# WiseSentry — full site content > AI call-quality monitoring for Australian businesses, by Wise Cloud Technology. > Reads call transcripts and alerts a manager only when the business itself let a > customer down. All data stored and processed in Australia; no call content is sent to > OpenAI, Google or Anthropic. Contact: 1300 321 083 · service@wisecloud.com.au > Wise Cloud Technology, ABN 94 649 433 974, 1/66 Appel St, Surfers Paradise QLD 4217 Generated from the live site. Canonical: https://wisesentry.com.au/ --- # Home WiseSentry · AI call-quality guardian The calls that cost you customers don't sound angry They sound calm, professional, even friendly — because a good agent stays calm while the customer is being let down. Sentiment scoring rates those calls positive. Your QA sample never sees them. WiseSentry reads what was actually said, and tells a manager only when your business is the problem. Book a demo → Call 1300 321 083 The evidence ## One month, one corporate travel agency WiseSentry reviewed every recorded call for a month. Here is what it found — and what a sentiment threshold would have found instead. 2,349 calls reviewed 45 raised to a manager (1.9%) 82 median sentiment of those flagged — positive 3 of the 45 scored 40 or below Read that third number again. The calls that needed a manager scored well on sentiment. If you were watching for low scores, you would have seen almost none of them. That is the entire reason WiseSentry exists. Figures from a 30-day review at an Australian corporate travel agency. Client not named; no call content is published. Why sentiment misses these calls → What a manager receives ## An alert you can act on in thirty seconds You get both. The alert is EMAILED to the manager responsible - with the reason it was flagged, the full transcript and a link to the recording - so nothing depends on anyone remembering to log in. The dashboard is there for when you want the pattern behind it. WiseSentry Standing guard over every customer call · Illustrative example SERVICE FAILURE · HIGH Why this was flagged: The customer rang for the third time about an approval email that never arrived. A callback was promised on the previous call and did not happen. The booking has now been cancelled. Call time Tue 12 Aug, 10:43 am Direction Inbound Agent Sample Agent (ext 205) Sentiment 82 / 100 Note: your phone system scored this call 82/100 for sentiment — i.e. positive. Sentiment-threshold monitoring would not have surfaced it. WiseSentry flagged it on content. ### Why that note matters Everyone on that call behaved well. The agent was polite and helpful. Nobody raised their voice. The score reflects the tone accurately — and completely misses that a customer has now been let down three times. ### What the manager gets - A plain-English reason, not a score - The full conversation, separated by speaker - A link to play the recording - Any follow-up calls on the same matter, attached to the same alert Comparison ## Three ways to review calls Most businesses do the first. Some buy the second. Only the third finds a polite call that went wrong. Manual QA sampling Sentiment thresholds WiseSentry Calls covered 1–2% 100% 100% Finds a calm call that went wrong Only if it happens to be sampled No — scores it positive Yes Separates your fault from a supplier's Depends on the reviewer No Yes, by design Groups repeat calls on one matter No No Yes Consistency Varies by reviewer and by day Consistent, but wrong Consistent Cost per 1,000 calls Hours of a team leader's week Low Low Arrives Weeks later, if at all Immediately Within minutes of the call ending How it works ## Set it up once — then it watches your back ### Reads every call WiseSentry transcribes and reviews each recorded call from your phone system, automatically, in the background. ### Judges it against your policy A purpose-built classifier decides whether the business let the customer down — and leaves routine calls alone. ### Checks its own work A second, independent pass re-reads borderline calls to suppress false alarms before anyone is emailed. ### Alerts the right manager Each department sees only its own calls. Managers get the reason, the transcript and the recording. The full pipeline → The rule that keeps it credible ## It only flags when you are the problem An alerting tool that cries wolf gets filtered to a folder within a fortnight. WiseSentry raises a call only when the business caused the problem, or its response to someone else's problem was inadequate. ### Flagged - A booking error your team made and did not fix - A promised callback that never came - Your own system or portal failing the customer - A customer left without help at the end of the call - Repeat contact about the same unresolved matter ### Not flagged - An airline cancelling a flight, handled normally - A hotel or supplier's mistake your team is fixing - Weather, strikes, airport closures - A customer venting about a third party - Ordinary trade-to-trade calls between businesses See the full policy → Data & sovereignty ## Your customers' conversations stay in Australia 🇦🇺 ### Stored and processed in Australia Recordings, transcripts and analysis stay onshore for their whole life. Nothing is shipped offshore for processing. 