> ## Documentation Index > Fetch the complete documentation index at: https://help.voxdash.com/llms.txt > Use this file to discover all available pages before exploring further. # Sample Records > Open one record of an AI-CATI study to see every call placed to it, play recordings, read transcripts, and review the answers and interview quality. A **sample record** is one person (one row of your sample) in an AI-CATI study. Its page shows everything that happened to that record: every call, its recording and transcript, and the answers given. Open a record from **Open** on the [Results](https://help.voxdash.com/AI-CATI/Results-and-Export.md) tab, from a pre-test run, or from a sub-sample run's records table. **Back to results** (or **Back to pre-test run**) returns you to the list with your filters kept. ## Record header | Field | Meaning | | - | - | | **Status** | See the table below. | | **Key type** | **Phone**, or **Composite (phone and external ID)**. | | **External sample ID** | The ID from your sample file, if it has one. | | **Enrolled** | When the record entered the campaign. | | **Number of calls** | How many calls have been placed. | | **Last event** | When something last happened to the record. | | **Phone** | Masked. Select **reveal** to show the full number and **hide** to mask it again. | If a callback is booked, a notice shows when: "A callback is scheduled for …". Select **Export record** to download this record's data. | Record status | Meaning | | - | - | | Pending | Not called yet; waiting for its first call. | | In progress | Called at least once and not settled yet. A call may be on the line right now. | | Callback scheduled | A callback is booked, and the record is waiting for it. | | Finished | Settled with a completed interview. | | Partial | Reached, but the interview stopped before the end. Settled and not called again. | | Closed | Settled without an interview, for example a refusal or all attempts used. Not called again. | | Waiting for capacity | No calling line came free in time, so the number was never dialled and nothing was spent. | ## Call history The **Call history** tab lists every attempt, oldest first. | Column | Description | | - | - | | **Attempt No.** | For example "2 of 6", against the study's maximum attempts. | | **Call time** / **Duration** | When the call was placed and how long it lasted. | | **AAPOR · Next call** | The call's AAPOR disposition and, if the record will be called again, when. | | **Voice** | **Play** or **Download** the recording. If nothing was recorded, the reason is shown, for example the respondent declined recording, consent was not asked, or recording is off for the study. | | **Transcript** | **View transcript**, when one is available. | **Dial and callback audit** lists each dial and callback with when it happened, who triggered it (the **AI interviewer** or the **calling rules**), the caller ID used and the outcome. ### Call transcript **View transcript** opens the conversation turn by turn: * **AI** and **Resp** (respondent) turns, each respondent turn with its STT confidence, and an **interrupted** marker where someone was cut off. * An audio player to play, pause and seek the recording, and **Download the recording**. * **Download as text** to save the transcript. * An option to **show technical events**: what the system recorded during the call, such as answers saved or changed, questions skipped because they do not apply, a respondent who did not qualify, the call outcome, and notes the system gave the AI (for example that an answer was unclear or the line went quiet). Answers flagged **Needs review** may not match what was said. > [!WARNING] > > Recordings and transcripts contain personal data. Download and share them only in line with your organization's data-protection rules and the consent respondents gave. ## Result The **Result** tab shows the answers per question (Q1, Q2, …) for the **fielded version** of the questionnaire. Each question shows its answer and a status badge: | Badge | Meaning | | - | - | | **Answered** | The respondent gave an answer. | | **Not answered** | The question was asked but no answer was recorded. | | **Skipped** | The question did not apply. | | **Not asked** | The question was never reached. The reason is shown, for example the survey logic skipped it, the call ended before it, a callback is booked, or the call dropped because of a technical failure. | | **Derived** | A computed value, never asked on the call. | Select **Download this record's result** to save the answers as a file. A partial record may have blank later questions. | Message | What it means | | - | - | | No interview has resolved for this record yet | There are no answers to show yet. | | This record broke off part-way and its partial answers have not been ingested | The partial interview was not saved. | | Answer ingest is not configured for this study | Per-question answers are unavailable. | ## Interview quality The quality panel summarises the record's calls: | Measure | Meaning | | - | - | | **Script adherence** | How closely the AI interviewer kept to the questionnaire's wording, in percent. | | **STT confidence** | How sure speech recognition was of what the respondent said (0 to 1). | | **Latency p50 / p95** | The typical and 95th-percentile time the AI took to reply. | | **Speedsters flagged** / **Straight-liners flagged** | Calls flagged for answering too fast, or for giving the same answer question after question. Zero is healthy. | | **Response-time anomalies** | Answers with an unusual response time. Zero is healthy. | | **Technical failure rate** | The share of the record's calls that ended in a technical failure. Zero is healthy. | ## Related pages * [Results and Export](https://help.voxdash.com/AI-CATI/Results-and-Export.md) * [Tests: Self-tests and Pre-tests](https://help.voxdash.com/AI-CATI/Tests-Pre-tests-and-Self-tests.md) This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.