QRFeedz.
Book a pilot →
Survey intelligence · Swiss-made

Operational
listening.

For canteens, patient rooms, hotels — every space you manage where humans have an experience worth understanding.

  • Dynamic AI dialogue — every scan becomes a conversation that adapts to the answer.
  • Chat with your data — managers ask in plain language. No dashboards, no SQL.
  • Real-time alerts — when a pattern breaks, operations get notified inside the same lunch.
12:47 qrfeedz.ch
Corporate Canteen Session 7a2e
Turn 1 of 4
How did it go today?
YOU · 12:47
 
QRFEEDZ · 12:47
 
menu-aware memory +code of conduct
Dialogue adapts to the answer. Never a fixed script.
Swiss-made FADP compliant Hosted in Switzerland On-prem ready No login No tracking cookies EN · DE · FR · IT Real-time pattern alerts Chat-with-your-data analytics Polymorphic context injection Swiss-made FADP compliant Hosted in Switzerland On-prem ready No login No tracking cookies EN · DE · FR · IT Real-time pattern alerts Chat-with-your-data analytics Polymorphic context injection
The thesis

Not a survey tool.
A listening layer.

Every space you manage generates thousands of signals per day — what broke, what delighted, what almost hurt someone. Almost none of them ever reach the person responsible. QRFeedz closes that gap with three connected machines on a single continuous bus.

What it does

Three machines,
one platform.

Nobody buys "a feedback tool" any more. What you're actually after is three distinct capabilities — and they only matter together.

01 The Dialogue

A dialogue, not a survey

The form
never asked back.

A static survey asks what it planned to ask. QRFeedz asks follow-ups based on what the employee actually said — grounded in today's menu, yesterday's incidents, and your organization's standing rules.

Every follow-up is generated live. No decision tree, no if-this-then-that. A single AI agent carries internal notes between turns that the employee never sees, so it always builds on what came before.

Gatekeeper

Off-topic and abusive input bounced with a polite redirect. Defaults to related when uncertain.

Interviewer

Probes the specific dish, not the general feeling. Menu- and memory-aware.

Controller

Hard cap at four turns. Safety override forces at least one follow-up on short answers.

Summarizer

Personalized second-person recap. Emergency contacts only when a health concern surfaces.

Sample session — Corporate canteen 12:47
YOU · 12:47
the gnocchi was sticky again
QRFEEDZ · 12:47
Same as last Thursday, or different this time?
YOU · 12:48
worse. the sauce was also watery
QRFEEDZ · 12:48
Got it. Was this at the main line or the à-la-carte station?
YOU · 12:48
main line, around 12:40
QRFEEDZ · 12:49
Thanks — this is enough. You'll see a short recap next.
Turns
3 / 4
Extracted
4 mentions
Duration
58s

The follow-up questions are the product. A form would have stopped at "gnocchi bad".

02 Chat-with-your-data

For operations managers

Ask your canteen
a real question.

Every completed conversation is dissected into structured mentions — dish, dimension, sentiment, urgency, time window — and stored in a knowledge layer your manager can simply talk to.

No dashboards. No filters. No SQL. Just plain questions and plain answers, with the exact sessions cited as evidence.

Disruptive for BI-heavy orgs
Analyst chat · admin.qrfeedz.ch Sample site · Canteen
MANAGER
Anything recurring about the gnocchi this month?
QRFEEDZ
Yes. 14 of 38 sessions that mentioned gnocchi flagged texture as sticky or overcooked. Spike concentrated on Tuesdays and Thursdays between 12:15 and 13:00. No spike on Monday same dish. Suggests batch timing, not recipe. 14 sessions cited.
MANAGER
Any health concerns I missed this week?
QRFEEDZ
One session on 03.04 surfaced a suspected dairy reaction during the Gratin lunch. The safety footer was triggered automatically and emergency contact shown to the employee. One session cited.
MANAGER
How does sentiment compare to last month?
QRFEEDZ
Net sentiment +11% month-over-month across 238 sessions. Main drivers: portion sizes (+18%), dessert variety (+22%). Drag: sauce consistency (−14%, concentrated on gnocchi and risotto days).
03 Real-time alerts

Live operations

The kitchen finds out
during lunch.