🔒 ### No third-party AI Call content is never sent to OpenAI, Google or Anthropic. The transcription and language models run in Australia, under our control. ✂️ ### Sensitive data redacted Card numbers, dates of birth and licence numbers are removed from transcripts automatically. Stated plainly: redaction applies to transcripts . If a card number is read aloud, it still exists in the original recording — exactly as it does in your phone system today. We would rather tell you that than let your IT team discover it. Read the detail → ## See WiseSentry read your own calls We will run it across a sample of your recordings and show you what it finds. If it finds nothing worth acting on, that is a useful answer too. Book a demo → Call 1300 321 083 --- # Why sentiment misses it Home / Why sentiment misses it The core argument Why sentiment analysis misses your worst calls Because the worst calls are handled beautifully. A capable agent stays warm and professional while your business quietly fails the person on the other end. Tone scoring hears the warmth. It cannot hear the failure. The measurement ## We checked this properly, on real calls Over 30 days at an Australian corporate travel agency, WiseSentry reviewed every recorded call in scope and raised 45 to management. We then looked at what the phone system's own sentiment score had said about those same 45 calls. 2,349 calls reviewed 45 raised to a manager 82 median sentiment of those 45 3 scored 40 or below On a 1–100 scale where higher is friendlier, the calls that genuinely needed a manager sat at a median of 82 . Only three of the forty-five fell into the range a "low sentiment" alert would catch. Put the other way round: if this business had been monitoring sentiment and acting on scores below 40, it would have seen 3 of its 45 real problems — and would have been reassured by the other 42. Why this happens ## Good service and a bad outcome sound identical ### The agent is doing their job They apologise, they stay calm, they promise to chase it. Every acoustic and lexical signal a sentiment model looks for says "this is going fine". ### Australians under-complain "No worries", "that's alright", "no problem at all" — said by someone who will quietly move their business elsewhere next quarter. ### The failure is in the facts A promised callback that never came is a fact about the past , not a tone in the present. No amount of listening to how it was said reveals it. ## And the inverse is just as expensive Sentiment doesn't only miss real problems — it manufactures false ones. A customer swearing cheerfully about an airline strike, to an agent who is fixing it brilliantly, will score badly. Chase that alert and you have spent a manager's afternoon on a call where everyone did well. Do that a few times and the alerts get filtered to a folder — which is how monitoring tools die. Comparison ## Where each approach actually looks Signal What it can see What it cannot see Sentiment score Tone, profanity, emotional intensity A calm conversation in which the business failed Keyword spotting Named words — "cancel", "complaint", "manager" The same meaning expressed in any other words QA sampling Deep detail — on the 1–2% of calls reviewed The other 98%, which is where the problem usually is WiseSentry What was actually said, judged against whether your business let the customer down Anything not said on a recorded call WiseSentry reads the transcript of the conversation. It does not analyse voice acoustics, stress or cadence — the signal it uses is meaning , not tone. What we found instead ## One theme, not forty-five scattered problems When you review on content rather than tone, the flags cluster — and the cluster is the actionable finding. In that month, 36 of the 45 flags were the same category of failure : service breakdowns, overwhelmingly traced back to one internal system that kept letting customers down in the same way. Not forty-five bad calls. One problem, surfacing forty-five times. That is the difference between a monitoring tool and a management tool. A score tells you a call felt bad. A pattern tells you what to go and fix on Monday. Being straight with you ## What we have not measured We can tell you what WiseSentry found. We cannot yet tell you what it missed. Precision — whether the calls it raises are genuinely worth raising — has been checked against a client's own verdicts, call by call. Recall — the share of all real problems that it catches — has not been measured, because doing so honestly means a human listening to a large random sample of calls that were not flagged. Any vendor quoting you a recall figure without describing that exercise is guessing. We would rather say so. ## Run it over your own calls The fastest way to test any of this is on your own recordings. We will show you what it