Ask the question. Rafiki finds the answer in your data.
50 specialised analysis tools, multi-step reasoning, and access to all your data: observations, welfare scores, enrichments, veterinary parameters, histories. Rafiki doesn’t guess — it analyses, cross-references, compares, and responds with facts.
- 50+ analysis tools
- Multi-step reasoning
- 3 depth modes
Animal Welfare Monitor
Animal Welfare Monitor
A true analyst, not just a chatbot
Most AI assistants simply respond from a text corpus. Rafiki queries your data in real time.
When you ask “How has Kibo been doing lately?”, Rafiki does not look for a generic answer. It launches a series of analyses: it reviews the latest behavioural observations, calculates the time budget over the last 30 days, compares with the previous month, checks weight trends and welfare scores, examines installed enrichments, and compiles keeper notes.
What makes Rafiki unique is its multi-step reasoning. It doesn’t just glance at the data once. It reasons, identifies what it needs to verify, launches new queries to dig deeper, and builds its answer step by step — in seconds rather than hours.
A complete analysis arsenal, mobilised on demand
- Behavioural analysis: activity budgets, trends, detection of significant changes
- Veterinary tracking: health history, treatments, welfare scores by category
- Enrichments: installation history, impact assessments, cross-referencing with behaviour
- Complete timeline: all park events, free-text search
- Tasks, objectives, rapid assessment groups, sectors and enclosures
- Rafiki automatically selects the right tools, in the right order, and combines their results
It doesn’t search for an answer, it builds an analysis
What sets Rafiki apart from a simple search engine is its ability to reason in multiple steps. When the answer isn’t in a single dataset, Rafiki chains analyses together.
For example, if you ask why an animal’s stereotypies have increased, Rafiki first verifies the trend, then looks for recent changes in the timeline, examines enrichments, and compiles keeper notes — delivering a structured response with probable causes and supporting data.
Depending on question complexity, Rafiki can chain up to 20 analysis steps in standard mode, and up to 100 steps in deep report mode.
From a quick answer to a comprehensive report
- Quick mode: response in seconds for factual checks
- Thinking mode: cross-analysis with up to 20 steps, period comparisons and correlations
- Report mode: exhaustive analysis up to 100 steps, structured document exportable as PDF
- Executive summary, per-animal analysis, trends, anomalies and recommendations
- Ideal for team meetings, accreditation files and quarterly reports
Available where you work, with the right context
Rafiki integrates directly into your workspace. From an animal record, it focuses on that individual. From an enclosure or sector, it adapts its analysis scope. From the park-wide view, it searches all data.
This automatic contextual framing means you don’t need to specify who you’re talking about — Rafiki already knows. It can also broaden its analysis when needed, pulling data from the enclosure or sector for context.
In addition to on-demand conversations, Rafiki generates daily reports for each sector: tasks, enrichments, observations, notes and attention points. Every conversation is archived and retrievable.
What Rafiki changes in your daily work
- Hours of analysis in seconds
Manually cross-referencing observations, weight, enrichments and scores over 6 months would take a full day. Rafiki chains up to 100 analysis steps in under a minute.
- Correlations invisible to the naked eye
A link between an enrichment and reduced stereotypies? A correlation between a group change and weight loss? Rafiki detects what no spreadsheet can show.
- Evidence-based decisions
Should you invest in new perches or food-based enrichments? Rafiki shows you the data supporting each option. Your decisions shift from intuition to evidence.
- Ready-to-use reports
Quarterly reports, accreditation files, monthly reviews: Rafiki generates structured, data-driven documents exportable as PDF. Writing time drops from 2 hours to 5 minutes.
- Accessible to the whole team
No statistics skills needed. Ask your question as you would ask a colleague. Rafiki understands natural language and responds in plain language.
- Your data stays yours
Each park has its own isolated space. No data is shared between facilities. Isolation is guaranteed at every query level.
Explore Rafiki
Rafiki draws from the entire AKONGO ecosystem
Rafiki accesses all data collected by other modules to build its analyses.
- Behavioural analysisActivity budgets, trends, and period comparisons feed directly into Rafiki’s responsesExplore module
- Veterinary dashboardHealth parameters, treatments, and clinical notes are accessible for cross-analysesExplore module
- EnrichmentsInstallation history, impact assessments, diversity — Rafiki cross-references with behavioural dataExplore module
- Data collection campaignsAll collected data is available for analysisExplore module
- Rapid Welfare AssessmentScores, objectives, and assessment groups accessible by RafikiExplore module
- Individual managementLife context, transfers, group changesExplore module
- TimelineThe park’s entire activity, browsable and searchable in free textExplore module
- Task managementTracking of completions and delaysExplore module
Give your data a voice
Learn more about our solutions
- Tools
Explore our comprehensive range of tools optimised for managing and analyzing animal welfare data.
- Training
Discover our training organisation and develop your expertise in ethology and animal welfare.