More feedback.
Richer feedback.
Real intelligence.
Whether you run periodic studies or operate a legacy VoC platform, you know the problem: feedback is delayed, poor and expensive. Zapio flips all three —
- DelayedWe effortlessly connect to your systems to ask in real-time, every day of the year.
- PoorAn NPS score becomes a 24-second voice note, enriched by AI follow-ups.
- ExpensiveFrom €160 a month.
OrangeTelecom
Sanlam AllianzInsurance
bpostPostal & parcels
P&V GroupInsurance
La Marocaine VieLife insurance
Le Pain QuotidienHospitalityNumbers, not promises.
reply rate, measured at an African retail bank.
Because it’s asked on WhatsApp, minutes after the visit — where your customer already is. No inbox to open, no link to click, nothing to install. Anyone without WhatsApp gets the same survey by SMS or email.
more verbatims, measured at a European telecom operator.
A conversation invites an answer; a form demands one. Replying to a message costs a customer almost nothing — so far more of them say something, and say more of it.
richer verbatims. “It was ok” becomes a story to act on.
A score gives you the sentiment. A 24-second voice note gives you the reason — and AI asks the follow-up right there in the thread, which a fixed questionnaire never can.
Find your fit
Zapio Studio
Feedback on autopilot. A report every Monday.
- 01
You send the extract, daily
Email the file to your dedicated Zapio address. We take it from there.
How much effort is it on our side?
An export and an email — a few minutes, once it’s routine. If your system can schedule the export, it becomes none.
What exactly do you need from us?
Whatever tool you use: one file per journey — branch visits, support calls — with two columns: the customer’s mobile number (or email) and the exact time. Anything else — branch, product, advisor — is optional, and lets us break the report down by it.
We don’t record who we serve — is this still for us?
Partly. Without a system that logs your customers — a CRM, a till, a loyalty system, a booking or ticketing tool, a call-centre log — we can’t reach them one by one. What we can run instead: QR codes at your locations and a feedback form on your website, so customers come to you. It works; it just asks fewer people.
- 02
Surveys fire in real time
Your questions are asked on the channel you prefer — WhatsApp, SMS or email. Short. Conversational. AI-powered.
Who writes the questions?
We do, with you: a short core — one score, one open question — and the AI follow-ups that turn a vague answer into a story. Nothing to build on your side.
Is surveying customers allowed?
Yes — asking your own customers about a service they just used is standard practice on any channel. What it needs is a line in your privacy notice: that you survey customers, on which channels, and how to opt out — we help you write it. One message, never a campaign; opting out takes one word.
What if we don’t want WhatsApp?
No problem. WhatsApp is the default because it gets the best results, but the same survey runs by SMS, email, QR code, tablet, in your app or on your website. Your channel mix, your call.
What if they don’t have WhatsApp?
We detect it and send by SMS instead — automatically. No mobile number but an email? Then by email. Nobody is skipped.
- 03
Analysis runs on its own
Every answer is read as it lands — voice notes transcribed — for what they felt, what they talked about, and what moved the score.
Is a human reading the answers?
AI reads every answer as it lands — sentiment, topics, drivers. We read the picture it draws and build your report on top of it.
Can we see the raw answers?
Yes. Every verbatim, voice notes included, is yours — we hand them over with the report.
In which languages?
Answers in English, French, Dutch, Spanish, German, Portuguese, Italian, Arabic and more than 100 others are understood as they are — spelling and grammar don’t matter, and a voice note works too.
Can we bring our past responses?
Yes — export them from your previous tool or study and they come in as the baseline, so your first report already says what changed.
- 04
You read the weekly report
What customers said, what changed, what to fix first — a PowerPoint in your inbox.
What’s actually in it?
Eight to ten slides: the scores and how they moved, the verbatims that explain them topic by topic, and the two or three things worth fixing first — ready to present as is, or forward.
Can we change the cadence?
Weekly is the default — a Monday report becomes a ritual, and rituals get read. Fortnightly or monthly on request.
What if we want to see the data live?
Then Zapio Platform is for you — the same data, in dashboards, alerts and tickets. And moving from Studio to Platform keeps everything.
Zapio Platform
Ask at every event. Close every loop.
- 01
Your systems send the events
Every interaction — a visit, a call, a delivery — reaches us as it happens. A daily SFTP drop gets you started; the API when you want true real time.
Is this a big IT project?
No. A daily SFTP drop gets you live; one API call per customer event when you want real time. Most start with the file.
What do you need per event?
Three things: a way to reach the customer (mobile or email), the journey it belongs to, and the exact time. Anything else — branch, product, agent — is optional, and becomes a filter in your dashboards.