finds and what it ignored. Book a demo → What gets flagged → --- # What gets flagged Home / What gets flagged The policy What gets flagged — and what deliberately doesn't The hardest part of call monitoring is not finding problems. It is refusing to report the ones that are not yours. WiseSentry applies one overriding test, and it is the reason managers keep reading the emails. The overriding test ## Did this business let the customer down? A call is raised only if one of two things is true. ### (a) You caused it A mistake, a wrong booking, a missed deadline, an unissued ticket, an instruction ignored, a promise broken — or your own system or portal failing the customer. ### (b) Your response was inadequate Someone else caused the problem, but the customer was left without help: no callback, no follow-up, no answer, no owner. Everything else is noise — however dramatic it sounds. An airline cancelling a flight, a hotel losing a booking, a strike, a storm, an airport closure. If your team is handling it normally, no alert is sent. That is not a limitation; it is the product working. Worked examples ## Same situation, opposite verdicts The distinction is easiest to see side by side. These are illustrative composites, not real customer calls. What happened on the call Verdict Why Airline cancels a flight. Agent rebooks the customer on the next service while they wait. Not flagged Someone else's problem, handled properly. Airline cancels a flight. Agent promises to call back with options. No callback is made; customer rings again two days later. Flagged Limb (b) — the response was inadequate. Customer cannot log in to your booking portal, so rings and is helped manually. Flagged Limb (a) — your system failed them, even though the agent recovered it. Customer is annoyed about their bank, and says so at length. Agent is warm and helpful. Not flagged Frustration aimed at a third party. Third call this fortnight about the same unresolved booking. Flagged Repeat contact on one matter is a failure signal in itself. Your staff member calls a supplier to chase something on a customer's behalf. Not flagged Trade-to-trade. Not a customer conversation. Categories ## How flags are labelled Every alert carries a category and a severity, so a manager can triage the inbox at a glance. ### Service failure The business did not deliver what it said it would. ### Booking or order at risk Something the customer has paid for or committed to is in jeopardy. ### Complaint A genuine complaint about your service — not about a third party. ### Escalation The customer has asked for a manager at your business. ### Duty of care Someone's welfare or safety is in play and your business has not met its obligation. ### Unresolved The call ended with the customer still exposed and no owner assigned. Categories are tuned per industry. The set above reflects a corporate travel deployment; yours would be defined with you during setup. Keeping it quiet ## How false alarms are kept down ### A second opinion Every candidate flag is re-read by an independent verification pass that asks a neutral question — not "is this flag correct?", which any model will simply agree with. ### Deterministic gates Some suppressions are rules, not judgement — personal calls, trade calls, calls too short to contain a conversation. Rules do not have off-days. ### One matter, one alert Follow-up calls attach to the original alert rather than generating a fresh one, so a single problem never floods an inbox. ### Reviewed against your verdicts During a pilot, you tell us which flags were right. Those verdicts tune the policy — on your definitions, not ours. In the first reviewed batch of a live deployment, four calls were raised and the client confirmed all four were worth their attention. Small numbers, honestly reported — but that is the bar: a manager should never regret opening the email. ## Define the policy on your own calls The rules above are a starting point. What counts as "your fault" in your business is a conversation worth having properly. Book a demo → Why sentiment misses it → --- # How it works Home / How it works The pipeline From recording to the right manager's inbox Seven steps, all automatic. Nothing to install on a desk, no change to how your team answers the phone. The alert comes to the manager by email, and the same calls are waiting in the dashboard whenever you want the wider picture. ## The seven steps ### Your phone system records Nothing changes here. WiseSentry works with the call recordings your system already produces. ### Transcription Audio becomes text on infrastructure we control in Australia. The audio is never sent to a third-party AI provider. ### Speaker separation The two sides of the call are separated so the transcript reads as a conversation — who said what, in order, with timestamps. ### Policy review A language model reads the whole conversation and judges it against a single question: did this business let the customer down? ### Verification A second, independent pass re-reads anything