payroll; underlying records not attached.
- I3, second-site team lead, notes section B: a supervisor reports that the tool missed exceptions twice in week six. The note summarizes this statement; no exact wording or incident log is available.
- I4, reseller, transcript paragraph 2: “The vendor slide says each site saves twenty hours a week.” Same underlying slide as I1; no pilot data or independent test.
- Read your context and instructions
- Compiled the expert interview synthesis
- Generated the data export
Fictional worked illustration from the public repository, not a recorded agent run or a verified client outcome. Any additional figures, thresholds and market signals are scenario assumptions, not observed results. No generated PDF or finished deck is included.
Next-round guide
Ask the operations lead to walk through one recent ticket and show queue timestamps, including manual checks. Ask finance which costs or hours could actually be avoided and which were redeployed. Ask the second-site supervisor to describe the two exceptions and the impact of the manual fallback. Ask for a contrary case: when was routing slower or less reliable than before?
Confirm consent, confidentiality and access to permissible records before interviews or recording. These are proposed questions, not sent invitations.
Evidence ledger
| ID | Provenance | Extract | Type and interpretation | Confidence |
|---|---|---|---|---|
| E1 | I1 ¶4 | “The routing queue was shorter on most mornings, but we still checked the sensitive tickets manually.” | Scoped firsthand perception of speed; manual control remains necessary | Directional |
| E2 | I1 ¶7 and I4 ¶2 | “The vendor slide says each site saves twenty hours a week.” | One underlying vendor claim repeated by two people; not two independent confirmations | Unverified |
| E3 | I2 ¶3 | “Paid overtime was unchanged in the pilot. We have not separated triage time from other support work.” | Attributed cost observation requiring payroll corroboration; not proof of zero operational benefit | Directional |
| E4 | I3 notes B | Paraphrase: two missed exceptions in week six | Secondhand summary; original wording and incidence denominator unknown | Unverified |
Conflict register and findings
Speed is promising but not quantified. E1 suggests a shorter perceived queue in one site, not measured throughput gains across all sites. E4 points to a quality risk that may differ by ticket type; retain it even though it is a minority account.
Twenty hours of savings is not corroborated. E2 has one commercial source and no period-matched measurement. E3 concerns paid overtime, which differs from time released or redeployed. The apparent disagreement may be about definitions; do not average “twenty” and “zero.”
The sample contains four participant roles, but two repeat one claim and one record is paraphrased hearsay. It cannot estimate population prevalence, rollout adoption or the percentage of sites that will benefit.
Decision and next verification
Do not recommend an unrestricted rollout or book cash savings from this evidence. Propose a staged pilot extension with baseline/post queue logs, time-on-task sampling, exception denominators and incident severity. Operations owns speed/quality measurement; finance validates avoided versus redeployed cost; the pilot sponsor approves the gate before expanding.
A gate cannot be invented from missing business tolerances. Ask the sponsor to agree quality and contribution thresholds. If critical exception controls cannot be demonstrated, pause expansion and retain manual checks. Handoff E1–E4 with their provenance and unresolved status to the transformation plan or decision memo.
Editable files: interview guide and evidence ledger.
interview-evidence.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
Prepare neutral interview questions and turn supplied notes into traceable findings. Preserve quotes versus paraphrases, contradictory views, source independence and decision-relevant evidence gaps.
What you get
SKILL.md, evidence-method and source-note references, editable interview-guide Markdown and evidence-ledger CSV, agent metadata and MIT LICENSE.
Primary deliverable: Interview guide, evidence ledger and findings memo.
How to use
Unzip the package into your compatible agent's skills folder. Ask the agent to use expert-interview-synthesis and supply your decision, evidence and constraints. Review the result before acting on it.
Open source and purchase transparency
Version 1.1.0. The same skill is available free under the MIT license in Consulting Skills Lab. This optional $5 purchase supports the creator and provides the packaged download through Agensi; it is not exclusive content and does not include an AI-model subscription, hosted service, consulting engagement or guaranteed future support. The MIT license is included in the ZIP.
Example and limits
The listing demo is a fictional worked illustration, not a recorded agent benchmark or a real customer result. Read the public example.
Works from supplied material. Does not recruit, record or contact participants. A small qualitative sample does not establish market prevalence or independently verify claims. Consent and confidential-data handling remain the user's responsibility.
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 11 days ago
- Passed all security checks, Safe to install