hxiq.ae/q/westgate-discharge-q3
A vanity slug per stage. Magic-link resume if the respondent pauses. Honeypot + bot protection at the edge.
AI command layer for every human experience signal
Turning human experience into intelligence.
HXIQ turns branded questionnaires, QR entry, chat-native surveys, and four reporting surfaces — Pulse, Explore, Suite, and Map — into structured feedback your operators can actually use.
// the gap
A 0–10 score
A 0–10 score
isn't a strategy.
NPS gives you a number. CSAT gives you a five-point scale. Neither tells you which property is degrading, which category is bleeding, or what to do about it before the next board review.
-
01
No context
A score has no shape. You can't see the drivers beneath it, the verbatim behind it, or the moment it broke.
-
02
No source
By the time the score lands in a dashboard, the response that caused it is buried in a CSV no operator opens.
-
03
No action
By the next review cycle, the moment is gone and the problem has compounded across the portfolio.
// a Monday morning at a multi-site operator
One operator. Three facilities.
One operator. Three facilities.
Two thousand responses last week.
This is the sequence HXIQ runs every time an operator opens the platform. The data is theirs. The questions are the same six. The path from a score to a decision compresses from quarterly to thirty seconds.
-
01 · 08:47
The health score moved.
Pulse opens to the headline KPI for the portfolio. Westgate Medical is down six points week-over-week. The discharge-experience category is the heaviest contributor. Twelve patient sessions have trended negative since Friday.
-
02 · 08:48
One filter, the whole stack moves.
Filter to Westgate Medical, last 30 days. Every chart, every export, every AI call now answers about that slice. No re-querying. No "is this the same data?" The filter signature is the contract.
-
03 · 08:49
The theme behind the dip.
Explore surfaces the cluster: "long wait at outpatient triage." Forty-two negative mentions in nine days, concentrated in the morning clinic window. Sentiment, volume, timeline — all attribute to the same root.
-
04 · 08:50
Ask in plain language.
HXIQ Pulse — ⌘K. "What changed at Westgate this month?" The assistant calls real reporting lookups, not a generative blur. Citations to the underlying responses. No raw SQL, no made-up numbers.
-
05 · 08:52
A root cause with a recommendation.
One synthesised insight: discharge delays at Westgate are cascading into bed turnover on the cardiology ward. Polarity: fix. Confidence: high. Impact: notable. Exported to a branded PDF before the morning huddle.
// capture
Feedback enters branded.
Feedback enters branded.
Stays structured.
HXIQ captures feedback the way an operator wants it captured — branded to their identity, scoped to a project, staged through a real journey. Five entry surfaces, one structured data model behind every one of them.
Per-channel codes. Track whether the waiting-room poster or the post-visit email drove the response. Same questionnaire, different attribution.
Enter your guest code
7
3
9
2
Forgot? Email a one-time code
Anonymous-to-the-respondent, attributable to the operator. Brute-force lockout after five tries. Recovery via 6-digit email code. Works for patient panels, study cohorts, or anonymous staff pulses.
How clear were your discharge instructions?
Mostly clear — the medication dosing wasn't explained well.
Conversational mode for surfaces where a form would die. Branching logic stays server-authoritative; the chat is a thin client.
-
01 · library
Write a question once. Reuse it everywhere.
The question library is versioned. Tag by category, accountability area, or industry pack. When wording changes, historical responses keep their original semantic meaning — no broken trend lines.
-
02 · stages
Real journeys. Auto-publish, auto-close, branch on conditions.
Stages run on schedule. Branching rules live in JSON on the questionnaire, not in the client. Respondents who pause get a magic-link resume — sessions live for seven days.
-
03 · branding
Instance brand by default. Project overrides where it matters.
Respondents see the operator's identity end-to-end — palette, copy, email shells, the assistant's name. Per-project overrides for sub-brands and white-label deployments. HXIQ never appears unless the operator wants it.
// read
Four ways to read the
Four ways to read the
same truth.
One filter signature follows you between surfaces. Pulse for the morning scan, Explore for the drill, Suite for the evidence, Map for the floor plan. What you select in one holds in the next — every chart, every transcript, every export answers the same slice.
