CLINICAL VOICE AI AGENT

LOLA®: the AI agent that
keeps patients
connected to their care
programs.

LOLA, Tucuvi’s clinical voice AI agent, engages with patients through natural conversation, following the workflows your organization defines. It collects relevant information and escalates situations that need your care team’s attention.

An AI agent that empowers clinical teams
https://vimeo.com/1239908611/aab137fb24?share=copy&fl=sv&fe=ci
NATURAL LANGUAGE INTERACTION

Natural interaction, with clinical control at every step

Patient record in the Tucuvi platform
https://vimeo.com/1239579169/886859ce38?share=copy&fl=sv&fe=ci
CAPABILITIES

Built around each program’s needs

Each program defines how LOLA reaches patients, how it handles their requests and what happens when contact isn’t made.

Patients answer in their own words and in their preferred language, with no app to download or portal to sign up for. Your team reviews results in its own language, with the original transcript always available.

Patient answering a call from LOLA on their mobile
Outbound outreach

LOLA proactively contacts patients for the follow-ups, reminders and specific tasks defined by each programme, and escalates situations that need the team’s attention.

Inbound requests

LOLA handles incoming calls from start to finish within the programme’s scope. Patients can return a missed call to continue their follow-up or call directly for supported requests, including access and demand management. When a request needs clinical judgement or falls outside the programme, LOLA routes it to the appropriate team.

Program-defined retry rules

Patients can return a missed call to LOLA and continue their follow-up. Programmes can also leave voicemail messages and configure how often LOLA tries again and when calls are made. If contact is not established, the team can review the patient’s status and decide what happens next.

Omnichannel

Programmes can combine phone and text, so patients can respond through the supported channel they prefer.

HYBRID ARCHITECTURE

Natural conversation, with safety by design

LOLA combines models that interpret patient responses with explicit protocol rules and natural, empathetic conversation. The protocol defines what LOLA can do, when it can proceed and when the team needs to take over.

Conversational flexibility with defined scope

LOLA generates responses suited to the conversation, while protocol rules govern permitted actions and escalation. Our declarative architecture makes these rules and outcomes explicit and easier to inspect and test, while supporting more natural and empathetic interactions.

Understanding, control and conversation

System 1 models interpret responses, while other task-specific models handle open-ended information. LOLA checks the information against the protocol and applies its rules. Large language models generate natural, empathetic responses, including clarification when needed, within the scope set by the protocol.

AI Sentinel with human-in-the-loop review

Before clinical information reaches the team, AI Sentinel checks the completed interaction for potential errors or inconsistencies. Cases that require further review are routed to Tucuvi’s review team. This keeps human-in-the-loop review manageable and scalable as volume grows.
Deterministic models capture clinical information with precision. LLMs facilitate natural, empathetic conversations. When combined into a single architecture, they enable professionals to receive reliable information while ensuring that patients feel truly heard.
Oscar García
AI Engineer
A virtual agent that supports clinical teams
https://vimeo.com/1087037724
CARE TEAM REVIEW & ACTION

Every interaction, ready for the team to act on

Every LOLA interaction is logged in the patient record in Tucuvi Dashboard, with its summary, transcript, audio and alerts. From the record, clinicians contact the patient, document their actions and update the review status.

Expand your care team's capacity with patient engagement across the entire care journey