Answers with the protocol read, and the query attached.
Ask Vivo answers questions about your trial data, contextualized by the study’s own documents. Every answer ships with human-reviewable reasoning steps and the query that produced it — and agents run under the user’s own permissions.
Not a chatbot. A source-backed reasoning layer.
Ask Vivo gives clinical development teams a governed AI interface to their own trial data, documents, protocol context, and operating workflows.
The protocol agent cites the study’s own documents, page by page
Every answer accompanied by reviewable reasoning steps and the query that produced it
Role-aware — every answer reflects your permissions and blinding status
Visual outputs, summaries, and generated tables on demand
Answers trace back to source records with evidence links
“Ask the trial. See the evidence.”
What is the rate of cardiac events in my study, and does the protocol require expedited reporting for them?
7 cardiac events across 5 subjects (rate 1.9/100 subject-months). Protocol §9.3.2 (p. 84) requires expedited reporting for Grade ≥3 cardiac events — 2 of the 7 meet the criteria; both flagged with source records attached for the safety team’s confirmation.
↗ Source: Safety database · Protocol v4.0 pp. 84–85 · Reasoning steps + SQL attached
Summarize missing data, pending verification, and open queries by site.
Sites 211 and 308 account for 62% of open items: 41 missing pages, 28 pending SDV, 19 open queries. Site 104’s queries aged >21 days doubled since last month.
↗ Source: EDC completeness module · DM review state · Query log
Real answers, from a demonstration trial.
The answer and its receipt: the protocol agent’s reasoning, shown beside the database agent’s answer. Demonstration trial, synthetic data.
Plain-language question, structured answer — with Explain Answer one click away. Demonstration trial, synthetic data.
Specialized agents reason over your trial like clinical experts.
Each agent queries the governed data model, applies domain reasoning, and surfaces prioritized findings with the evidence attached. Agents inherit the user’s permissions — one instance per role, blinding-aware by design.
HIGH
Enrollment gap & screen failures
Site 8 has 32 screen failures — top causes identified from screening logs. Recommendation: CRA review at Site 8; redistribute recruitment to two lower-failure sites with capacity.
CRITICAL
SAE documentation gaps
Cross-references ICH E2A reporting obligations: flags SAEs missing narratives or causality assessment while reporting windows are still open — before they become findings.
MEDIUM
Discrepancy concentration
Site 3 far exceeds the trial mean for discrepancies; two sites identified as best-practice models. Reusable rules proposed for the recurring patterns.
SIGNAL
Subgroup hypothesis generation
Surfaces trends and hypotheses on endpoints and subpopulations — divergences by baseline strata, deviation-driven population gaps — each with the analysis attached for review.
The questions teams ask Vivo every day.
How many patients enrolled in the UK last month for my study?
What is the rate of cardiac events in my study?
Which endpoints have changed most from baseline?
Summarize missing data, pending verification, and open queries by site.
What are the safety trends in my biomarker-driven subpopulations?
Rank sites by protocol deviations, dosing issues, and data quality.
Give me a detailed medical history for a set of patients.
Are deviations documented and resolved for inspection?
How many patients did Site 103 recruit across all oncology studies this year?
“Vivo is our new best friend.”
Ask your trial a question. See the evidence come back.
Governed, auditable, role-aware AI over your live trial data — in production with sponsors today.