-
Pulse
What changed?
The executive scan. Portfolio health, the categories that moved, the problems demanding attention right now, and the AI-surfaced root causes underneath them. Charts primary, AI shortest. The surface an operator opens at 8:47 on a Monday.
- Portfolio KPI · week-over-week delta
- Problem-first attention row
- Trend line · brushable
- AI insight pill · cached + refresh-aware
-
Explore
Why?
Sentiment split, theme clusters, per-question detail, brushable timeline. The operator's middle-of-day surface — where a 6-point drop becomes "long wait at outpatient triage, concentrated late-morning, on the cardiology ward." AI mid-depth, evidence at hand.
- Sentiment high/low split
- Theme buckets · counts · sample quotes
- Timeline · brush to filter
- Per-question drill
-
Suite
Show me.
The evidence layer. Every response, every transcript, every custom field. Export the active slice to CSV, XLSX, or print-ready PDF — branded, filter-faithful, ready for the board pack. This is what a department head opens when leadership asks "are you sure?"
- Full response ledger · per-respondent transcripts
- Custom-field columns · respondent search
- Branded CSV / XLSX / PDF export
- Filter signature follows the export
-
Map
Where on the floor plan?
Upload an SVG site plan. Assign each region to a category. Watch health colour live — green, amber, red — driven by the same scores Pulse and Explore see. Built for multi-site operators: hospital wards, clinic floors, campus buildings, retail footprints.
- SVG site plan · region-to-category mapping
- Score thresholds · green / amber / red
- Per-zone tooltips · response count or "no data"
- Floor swap · multi-property navigation
One filter signature. Stage, category, custom fields, distribution tags, date range, completion state, sentiment band, respondent search — every surface and every export reads the same hash. No "is this the same data?" arguments.
// ask
Ask in plain language.
Ask in plain language.
Get answers from your data,
not a generative blur.
HXIQ Pulse is the ⌘K assistant inside every dashboard. It doesn't write SQL. It doesn't invent numbers. It calls real reporting lookups — the same functions that paint your charts — and answers with citations back to the underlying responses. Operators stop guessing whether an AI summary is real.
-
10 · lookups
Ten governed lookup types. No raw SQL. Ever.
list_questionnaires · questionnaire_metrics · category_scores · category_detail · question_metrics · stage_funnel · nps_theme_correlations · portfolio_projects · respondent_leaderboards — the same functions that paint the charts answer the chat.
-
cite · or · don't
Every claim points to a chart, a metric, or a verbatim.
Answers carry inline citations. Click to jump to the source view with the same filter applied. If an answer can't be cited, the assistant says it can't be cited — instead of inventing one.
-
scope · role
Scoped to what the asker is allowed to see.
The assistant runs as the authenticated user. Viewers see what viewers see. Roles filter at the auth layer, not in the prompt. The AI can't accidentally surface a category a role wouldn't have opened manually.
-
kill · switches
Ten feature flags. Disable any of them instantly.
Conversational assistant, dashboard insights, per-question commentary, theme extraction, anomaly alerts, NPS–theme signals, email digests — each independently togglable from Admin → Intelligence Usage. PII redaction sits in front of every prompt.
// synthesise
Themes cluster. Clusters become causes.
Themes cluster. Clusters become causes.
Causes become a memo.
Cross-category root-cause synthesis is the new capability. HXIQ doesn't stop at "your score went down." It clusters the verbatims, finds the connections between categories that single-surface dashboards miss, grades the cause, and writes the recommendation the operator needs to act on.
-
cross · category
Finds connections single-surface dashboards miss.
A drop in discharge scores might be a staffing problem, a pharmacy problem, or a ward-coordination cascade. HXIQ correlates themes across categories, identifies the upstream cause, and writes one synthesised insight instead of three disconnected dashboards.
-
graded · always
Every cause is graded — polarity, confidence, impact.
Polarity: fix, amplify, monitor, or protect. Confidence: low, medium, high. Impact: major, notable, worth-watching. Operators see what's worth doing today and what can wait until next week.
-
anomaly · detection
Automated alerts when scores drift outside expected bands.