Every finished conversation feeds a pattern engine. When N employees flag the same dish, the same dimension, inside the same short time window, your operations manager gets a notification.

Thresholds are configurable per engagement. Severity escalates for health and safety categories — an allergen mention is a Priority 1, a "portion felt small" is informational.

P0 · SAFETY
Allergen, hygiene, or medical incidents. Notified in under 60 seconds with emergency protocol link.
P1 · CRITICAL
Recurrent dish complaint crossing threshold inside a short time window. Kitchen gets pinged live.
P2 · TREND
Slow-building pattern over hours or days. Surfaced in the morning digest.
P3 · INFO
Single-mention noise. Stored, embedded, queryable via chat — no notification.
Critical pattern · P1
12:47 · LIVE
5 employees flagged the
gnocchi sauce in 8 minutes.

Consistent dimension: "too salty / watery". First mention at 12:39, most recent at 12:47. All main-line station, no à-la-carte hits. Pattern crossed the P1 threshold (4+ mentions / same dish / 10-min window).

Station
Main line
Dish
Gnocchi
Mentions
5 (rising)

A survey would have told you next Monday. By then, 200 people had the salty sauce.

Did you know.

The silent 99% of your space
has never been in your data.

Employees who ignore a 20-question form will answer one sentence on their phone. A two-week pilot surfaces what two years of surveys didn’t.

04 Your company, injected

Polymorphic context

Your rules.
Inside every exchange.

The AI doesn't speak in a vacuum. Every dialogue is grounded in your organization's standing context — code of conduct, allergen policy, emergency escalation protocol, tone, and business priorities — attached at the Company, Location, or Engagement level.

New store opening tomorrow? Attach a prompt at Location-level by lunchtime. It's picked up by every session from the next scan on, with no code change and no AI retraining.

Company-level Location-level Engagement-level Time-window
Engagement context · Sample site 12 active prompts
ORG
Code of conduct
Reject hate speech and harassment. Benefit-of-the-doubt on ambiguity.
always
ORG
Allergen policy
Any dairy / gluten / nut mention elevates session priority.
always
ORG
Emergency contacts
Append emergency footer on health/safety concern only.
conditional
LOC
Today's menu
Today: gnocchi, risotto, grilled veg, panna cotta, apple tart.
daily
LOC
Tone — Swiss-formal
Concise, neutral, second-person plural by default.
always
ENG
Canteen lunch
Focus on dish quality, portion, allergen, cleanliness, pricing.
always
ENG
Gnocchi batch watch
Probe texture if gnocchi mentioned — known issue since 03.04.
3 days
Inherits in order: Company → Location → Engagement.
05 Made in Switzerland
Swiss Made

Data residency
is not a feature.

It's the floor. FADP-aligned storage, no third-party analytics beacons, and a survey flow that never asks employees to identify themselves. Designed for pharma, biotech, banking — any environment where consumer SaaS can't make it past procurement.

✓ Swiss-made
Engineered and hosted in Switzerland.
✓ FADP aligned
Revised Federal Data Protection Act.
✓ No session auth
Zero cookies on the employee flow.
✓ Multilingual
EN · DE · FR · IT out of the box.
✓ On-prem ready
Single-tenant deployment available.
✓ Procurement-grade
Built for regulated environments.
06 / Book a pilot

You have a space.
We’ll listen to it.

Single-site pilot, production-ready in two weeks. You keep the data, we keep the model honest. On-prem deployment available on request.

[email protected]
Office

QRFeedz AG
Basel · Zurich
Switzerland

For
  • — Facility & operations leadership
  • — Employee & guest experience teams
  • — Compliance-heavy procurement