borderline and suppresses it if it does not hold up. This is the step that keeps the inbox trustworthy. ### Incident grouping A call is held briefly after it ends, so follow-up calls about the same matter attach to the same alert instead of arriving as separate emails. ### The alert An email to the manager responsible, with the reason, the full transcript, the call metadata and a link to play the recording. Timing ## How quickly you hear about it Minutes, not seconds — and that delay is deliberate. WiseSentry waits until a call has been over for a short settling period before it decides anything. That is what allows a transfer, a callback, or a second call about the same problem to be recognised as one matter and delivered as one alert with the follow-up legs attached. We do not claim real-time. A tool that emails the instant a call ends would send you three separate alerts about the same customer within the hour. The settle window is a feature, and it is measured in minutes. Departments ## Each team sees only its own calls Most businesses of any size are several businesses in a trench coat — branches, queues, product lines, states. WiseSentry is organised the same way. Calls are scoped to a department, alerts route to that department's manager, and access to the dashboard is granted per department. A branch manager sees their branch. A national operations lead can be granted several, or all. - One department to start with, more added as you roll out - A user can hold one department, several, or the whole organisation - Nobody sees another client's calls, ever — enforced at the query, not by convention What you need ## Prerequisites ### Call recording turned on WiseSentry reviews recordings. If a call is not recorded, it cannot be reviewed — and you should be telling callers you record, regardless. ### A supported phone system 3CX today. Other platforms on request — talk to us about what you run before assuming either way. ### Someone to receive alerts A named manager per department. An alert nobody owns is an alert nobody acts on. Integrations → ## See the pipeline run on your calls Book a demo → Call 1300 321 083 --- # Data and security Home / Data & security Data & sovereignty Where your customers' conversations actually go Call recordings are among the most sensitive data a business holds. This page answers the questions your IT and compliance people will ask — including the ones with awkward answers. Residency ## Stored and processed in Australia ### Onshore, end to end Transcription and analysis run on a machine Wise Cloud purchased and Recordings, transcripts and derived analysis are stored and processed in Australia for their whole life - not just at rest, and not just in transit. ### No third-party AI providers Call content is never sent to OpenAI, Google or Anthropic. The speech and language models run in Australia, under our control. ### Nothing crosses a border Because there is no external AI API in the path, there is no cross-border disclosure to assess under Australian Privacy Principle 8. This is worth being precise about, because most call-analytics products are a friendly interface over a US AI API. If your provider cannot tell you which company's servers processed a given call, the honest answer is that it left the country. Redaction ## What is removed, and what is not Being exact here matters more than sounding impressive. Transcript Original recording Card numbers Redacted Unchanged Dates of birth Redacted Unchanged Licence numbers Redacted Unchanged Stated plainly: if a customer reads a card number aloud, that audio still exists in the recording — exactly as it does in your phone system today. WiseSentry does not modify your recordings. We would rather you heard that from us than found it in a penetration test. This is why recording links in alerts expire, and why access is logged. See below. Access ## Who can see what - Per-department access. A user is granted specific departments of their own organisation — one, several, or all of them. - Default deny. An account with no grant sees nothing, not everything. - No cross-client access, ever. Isolation is enforced in the database query on every read, not by convention or by a shared key. - Authenticated sign-in in front of the dashboard, with no password for your staff to choose badly. - Access logging. Who opened which call, and when. Compliance ## Australian obligations ### Privacy Act 1988 & the APPs Call recordings containing personal information are covered. WiseSentry is a processor of data you already hold — it does not introduce a new collection. ### APP 8 — cross-border disclosure Not engaged by WiseSentry's processing, because no call content leaves Australia. Your own recording retention remains your obligation. ### You must still tell callers Recording calls requires notification, and that duty is yours, not ours. WiseSentry reviews recordings you are already lawfully making. ### New Zealand For NZ operations, the Privacy Act 2020 and IPP 12 apply to the same questions. Ask us how a trans-Tasman