Who builds the connection?
We do, with your IT. You point us at the export or the API; we handle the mapping, the tests and the go-live. A few hours of their time, not a project.
- 02
Surveys fire in real time
The moment the event lands, the survey goes out — on WhatsApp by default, or the channel you choose. Short. Conversational. AI-powered.
Can we keep our questionnaire?
Yes. Your NPS, CSAT or CES questions run as they are — the AI follow-ups add the story after the first open answer; they don’t replace your metric.
Which channels?
WhatsApp by default because it gets the best results, then SMS, email, QR code, tablet, in your app or on your website — per journey, your call. Any customer without WhatsApp is detected and reached by SMS, or by email if that’s what you have; nobody is skipped.
Is surveying customers allowed?
Yes — asking your own customers about a service they just used is standard practice on any channel. What it needs is a line in your privacy notice: that you survey customers, on which channels, and how to opt out — we help you write it. One message per event, never a campaign; opting out takes one word, per channel.
Won’t we over-survey people?
No — a frequency cap per customer is built in, so someone with three interactions this week is asked once. You set the cap.
Can we still build the surveys ourselves?
Yes — the builder is yours: questions, logic, languages. We set the first one up with you; after that it’s as hands-on as you like.
- 03
Analysis runs on its own
Every answer is read as it lands — voice notes transcribed — for what they felt, what they talked about, and what moved the score.
Where does the data go?
Nowhere outside Europe — hosted and processed in the EU, end to end, on European AI endpoints — and your data never trains a model.
Can we shape the topics ourselves?
Yes. The taxonomy — themes, topics, terms — is yours to edit; the AI sorts every answer into it and scores the sentiment per topic.
In which languages?
Answers in English, French, Dutch, Spanish, German, Portuguese, Italian, Arabic and more than 100 others are understood as they are — spelling and grammar don’t matter, and a voice note works too.
Can we dig into the data?
Everything is clickable: a score, a topic, a site — click it and the answers behind it open, voice notes included. Slice by journey, site, product, agent or any column you send, and compare two segments side by side.
Can we bring our past responses?
Yes — export them from your current tool and they come in as the baseline, so the comparison starts on day one.
- 04
Your team works the data
Dashboards you explore yourself — each manager sees their own sites — alerts when a score slips, tickets to close the loop, and Nalu to ask your data a question in plain language.
How do alerts work?
You set the trigger — a detractor, a score under a threshold, a topic spiking — or just describe it in plain language, “any mention of a rude advisor”, and the AI filter catches it. The right person gets it the moment it happens: email, or a webhook into your own tools.
How do tickets work?
A bad answer becomes a task: assign it, reply to the customer in the channel they used — WhatsApp, SMS, email — close it, and see it in the numbers.
What’s Nalu?
Zapio’s AI data assistant. Ask your data a question in plain language — “which site lost the most points this month, and why?” — and it answers the way an analyst would: the number, the reason, and the verbatims that prove it. Included.
Can we export the data?
Always. The raw data — every answer, every field — as Excel, CSV or SPSS, from any view. And branded PowerPoint reports on a schedule, per audience. Your data stays yours.
Can it push data into our other tools?
Yes — into any tool that’s open to it. Your CRM, your BI, your contact-centre platform: if it has an API, we connect to it and push what matters where it’s needed. Alerts already reach email and webhooks out of the box.
Where do we start on day one?
One journey, one dashboard we set up with you, and the two or three alerts that matter most. Add journeys as the first one proves itself.
How do we compare it with our platform?
Run the same journey in both for a month, then put the numbers side by side: reply rate, verbatim depth, time to first insight. Decide on that.
Zapio Platform
Yes. Yes. And yes, that too.
As many as you want — users are never a line on the invoice. Every site manager, every regional lead, every partner gets their own login, scoped to what they should see. Reach is the point, and we don’t charge for it.
Really unlimited?
Yes — people never appear on the invoice. The price follows journeys, volumes and, for a network, the number of sites you analyse — not the number of people reading. Two hundred site managers logging in cost nothing more than twenty.
Do partners count?
Yes — external partners, agencies and franchisees get their own logins too, scoped to their network.
Zapio on a phone or tablet. A new answer or review pings the right person, one tap opens it, and they reply to the customer from wherever they are — dashboards in your pocket, the desk designed for a phone.
Who can use it?
Anyone you give a login to — head office and the CX team, regional and site managers, front-line staff, external partners — each seeing exactly their own scope: their sites, their journeys, their tickets.
What can they do in it?
Everything the web does — dashboards, answers, tickets. On the phone the desk comes first: the list, the answer, the reply, in three taps.
