What admissions teams ask us.

Violeta is an AI voice and WhatsApp agent for admissions teams. It calls every person who fills in a form or responds to a campaign, answers their questions with your institution's information, qualifies them against your criteria and hands them to your advisors, either as a live transfer or as a meeting booked in your CRM.

It is built and run by Stackbone, a company that puts AI agents into production for businesses.

It is the time between a prospective student asking for information and getting the first contact. In sales, firms that tried to contact a lead within the first hour were nearly seven times as likely to qualify it as firms that tried an hour later.

In education, the response is often late. In a test of 1,000 inquiries to online programs at 182 institutions, 44% got no response, and the median among those that did was 3 h 18 min. And 76% of online students surveyed by RNL say they would likely enroll in the first program that responds.

No published study measures the effect of speed on enrollment directly. That is why we propose measuring it in your operation, against a control group.

Sources: Oldroyd, McElheran and Elkington, "The Short Life of Online Sales Leads", Harvard Business Review, March 2011 (1.25 million leads, 42 firms). UPCEA, "Enrollment Process Review: Secret Shopper Analysis", December 2025. RNL, 2024 Online Student Recruitment Report.

No. It handles the first contact and the filter: it calls, checks the details, answers the common questions and gauges interest. Your advisors get the people who are ready to talk and spend their time enrolling students, not chasing cold leads.

As soon as the lead lands in your CRM, if it falls inside the calling window you set for the lead's local time. If it arrives outside that window, it waits in the queue and gets a call as soon as the window opens.

It is designed for them. You set the maximum number of simultaneous calls, the queues and priorities (today's leads, by age, by propensity) and the retries for people who don't pick up. The agent follows those rules on every contact, without you hiring or training advisors for the peak.

Yes, through their APIs, with the systems you already use: the agent reads and writes in them and your data stays there. For an online business school it runs on Zoho CRM and Typeform.

For Genesys Cloud, Microsoft Dynamics 365, Salesforce or any other system with an API, the integration is built as part of the project. We don't ask you to change your CRM or contact center.

Yes. Voice and WhatsApp are one flow with the same lead information: if the lead doesn't pick up, the agent can continue on WhatsApp with Meta-approved templates, and if they reply on WhatsApp, it can offer them a call.

If an advisor is available, the agent transfers the call live to the right queue, with a summary of the conversation. If not, it books a meeting with an advisor from your CRM and confirms it with the prospective student. Everything is logged in the CRM.

Yes. It introduces itself as a virtual assistant at the start of every call and, if asked whether it is a person, confirms it is an AI and offers to pass them to an advisor.

The sample calls on this site are in Spanish and English. Other languages depend on the voice provider you choose for your operation.

Yes. Scripts and settings are edited from the agent's dashboard, versioned, and applied without redeploying it. You can have a different script or criterion per program or segment.

Where you choose: on Stackbone Cloud, in your own AWS, Azure or Google Cloud account, or on your own servers. When it runs on your infrastructure, the runs, data and keys stay inside the agent's container, with its own database; Stackbone's control plane does not hold them.

Details on the Trust page.

Every layer is swappable: speech-to-text, the language model (LLM) and the voice. You use the provider you prefer, on your account and your bill, and you can change one without rebuilding the agent or its integrations. That is why an online business school chose us in a competitive demo.

Every call is recorded and transcribed, and every conversation is kept as a run with its trace: each step, each query to your systems, with what went in and what came out. You can also set rules that mask personal data, block a topic or hold an action until a person approves it.

It depends on the country and on the consent the lead gave. In the United States, the FCC ruled in February 2024 that AI-generated voices count as artificial voices under the TCPA, so these calls need the recipient's prior consent. In Europe and Latin America, each country's data protection and commercial communication rules apply.

Your legal team sets the consent basis and the agent applies it: it only calls people who have given it, honors do-not-call lists and identifies itself as an AI. This is not legal advice.

We measure it the way you would measure a new team: part of your leads follows your current process and part goes through the agent. We compare time to first contact, contact rate, qualified leads handed to advisors and cost per effective contact.

We start with one flow and one queue, not the whole operation. For the online business school, the plan was six weeks in five deliveries, from the platform being ready to the first call and the volume and calling-hour rules. Your timing depends on the systems to integrate.

It depends on lead volume, channels and the systems to integrate. In the demo we review your operation and send you a proposal. Book a demo.

On a voice platform you choose the LLM, but speech-to-text and voices stay with the provider, the agent runs in their cloud and a third-party partner usually implements it. With Violeta every layer is swappable, it can run in your cloud account or on your server, and one team builds and runs it. The same base then serves WhatsApp and other workflows.

Violeta is a product of Stackbone, a company that runs AI agents in production with control, monitoring and audit built in. The team works from Madrid and Montreal.