deployment is structured. This page is a plain-English summary, not legal advice. Your obligations depend on your industry and your own recording practices. ## Send us your security questionnaire We would rather answer it properly up front than discover a blocker after a pilot. Get in touch → Call 1300 321 083 --- # For corporate travel Home / For corporate travel Corporate travel A business built on promises made by phone "I'll get that approved for you." "I'll call you back before close." "That's confirmed, you're all set." Corporate travel runs on commitments given verbally, at speed, under pressure — and the ones that quietly fail are the ones that lose accounts. Why this industry ## The failure mode is specific — and invisible WiseSentry has been running in production on an Australian corporate travel agency. The patterns below are what a month of real calls actually looked like. ### Approval chains that break silently A booking sits pending approval. The notification never reaches the approver. Nobody notices until the booking auto-cancels — often the night before travel. ### Promised callbacks that don't happen The agent means it. The shift ends. The traveller rings again two days later, and the second call sounds just as polite as the first. ### Suppliers chasing your customer A hotel rings the traveller directly for card details your team never supplied. The first your management hears of it is at renewal. In one 30-day review, 36 of 45 flagged calls were the same category of failure — service breakdowns concentrated around one internal system. Not forty-five unrelated bad days. One fixable problem, surfacing over and over. Tuned for the trade ## What the policy already understands Generic call analytics gets travel wrong in predictable ways. This deployment has been corrected on real calls for each of them. Situation What a generic tool does What WiseSentry does Agent rings an airline's trade desk to chase a ticket Flags it — sounds like a frustrated customer call Recognises trade-to-trade and ignores it Traveller vents about a hotel's rude staff Flags it as a complaint against you Identifies the criticism is aimed elsewhere A flight is cancelled and your agent rebooks it well Flags it — high drama, unresolved language Not your fault, handled properly — no alert Caller asks for "your manager" — meaning their own office manager Flags as an escalation Checks whose manager before escalating Your booking portal fails, agent fixes it manually Sees a helpful call, no flag Flags it — your system failed the customer Each row reflects a real correction made during a live deployment, on the client's own verdicts. Duty of care ## The calls that are about more than money Corporate travel carries an obligation to the traveller, not just to the account. A traveller stranded overnight, housed somewhere unsafe, or left without a booking in a foreign city is a duty-of-care matter — and one your client's HR team will ask about. WiseSentry treats welfare separately from commercial risk, and raises it when your business has not met its obligation. Of the categories in a live deployment, this is the smallest by volume and the most defensible to put in front of a board. ## Run it over last month's calls The most convincing demo is your own recordings. We will show you what a month of your calls contains — including the calls that scored well. Book a demo → Call 1300 321 083 --- # Integrations Home / Integrations Integrations What WiseSentry plugs into WiseSentry reviews the call recordings your phone system already makes. No new hardware, no handset changes, and nothing for your team to do differently. ## Supported today ### 3CX Fully supported and in production. WiseSentry reads recordings and transcripts directly from your 3CX instance, on-premises or hosted, and writes results back where your team already looks. Wise Cloud is a 3CX specialist — we run and support 3CX estates as our core business, which is why this integration is the deepest one we have. ### Everything else Ask us. If your platform can produce call recordings and let us reach them, the rest of WiseSentry is platform-agnostic — the transcription and review layer does not care where the audio came from. We would rather scope your platform honestly than list a logo we have not actually connected. Why this page is short. Plenty of vendors list twenty integrations and support three. We list what is in production. If yours is not here, tell us what you run and we will give you a straight answer about effort and timing. Related ## Transcripts into your CRM Separately from WiseSentry, Wise Cloud can sync call transcripts into the systems your team already works in — Salesforce, HubSpot, Zoho or Microsoft 365. That is a different product to this one. WiseSentry is about surfacing the calls that need a manager; transcript sync is about putting the record where the account owner will see it. They work well together, and neither requires the other. See Wise Cloud's other services → ## Tell us what you run Two minutes