What is Ask Ella?
Ella is Citeline's natural-language AI assistant, built directly into Trialtrove+, Sitetrove+, and Pharmaprojects+. Ask a question in plain language and Ella queries the underlying Citeline database.
All answers are grounded in expert-validated Citeline data, with 95% validated accuracy across the Citeline database.
Standard vs. Deep Research
Standard Ask a question, get a direct answer.
Best for:
Counts and quick data checks
Status of a trial, drug, or site
Exploring an indication or pipeline
Answering a one-off stakeholder question
Deep Research Ella analyzes, synthesizes, and delivers a full structured report — with integrated charts, graphs, and maps — from the Citeline data you're entitled to.
Best for:
Site and investigator landscape reports (Sitetrove+)
Pipeline and competitive landscape reports (Pharmaprojects+)
Replacing manual multi-export workflows
Deliverables that need source-backed analysis
Prompting best practices
These apply across all three products and both modes.
Be specific about what you want Name the indication, phase, geography, or timeframe. "Phase III oncology trials in NSCLC recruiting in the US" returns far better results than "cancer trials."
State the output format for Deep Research Tell Ella what the report is for: "Give me a competitive landscape report for BD review" helps it structure the output and include the charts and comparisons you'd actually use.
Use follow-up questions to refine After an initial Standard answer, dig in: "Break that down by sponsor type" or "How does that compare to 12 months ago?" Ella holds context from earlier in the conversation.
Name the drug, company, or condition precisely Use the exact trial name, INN, or company as it appears in Citeline data. Abbreviations and brand names work, but the more precise the term, the more precise the result.
Ella draws from Citeline's databases — it can't pull in competitor press releases or news outside Citeline. Use Ella for the Citeline data layer.
Trialtrove+
Standard — example prompts
"How many Phase II trials in HER2-positive breast cancer are currently recruiting in Europe?"
"Show me completed trials for [drug name] across all indications."
"What is the current status of the SUCCESSOR-2 trial?"
"Find trials with a primary endpoint of overall survival in hepatocellular carcinoma."
Deep Research — example prompts
"Generate a competitive landscape report for Phase II and III trials in metastatic pancreatic cancer, including key sponsors, interventions, and primary endpoints."
"Give me a trial activity report for [company name] across all oncology indications over the last 3 years."
"Create a report on the CAR-T trial landscape in haematologic malignancies — include phase breakdown, sponsor mix, and enrolment geography."
"Give me a report on paediatric rare disease trials in Phase II and III globally, broken down by therapeutic area."
Tips: Specify phase, status (planned / ongoing / completed), and geography together for tighter results. Use follow-ups to filter by enrolment size, endpoint type, or sponsor type.
Sitetrove+
Standard — example prompts
"Which sites in Germany have the most experience recruiting in Type 2 Diabetes trials?"
"Find oncology sites in the US that have run Phase II trials in the last 3 years."
"Show me investigators with 10+ trials in NSCLC."
"Which sites have recent experience with rare disease pediatric trials?"
Deep Research — example prompts
"Generate a site landscape report for Phase III Crohn's Disease trials in Western Europe."
"Give me an investigator report for [investigator name] including trial history, site affiliations, and therapeutic areas."
"Produce a feasibility report for enrolling in a rare CNS disorder trial across the US and UK."
Tips: Lead with the indication, geography, and phase — these anchor the site and investigator filters. Use follow-up questions to de-risk: "Which of those sites had enrolment delays in the past 2 years?"
Pharmaprojects+
Standard — example prompts
"What drugs are in Phase II or III for idiopathic pulmonary fibrosis?"
"Which companies have the most assets in the GLP-1 space?"
"Show me the pipeline for [company name] in oncology."
"Are there any KRAS-targeted therapies that recently entered Phase III?"
Deep Research — example prompts
"Give me a competitive landscape report for the NASH/MASH pipeline, including late-stage assets and key sponsors."
"Generate a pipeline analysis for precision oncology assets targeting EGFR in NSCLC."
"Produce a white space report for rare autoimmune indications with fewer than 5 late-stage assets."
"Create a BD landscape report covering early-stage CNS assets in-licensed in the past 24 months."
Tips: Scope the report upfront: "Include Phase I–III, all modalities, global geography" avoids mid-report gaps. Pair with follow-up questions for competitive depth: "Which of those sponsors have partnered assets?"
What a Deep Research report includes
Reports return in minutes and are structured for immediate use:
Charts and graphs — pipeline breakdowns, phase distributions, sponsor comparisons
Maps — site and investigator geographic distributions
Structured narrative — analysis written for direct use in decks or internal reviews
Source attribution — grounded in Citeline's expert-validated records
Follow-up question strategy
Ella holds context across the conversation. Use follow-up questions to layer in detail without starting over.
Narrow: "Filter that to US sites with enrolment in the last 18 months."
Compare: "How does that compare to the previous year's data?"
Reformat: "Present that as a table I can copy into a slide."
Dig in: "Why does Site X have so many more trials than the others?"
Expand: "Include Phase I assets in that pipeline analysis too."
Validate: "Which of those sites would you recommend for a rare disease trial given their enrolment track record?"
Your Data & Privacy
Q: Does Ask Ella use client-entered questions or prompts to train its AI models?
A: No. Questions and prompts submitted to Ask Ella are not used to train or improve the underlying AI models. Ask Ella uses the submitted question only to understand the user's request and retrieve and synthesize relevant information from Norstella's proprietary and licensed data sources. Its responses are grounded in these trusted sources and are not learned from client-entered inputs.
Entitlement and availability
Questions? Reach out to your Customer Success Manager or the Client Enablement team.