A worker job runs Welch's t-test, Mann–Kendall trend tests, Cohen's d, and Wilson intervals on category scores, completion rates, and NPS. When a real change crosses a confidence threshold, the operator gets a flag — not a notification storm.
-
domain · aware
One engine. The right language for every operator.
The AI infers the domain from the questionnaire — healthcare, multi-site clinical, public services, hospitality, retail, education, fitness, professional services, food service, generic — and reframes every generated sentence with the right nouns and recommend verbs. Patients, not customers. Discharge, not checkout. The platform doesn't make you translate.
// govern
Built for operators with
Built for operators with
regulatory weight.
HXIQ is designed for organisations that can't afford an AI to hallucinate, a dashboard to leak, or an audit log to have gaps. The governance surface is not an afterthought; it's the spine of the platform.
PII redacted before any LLM call
A regex layer strips emails, phone numbers, and card-pattern data before text leaves the boundary. Tokenised, replayable. Configurable per instance, default on.
Ten AI feature kill switches
Conversational assistant, dashboard insights, per-question commentary, theme extraction, anomaly alerts, NPS–theme signals, email digests, root-cause synthesis, cross-question correlation. Toggle each one off without redeploying.
Append-only audit log
Every state-changing action — who, when, before/after, IP hash, user agent. Monthly archival to S3 in immutable batches. Never cleared by the maintenance wipe.
Role-based access at the auth layer
Four roles: super_admin, admin, editor, viewer. Capabilities checked per route. The AI runs as the authenticated user — it cannot accidentally surface data the user wouldn't have opened manually.
Configurable retention sweeps
Set days-to-retain for in-progress sessions, completed sessions, and audit log per instance. Nightly cron purges; archive batches survive on cold storage if compliance requires.
Token budget · per-feature cost telemetry
Monthly LLM budget cap per instance. Per-feature, per-model, per-day token usage tracked. HXIQ-paid spend cleanly split from tenant BYOK spend on the platform admin plane.
HIPAA BAA · regional residency · DPAs
Healthcare BAA on request. Region pinning confirmed per deployment. Custom DPAs available for enterprise deployments. Encryption in transit and at rest by default.
Circuit breaker on every AI surface
If the provider has an outage, dashboards show the last cached insight with a "refreshed N minutes ago" badge. The product degrades gracefully — it never lies about being live.
Compliance paperwork — SOC 2 program in progress, GDPR-ready posture by default, residency / BAA / DPA confirmed during onboarding. Penetration testing reports available on request for enterprise deployments.
Running with operators
who can't afford theatre.
Live across aesthetic healthcare, longevity, and eco-luxury hospitality — operators measuring outcomes that are too subjective for an NPS number to capture.
Capturing patient sentiment across the 3-phase care protocol — pre-treatment diagnostics, in-procedure, and post-treatment follow-up. Aesthetic outcomes are subjective; HXIQ structures them so the clinical team sees protocol drift before it shows up in reviews.
Catches protocol drift before it surfaces in reviews
Member feedback flows in across nine longevity programs — gut health, hormone optimisation, sleep, focus, performance. Monthly check-ins and quarterly retests land in one continuous record so physicians see how each protocol is responding for executive and athlete members.
Aligns each protocol to how the member is actually responding
Captures stay quality across pre-arrival, in-stay, and post-departure stages. Property maps surface zone-level health across the lodges so operators can see which villa block is drifting before the GM asks.
Zone-level health visible across every propertyPer-project. Not per-seat.
Operators are not a metered resource.
Charge around the project and response volume. Invite every team member who needs to build, manage, or read. We don't tax operators for showing up.
The full Launch experience, free for 30 days — no credit card. Spin up your branded workspace in minutes.
- 1 active project
- No credit card required
- Full questionnaire engine + Pulse · Explore · Suite · Map
- QR entry + chat-native respondent surveys
- CSV / PDF export
- Upgrade in-app anytime
One project, one operating loop. The fastest way to stand up a branded measurement programme.
- 1 active project
- Question library + staged questionnaires
- QR entry + respondent access
- Pulse · Explore · Suite · Map reporting
- Branded chat-native respondent surveys
- CSV / PDF export
For operators running parallel programmes — multi-property, multi-clinic, or multi-line of business.
