The AI agent that stands at your counter
A tablet at reception, a face that answers. Around two seconds to the first answer. No video recording, nobody recognised twice.
EchoCall Presence puts your AI agent on site. A tablet at reception, in the foyer or on the counter shows a face that can be spoken to and answers - the same agent that already handles your calls, on the same knowledge base. A visitor has to type nothing, install nothing and choose nothing, and the device sleeps while nobody is there.
to the first answer
Measured from waking the device to the first sound, not to the first subtitle.
The visitor is not filmed
The device sends audio only. No facial recognition, and nobody is recognised again.
the same voice as on the phone
Presence uses your existing voice agents, German dialects included.
Why waiting time decides everything at a counter
On the phone a caller waits. In front of a screen nobody waits: whoever gets no answer within two seconds turns away, and the conversation never happened. That makes the first second at a counter device not one metric among many, but the one that decides the rest.
Our own first build took nine seconds before anything could be heard, because the face had to finish rendering first. Today our own voice takes over immediately and the face joins at the change of turn. What remains, measured, is around two seconds to the first sound.
The second decision is about cost. An avatar that simply runs costs roughly 65 euros per device per day, even with nobody in front of it. That is why Presence has no standing session: the device sleeps, wakes when spoken to, ends the conversation by itself and goes back to sleep.
What is included
Everything a talking device at a counter needs, including the paperwork that has to hang next to it. You put up a tablet; the rest you manage where your telephony already lives.
- Paired in ten minutes
Create the device, enter a six-character code on the tablet, done. The device checks its own microphone, speaker and network and reports the result back.
- A choice of faces, or your own
A selection is ready to use. Your own face, for example that of your receptionist, you submit with the name and the consent of the person shown.
- Opening hours and a daily limit per device
Hours, the maximum length of a conversation and minutes per day apply per device and follow the clock at the location, not the server.
- Counter notice template in three languages
A ready-made sheet per language for Article 13 GDPR and Article 50 of the AI Act. Fill it in, print it, hang it next to the device.
How it works
From unpacking the tablet to the first conversation at the counter in one morning.
Pick the agent
Presence uses your existing voice agents. The agent that answers your phone can be the same one standing at your counter.
Pair the tablet
Create the device in EchoHub, enter the code on the tablet, allow the microphone. The device reports its self test back within twenty seconds.
Set the limits
Opening hours, silence before the end, maximum length and minutes per day. The defaults are a sensible place to start.
Hang the notice beside it
Fill in the template and print it. Only then does the device go into service, so that no visitor finds out afterwards who they were talking to.
Where a device at the counter changes something
Wherever people arrive and wait: at reception, in a waiting room, in a branch. The more sensitive the place, the more closely the notice and the legal basis deserve a look, and the more it matters that a person is available instead.
Runs with what you already have
Presence is not a second system beside your agent, it is one more place where the same agent speaks. Every connection it uses on the phone it uses at the counter.
Calendar
Google Calendar, Outlook, Calendly, Cal.com
CRM
Salesforce, HubSpot
Telephony and chat
The same agent and knowledge base, on the phone, in the web widget and at the counter
Automation
n8n, Zapier, Make, plus a full REST/webhook API
How Presence compares
An avatar on a screen is quick to build and expensive to run. The differences are not in the face but in the waiting time, in the billing, and in what the visitor gets to know.
Waiting time at the device
Cost while idle
Informing people on site
Video recording
Channels
The most expensive thing we learned while building this is that nobody waits in front of a screen. So the voice speaks first and the face joins at the change of turn - around two seconds, measured at the first sound.
EchoCall Product Team
EchoCall LLC
Frequently asked questions
No. The tablet sends audio only. No video recordings are made, there is no facial recognition and no biometric analysis, and nobody is recognised from one conversation to the next. What is recorded is the same as for a phone call: the written transcript, kept for as long as your agent is configured to keep it.
It sleeps. Your idle text is on the screen and it costs nothing. Presence has no standing session: the device wakes when spoken to, ends the conversation by itself after a while of silence or after a fixed maximum length, and goes back to sleep.
A notice covering three things: the information required by Article 13 GDPR at the point of collection, the disclosure required by Article 50 of the AI Act that an AI system is speaking, and the third-country transfer for rendering the face. We provide a ready-made template in three languages. Eight fields you fill in yourself, because only you know who the controller is and how long you keep transcripts.
A tablet with a microphone, a speaker and Wi-Fi, Android or iPad. Plus a power supply, because a device that stays awake all day does not live off its battery. You switch off the screen lock and automatic shutdown, and on first start you allow access to the microphone.
Yes. You submit a face of your own, giving the name of the person shown and confirming that they agree. No name means no consent, and no consent means no face.
That is the most important sentence on the notice, and it sits right at the top in a box of its own. A device that is the only option puts people under pressure. One that is visibly one of two options takes that away. Which person is available you write into the template.
What was actually spoken is counted, from the moment the device joins to the end of the conversation. Part minutes count in full, a conversation that never got off the ground costs nothing, and every conversation is billed exactly once. Your Presence plan includes a fixed number of minutes per month and a number of devices; beyond that you top up minute packages.
Put your AI agent at the counter
Presence builds on your existing voice agents. Start with 30 voice minutes and 30 chat conversations of starting credit at hub.echocall.de, valid for 14 days - card required for verification, €1 automatically refunded.