Do we download an app?
No store needed: it installs from the browser on iPhone and Android — “Add to Home Screen” — no IT rollout, and it updates itself.
Won’t they be flooded?
Three guards: they only get what’s in their scope — their sites, nothing else; which events reach the phone is configurable per team — a new answer, a detractor, a review, a ticket; and each person can switch notifications off per device.
Set up like you are: brands, countries, regions, sites — each level with its own look, templates and dashboards, the group seeing across all of it. Who sees what and who does what is yours to set, down to a single site. Every change is logged.
And if we leave?
You leave with everything: every answer and every field, in open formats, at no cost — and an exit plan written into the contract. No lock-in, by design.
Who controls what people see?
Your admins. Roles say who sets up, who reads, who answers customers; scopes say which sites or journeys they see — a manager their sites, a partner their network, head office everything — enforced in the data layer, not on the screen.
Is there an audit trail?
Yes — who changed what, when: members, scopes, forms, dashboards, exports. Kept, and exportable on request.
Hand-built, or drafted by the AI from your own past verbatims — and never frozen: as customers bring up new themes, the AI proposes them and you decide. Yours to shape, AI to grow.
We already have one — can we keep it?
Yes. We load it as is, themes, topics and terms in your words, and the AI sorts against it from day one — proposing additions only where your comments don’t fit.
And when it changes?
The past answers are re-read against the new version — nothing is lost, and the trends stay comparable.
Does it work across languages?
One taxonomy, every language: an English complaint and a French one about waiting time land in the same topic, so trends compare.
Can we have several taxonomies?
Yes — per brand, per journey, as many as you need. Every source, a journey survey or a review platform, is read against the taxonomy you assign to it.
Google Maps, Trustpilot, Google Play and App Store reviews; Facebook, Instagram, X and LinkedIn — into the same analysis pipeline as your survey responses, on the same dashboards.
Can we reply from Zapio?
Yes. A review or a comment lands as a ticket next to your survey answers; you reply from the same desk, and the reply goes back to the platform it came from.
Can you add other platforms?
Name them. If the platform opens its data, we assess it and add it — our product team builds client-led, and new sources are added on request.
An open platform: it pulls from and pushes to the systems you already run — your CRM, your BI, your contact-centre platform, your data warehouse. If it has an API, we connect to it.
Which tools?
Any with an API — Salesforce, HubSpot, Power BI, Tableau, Genesys, Snowflake, Zendesk, Slack, to name the usual ones. Tell us yours.
What goes where?
What matters, where it’s needed: a score and the verbatim on the customer record, a case on a detractor, every answer in your BI.
Is there an API?
Yes — signed, idempotent, with keys you manage: one call per customer event, scheduled up to seven days ahead, and a webhook back on anything you define.
A team of CX and AI people, client-led by design: we set it up with you, train your people, and stay close after go-live — a direct line, not a ticket queue. What clients ask for is what we build next.
Who do we talk to?
The people who build it — by email, phone or chat, and you’ll know us by first name. A named contact and tighter SLAs when you want them.
Is there an SLA?
Yes, in writing: 99.5% uptime, response times by priority, and the metrics that matter for feedback — time from event to invitation, delivery rate, dashboard freshness.
Do you build custom things?
Yes. When it serves more than one client it becomes product; when it’s yours alone, it’s a scoped project with a clear price.
And after go-live?
Training first — live sessions, recordings, guides, included — then closer than before: new journeys as you grow, quarterly reviews, a product shaped by what you ask for.
A one-off study when you have a question; a form when you need one — a sign-up, an event survey, a contact form. Send it to a list, share a link or a QR, or embed it in your site or app.
Is this self-serve?
Yes — build it, send it, read the results: scores, answers and the AI’s read of the open text, per form. No project; we’re a message away if you want a hand.
Can it replace our survey tool?
For most teams, yes: the builder, the distribution, the analysis — in the same platform as your journey feedback instead of a second tool.
Security
Built for regulated buyers.
- Hosted in the EUData and compute in Paris, end to end.
- Deployable anywhereNeed it elsewhere? Another country’s cloud, or your own servers — the same platform.
- EncryptedAt rest and in transit, keys rotated.
- AI stays in EuropeEvery model call on EU endpoints — and your data never trains a model.
- Your data, nobody else’sStrict separation between clients at the database layer — and tested for it.
- Your login, your rulesSSO with your identity provider, MFA, role-based access.
- GDPR, properlyErasure on request. Retention you set. Nothing kept longer than it should be.
Your security team will want more than seven lines. We’re ready for that.
Talk security with us