- Up to 3 active projects
- Everything in Launch
- AI insights · theme extraction · NPS–theme signals
- HXIQ Pulse (⌘K) cross-project lookup
- Reminder rules + scheduled dispatches
- Per-project palette + branded exports
For larger operators with multiple stakeholders, regional teams, and stricter access requirements.
- Up to 5 active projects
- Everything in Growth
- Role-based access + admin controls
- Audit log + activity history
- Worker-backed jobs at higher volume
- Priority support
For operators with regulatory weight. Custom residency, BAA, dedicated capacity, no per-seat tax.
- Unlimited projects & responses
- Everything in Scale
- HIPAA BAA · custom DPAs · regional pinning
- OpenAI key governance · seven intelligence kill-switches
- Custom retention policy
- Dedicated launch support
All tiers include the same questionnaire engine, branded respondent flows, and exportable data. Compliance paperwork is handled during onboarding where required.
The questions operators
ask first.
01 How is this different from NPS or CSAT tools?
NPS gives you a number. CSAT gives you a five-point scale. HXIQ gives you a full operating surface: reusable questions, staged questionnaires, branded chat-native respondent flows, Pulse · Explore · Suite · Map reporting, respondent transcripts, CSV/PDF exports, and optional AI analysis on the same filtered data.
02 How long does setup typically take?
Onboarding starts by configuring a real project with your branding, your questions, stages, respondents, QR assets, and reporting. Most operators are running a real measurement loop within a working day. Enterprise timelines depend on residency, compliance paperwork, and launch support.
03 Where is respondent data stored?
The platform is designed around instance and project-level configuration for data handling, retention, and regional posture. Exact hosting regions are confirmed during onboarding and deployment.
04 Can we white-label the respondent surveys?
Yes. Instance branding and project-level overrides cover organisation name, palette, logos, welcome copy, thank-you messaging, and the respondent assistant persona (default HXIQ Pulse). Surveys show your brand; login and entry pages include a subtle Powered by HXIQ attribution.
05 What AI features can we govern, and how granularly?
Ten independently togglable feature flags: conversational assistant, dashboard insights, per-question commentary, per-questionnaire summary, per-project trend, respondent summary, anomaly alerts, free-text theme extraction, cross-question correlation, and email digest narratives. Each one is on/off from Admin → Intelligence Usage. PII redaction runs in front of every prompt and is configurable per instance.
06 How does HXIQ Pulse avoid invented answers?
The conversational assistant cannot run raw SQL. It calls one of ten governed reporting lookups — the same functions that paint the dashboard charts. Each answer carries inline citations to the underlying chart, metric, or verbatim. If an answer can't be cited, the assistant refuses to invent one. Permissions filter at the auth layer, not in the prompt, so a viewer cannot use the chat to reach data they couldn't open manually.
07 What does the root-cause synthesis actually do?
A worker job clusters free-text answers, identifies themes that cross category boundaries, and writes a single insight with polarity (fix / amplify / monitor / protect), confidence (low / medium / high), impact (major / notable / worth-watching), and a concrete recommendation. Below ~5 free-text responses the system shows preliminary signals instead of a synthesised cause — operators see "we don't have enough yet" rather than a fabricated narrative.
08 How does pricing scale across HXIQ Launch, Growth, Scale, and Enterprise?
Tiers are gated by project count: Launch (1), Growth (3), Scale (5), Enterprise (unlimited). Every tier includes the full reading surface — Pulse, Explore, Suite, Map — and the ⌘K assistant. Enterprise adds residency pinning, HIPAA BAA, custom DPAs, and dedicated launch support. Pricing is per-tenant, negotiated against project count, response volume, and compliance scope. There's no per-seat tax.
// start the loop
Stop guessing.
Stop guessing.
Start synthesising.
A working HXIQ environment in 24 hours. Healthcare-first onboarding (BAA, residency, PII redaction) available from day one — also wired for hospitality, retail, education, and any operator with regulatory weight. Drop your work email — we'll set up the loop.
No static deck. We configure a working HXIQ environment in 24h. Got it. We'll be in touch within a working day. Something went wrong. Email us directly instead.