on the phone will tell you whether this is straightforward for your setup or not. Get in touch → Call 1300 321 083 --- # FAQ Home / FAQ FAQ Questions, answered straight Including the ones with awkward answers. If something here is missing, ring us and ask — we would rather tell you now than after a pilot. ## The basics What is WiseSentry? WiseSentry reviews every recorded call your business makes and alerts the responsible manager when your business has let a customer down. It reads the conversation and judges the substance, rather than scoring the tone. How does it decide which calls to flag? One overriding test: did this business cause the problem, or was its response inadequate — no callback, no follow-up, customer left hanging? Everything else is left alone, including dramatic problems caused by airlines, hotels, suppliers or weather that your team is handling normally. Full policy → Does it work with my phone system? 3CX is fully supported and running in production. Other platforms are assessed on request — if your system produces call recordings we can reach, the rest of WiseSentry is platform-agnostic. We would rather scope it honestly than list a logo we have not connected. Integrations → Who receives the alerts? A named manager per department. Calls are scoped to a department, and dashboard access is granted per department too — a branch manager sees their branch, an operations lead can be granted several or all. How much does it cost? Pricing is tailored to your call volume and number of departments. Book a demo and we will put a figure together against your actual usage rather than a made-up tier. ## The awkward ones Is it real-time? No — and that is deliberate. A call is held for a short settling period after it ends so that a transfer, a callback or a second call about the same matter can be recognised as one incident and delivered as one alert. Alerts arrive within minutes, not seconds. A tool that emailed instantly would send three alerts about the same customer in an hour. Does it analyse voice tone, stress or emotion? No. WiseSentry reads the transcript . There is no acoustic analysis of cadence, volume or stress. The signal it uses is meaning, not tone — which is precisely why it finds calls that sound perfectly pleasant. Why that matters → Does it redact credit card numbers from recordings? From transcripts , yes — card numbers, dates of birth and licence numbers are removed automatically. From the audio , no. If a number is read aloud it remains in the original recording, exactly as it does in your phone system today. WiseSentry does not modify your recordings. Detail → How accurate is it? Precision has been measured against a client's own call-by-call verdicts — in the first reviewed batch, four calls were raised and all four were confirmed worth attention. Recall has not been measured. Establishing it honestly means a human listening to a large random sample of calls that were not flagged. Any vendor quoting you a recall figure without describing that exercise is guessing. Will it replace our QA team? No. It changes what they spend their time on. Instead of sampling 1–2% of calls at random and hoping, they review the calls that a policy has already identified as worth reviewing — and they still make the judgement about what to do. Do we need to tell callers we are recording? Yes, and that obligation is yours regardless of WiseSentry — it reviews recordings you are already lawfully making. If you are not currently recording, that is the first conversation to have, and it is a legal one before it is a technical one. What happens to our data if we stop using it? Ask us this during the pilot and get the answer in writing, from us or from anyone else you evaluate. Ours: your recordings never leave your phone system's control, and the analysis we hold about them is deleted on request. ## Ask us something harder Get in touch → Call 1300 321 083 --- # About Home / About About Built by people who run phone systems WiseSentry comes from Wise Cloud Technology, an Australian telephony specialist. We did not start with an AI product looking for a problem — we started with customers whose phone systems we already run, and a problem we kept watching them have. ## Why it exists Every business we support records its calls. Almost none of them review those recordings in any meaningful way. The recordings sit there — a complete, honest account of how the business actually treats its customers — and nobody has the hours to listen. The tools on the market scored calls for sentiment. When we tested that idea against real recordings, it did not work: the calls that genuinely needed attention scored well , because good agents stay calm while the business fails someone. So we built something that reads what was said instead. WiseSentry is deliberately narrow. It does not do agent scorecards, workforce management, or predictive analytics. It answers one question — which calls should a manager look at today? — and tries to answer it well enough to be trusted. ## How we build ### Onshore by design We run the models ourselves in Australia rather than renting a US API. It costs more effort, and it is the only way to answer the residency question with a straight yes. ### Measured, not asserted Every claim on this site traces to something we counted on real calls. Where we have not measured something — recall, for instance — we say so. ### Corrected by clients The flagging policy has been rewritten repeatedly because a client told us a flag was wrong. That feedback loop is the product. ## Talk to the people who built it No SDR script. You will speak to someone who can answer a technical question properly. Get in touch → Call 1300 321 083 --- # Fact sheet Home / Fact sheet Fact sheet WiseSentry, stated plainly A flat, unambiguous summary for anyone — human or machine — that wants the facts without marketing language. Also available as /llms.txt and /llms-full.txt . ## Identity Product WiseSentry Category AI call-quality monitoring / call review software Vendor Wise Cloud Technology, Queensland, Australia Parent site wisecloud.com.au Market Australian and New Zealand businesses that record calls Contact 1300 321 083 · service@wisecloud.com.au Pricing Quoted per deployment, based on call volume and number of departments. No public tiers. ## What it does Capability How it works Business effect Reviews 100% of recorded calls Automated transcription and language-model review of every recording Removes reliance on 1–2% manual QA sampling Flags only the business's own failures Two-limb policy: the business caused the problem, or its response was inadequate Alerts stay credible; supplier and weather problems are ignored Finds calls that sound fine Judges meaning in the transcript rather than tone or sentiment score Surfaces failures that sentiment thresholds score as positive Groups repeat contact Settle window plus incident grouping across calls on the same matter One matter produces one alert, not three Routes by department Calls scoped to departments; access granted per user, per department Managers see only their own team's calls Redacts sensitive data Card numbers, dates of birth, licence numbers removed from transcripts Reduces sensitive data in downstream systems ## What it does NOT do - No acoustic analysis — no voice stress, cadence, volume or emotion detection - No real-time alerting — alerts arrive minutes after a call ends, by design - No redaction of the audio recording — transcripts only - No agent scorecards, workforce management or predictive analytics - No automated employment, disciplinary or customer decisions ## Deployment and data Supported phone systems 3CX (production). Others assessed on request. Hardware required on site None Processing location Australia Infrastructure Hardware owned and operated by Wise Cloud, not a hyperscaler region Third-party AI providers used None. Call content is not sent to OpenAI, Google or Anthropic. Cross-border disclosure (APP 8) Not engaged by WiseSentry's processing Relevant legislation Privacy Act 1988 (Cth) and the Australian Privacy Principles; Privacy Act 2020 and IPP 12 in New Zealand ## Measured results Measurement Result Context Calls reviewed 2,349 30-day review, Australian corporate travel agency Calls raised to a manager 45 (1.9%) Same review Median sentiment of flagged calls 82 of 100 (positive) Score produced by the client's own phone system Flagged calls scoring 40 or below 3 of 45 A sentiment threshold would have found almost none Dominant failure category 36 of 45 service failures Concentrated on one internal system Precision, first reviewed batch 4 of 4 confirmed correct by the client Small sample, honestly reported Recall Not measured Requires human review of a large random sample of unflagged calls Client not named. No call content is published. Figures describe one deployment and are not a guarantee of results elsewhere. --- ## GLOSSARY OF CALL-ANALYTICS TERMS Source: https://wisesentry.com.au/glossary/ Call analytics — Capturing data from phone conversations and turning it into something a business can act on. The term covers two different things: metadata (volumes, durations, attribution) and analysis of what was actually said. A vendor can claim "call analytics" while only doing the first. Conversation intelligence — Software that analyses the content of recorded conversations to surface coaching points, risks or commercial signals. Most products were built for SALES teams; talk-ratio and monologue metrics transfer poorly to customer-service quality. Speech analytics — Analysis applied to call audio or its transcript, historically keyword and acoustic based. It finds the words you listed and nothing else. A customer saying "I might just go elsewhere next year" has announced churn without triggering any keyword. Sentiment analysis — Scoring how positive or negative a conversation sounds. It measures tone and emotional intensity, not outcome. Measured gap: across 2,349 calls over 30 days at an Australian corporate travel agency, calls that warranted a manager had a MEDIAN