Give your team a useful record of each eligible AiDial conversation, with call summaries and configured protection for personal information. Choose what is retained, who receives it and how staff follow up.
Call transcription turns a spoken conversation into a written record. For calls handled by your AiDial receptionist, it helps staff understand the enquiry, review what was said and decide the next step.
Redaction masks selected personal information in saved text. A summary gives staff a shorter account of the conversation. These features work together, with consent, retention and delivery settings agreed for your project.
Start with a practical question: what does your team need after a call? We configure the record and notifications around that purpose, whether it is returning an enquiry, reviewing an appointment request or following up after a staff handover.
An eligible call can produce a speaker-labelled transcript showing the conversation between the caller and AiDial. Your team can use that record to review the enquiry and see which details were discussed, without relying on a short message alone.
AI-generated summaries highlight the reason for the call and the next steps discussed. Where analytics is enabled, sentiment can provide another signal for staff review. These are aids to judgement: a transcript or summary may contain recognition errors, and a conversational statement does not independently prove that a booking or another action succeeded.
The record covers the AiDial conversation. After a caller is handed to your team, do not assume that the onward human conversation is included. See AI call transfer and human handover for how that boundary works.
Personally identifiable information, or PII, can appear naturally during a call. Depending on your project configuration, AiDial can mask names, email addresses, phone numbers and supported Australian identifiers in persisted transcripts and related text outputs.
Coverage is configured for your workflow. Australian formats such as TFNs, Medicare numbers and BSBs can be included, with project-specific patterns for relevant variations. We test representative wording and formatting rather than assume that one rule fits every business.
Redaction takes place when text is prepared for storage and applicable downstream outputs. It does not mean that personal information is never processed during the conversation. Detection has limits, and masking a transcript does not remove the same information from an audio recording.
In this illustrative example, the saved transcript keeps the caller’s request while replacing the name and contact details with placeholders. Your actual placeholders and the information masked depend on the agreed configuration.
The goal is to leave staff with useful context and reduce unnecessary exposure in saved text. If a member of staff needs an original detail for an authorised purpose, agree the permitted access or follow-up route separately.
This is an explanation of the text workflow, not a product screenshot or a promise that every sensitive detail will be detected.
Before redaction
Caller: I am Alex Example. Please email alex@example.com or call +61 4XX XXX XXX about changing my appointment.
Saved with the selected details masked
Caller: I am [NAME]. Please email [EMAIL] or call [PHONE] about changing my appointment.
Fictional details; X replaces phone digits. Actual detection and placeholder labels depend on configuration.
The transcript workflow follows the consent and retention rules configured for your project. When consent is required, eligible transcript storage and delivery depend on it. Decline handling can be configured to continue without retained recording or transcription, end the call or offer a staff handover.
Where redaction is enabled, the readable saved transcript masks the configured information. Depending on your project’s compliance configuration, an encrypted original can also be retained for an authorised purpose. A redacted copy and an encrypted original have different access and disclosure implications; both need clear rules.
Agree who may receive exports, whether originals are needed and how long the relevant records should be retained. AiDial’s controls support that workflow; your organisation determines its obligations and approved use. Read more about AiDial privacy and compliance controls.
Choose how your staff receive eligible post-call information. Depending on the configured services, options include a transcript or summary by email, a short call summary by SMS, and consent-aware delivery through an integration.
Recipients, output format and delivery channels are agreed during setup. Redacted, original and summary-only exports have different requirements. Delivery depends on the configured channel, the available record and the relevant privacy gates; it is not a guarantee of an instant notification after every call.
For example, a practice team may need the reason for an appointment enquiry and a clear follow-up task. A property team may need a summary of an urgent maintenance request. Discuss what staff need to act, then connect that workflow to your existing business systems or SMS follow-up.
Call transcription turns a spoken conversation into a written record. For eligible calls handled by AiDial, the transcript can show what the caller and AI assistant said, helping your team review the enquiry and plan its follow-up.
This page describes transcription for conversations handled by the AiDial platform. It does not offer a general audio-upload transcription service. If you have another call source or a specific integration requirement, discuss the connection and scope with our team.
PII redaction masks selected personally identifiable information in text. In an AiDial transcript workflow, configured details can be replaced with placeholders so the readable saved record keeps useful context while exposing fewer personal details.
Configured coverage can include names, phone numbers, email addresses and supported Australian identifiers such as TFNs, BSBs and Medicare numbers. The selected types and any custom patterns depend on your project. Coverage should be tested against representative calls and formatting.
The active conversation can process unredacted information. Redaction is applied when eligible text is prepared for persistence and applicable outputs. It should not be described as preventing all personal information from entering the voice-processing workflow.
No. Masking information in saved text does not remove it from an audio recording. Recording has separate configuration, consent, storage and retention controls, which should be agreed alongside your transcript workflow.
No detection system should be treated as a guarantee that every sensitive value will be found. Speech recognition, wording and formatting can affect detection. Choose appropriate coverage, test realistic examples and keep collection and access limited to the business purpose.
Transcripts can contain recognition errors, especially with background noise, names or unusual terminology. Staff should review details that matter before acting. A transcript or AI summary is a useful record, not independent proof that a requested action succeeded.
Yes, eligible post-call email can be configured for summary-only delivery. SMS call summaries are another option where that service is enabled. The content, recipients and privacy requirements are agreed for your project.
Eligible transcript email can be enabled after a call, using configured recipients, templates and output formats. Delivery depends on the available stored record, the email service and applicable consent and retention checks. It is not enabled for every project by default.
Depending on your project's compliance configuration, an encrypted original may be retained alongside the readable redacted record. Its purpose, retention and permitted export need to be agreed. Encryption and redaction protect information in different ways.
The configured decline workflow can continue without retained recording or transcription, end the call, or offer a handover to staff. The required notice, consent mode and handling should be agreed before activation. Do not assume that one configuration is suitable for every organisation.
Retention depends on the applicable project and data-retention configuration. Agree the purpose and retention requirements during setup, including whether encrypted originals are required. We do not promise one universal retention period for every business.
Do not assume so. The record described here covers the AiDial conversation and its handover boundary. Onward conversations handled by your staff or phone system can have separate recording and transcription arrangements.
Eligible call data can be connected to supported systems or approved workflows. The specific connector, event, format, recipients and privacy controls need to be assessed and configured. See the AiDial integrations page or discuss your required workflow with our team.
Start with the calls AiDial will handle and what staff need afterwards. Agree the transcript and summary outputs, redaction coverage, consent handling, recipients and retention. Then test representative calls and staff follow-up before activating the workflow.
Tell us which calls AiDial will handle, what your team needs afterwards and which information should be protected. We will help you scope the configuration and checks before activation.