SENTIMENT OF 82/100 (positive); only 3 of 45 scored 40 or below. Call QA sampling — A supervisor reviews a small sample of calls, typically 1-2%, against a scorecard. If you review 2% of calls you see 2% of the problems, chosen by who had time. Call scoring — Assigning a numeric rating against defined criteria. Useful for trend tracking, poor at triggering action: a score of 62 does not tell a manager what to do on Monday. Speaker diarisation — Separating a recording into who spoke when. Without it, "I'll call you back this afternoon" is a floating sentence — customer request or agent promise are opposite meanings. Stereo recordings diarise reliably; mono must be inferred and can collapse. Escalation detection — Identifying calls where the customer is asking for a manager or signalling the matter has outgrown the handler. Most Australians escalate without using the word "manager". Must distinguish escalation AT your business from venting about a third party. Repeat contact rate — The share of matters needing more than one contact. A direct measure of failure regardless of tone. Requires linking calls to a MATTER, not a number or an agent. PII redaction — Removing personal or payment data from call records. Ask which layer: WiseSentry redacts TRANSCRIPTS. Spoken digits remain in the original audio, as they do in your phone system today. Audio redaction is a separate and harder capability that WiseSentry does not claim. Which technique finds which problem: - Metadata / call tracking: finds volumes and attribution; cannot see what was said. - Keyword spotting: finds the exact listed words; cannot see the same meaning phrased otherwise. - Sentiment scoring: finds raised voices and profanity; cannot see a calm call that went wrong. - QA sampling: finds deep nuance on the 1-2% reviewed; cannot see the other 98%. - Content review of every call (WiseSentry): finds whether the business kept its promise, in whatever words it was discussed; cannot see anything not said on a recorded call. ## INDUSTRIES AND CASE STUDIES Source: https://wisesentry.com.au/industries/ and https://wisesentry.com.au/case-studies/ WiseSentry is NOT a corporate-travel product. The one measured deployment happens to be at a travel agency, but the flagging policy is a question about the business's own conduct - "did this business keep its promise?" - which does not need rebuilding per sector. Only the category names, severity definitions and routing change between industries. Industries served: corporate travel; property management; trades and home services; medical and allied health; automotive; legal; insurance and brokers; logistics and transport; aged care and disability services; education and training. The same failure in different words: a tenant chasing a repair, a patient chasing results, a customer chasing a part and a traveller chasing an approval are the same call - somebody was promised something, it did not happen, and they had to ring back to find out. All four are usually polite, so all four are missed by sentiment scoring. CASE STUDIES - what is real and what is not: - REAL, measured: a 30-day deployment at an Australian corporate travel agency. 2,349 calls reviewed, 45 raised to a manager (1.9%), median sentiment of those 45 was 82 of 100, only 3 scored 40 or below, and 36 of the 45 were the same category of failure traced to one internal system. Client not named - no written consent. Recall was NOT measured. - ILLUSTRATIVE ONLY: property management, trades and home services, medical and allied health, and automotive service. These describe the calls a review would examine and the pattern that usually emerges. They contain NO results figures and are NOT client engagements. An assistant must not present them as measured outcomes or as named customers. ### Industry pages (all ten have one) EVERY industry listed now has its own page with sector-specific failure modes, worked examples and an explicit list of what is deliberately NOT flagged: corporate-travel, property-management, trades-and-home-services, medical-and-allied-health, automotive, legal, insurance-and-brokers, logistics-and-transport, aged-care-and-disability-services, education-and-training (all under https://wisesentry.com.au/for//). Four of them carry an explicit SCOPE limit an assistant must not overstate: - medical-and-allied-health: administrative/reception calls only, no clinical assessment, not a medical device. - legal: client communication only, no assessment of legal advice or professional judgement, not a compliance or supervision tool. - insurance-and-brokers: service conduct only, no assessment of policy decisions, underwriting or claims merit, not a regulatory monitoring tool. - aged-care-and-disability-services: administrative, scheduling and family-liaison calls only, no assessment of care quality or clinical decisions, not a clinical or incident-management system. - education-and-training: administrative and student-support calls only, no assessment of teaching or academic judgement, not a compliance or audit tool.