Solutions
Build an evidence environment around the knowledge you already have.
Each area below can be approached on its own. Most engagements start with one collection, one question or one workflow.
Evidence Compass is a reusable Innovaforte capability for combining structured data and documentary evidence. Sources, terminology and evaluation methods are adapted to each sector, with particular depth in health, public policy and life sciences.
Connect data and sources
Demonstrated in the pilot · configured per engagement
Integrate international and national statistical sources, official APIs, databases, institutional repositories and existing business systems — and provide structured exports or APIs to other systems. Each connection is source-specific: access rights, definitions, update frequency and data quality determine what can be connected.
Examples
- Official statistics and indicator APIs
- Institutional repositories and databases
- Structured exports and APIs for other systems
Import and structure documents
Demonstrated in the pilot · configured per engagement
Work with reports, PDFs, spreadsheets, CSV, structured web tables and suitably configured document formats. Extract relevant fields, findings and tables; map metadata and terminology; keep source references; and have reviewers validate the result. Not every document or scanned image can be processed automatically.
Examples
- Field, finding and table extraction
- Metadata and terminology mapping
- Reviewer validation
In the Violence Prevention pilot, selected official Moroccan publications are processed through structured report extraction where an API is not used. Values retain their publication context and page/table references and are checked before use.
Evaluate and compare evidence
Configured per engagement
Appraise relevance, study design, population, methodology, jurisdiction, dates, limitations and comparability using criteria suited to the domain. Identify conflicting findings, duplicate records, different publication versions and gaps. Automated assistance supports expert review; it does not certify scientific validity, legal correctness or compliance.
Examples
- Domain-specific appraisal criteria
- Conflicts, duplicates and versions
- Evidence gaps
Explore, ask and synthesise
Demonstrated in the pilot · configured per engagement
Search, filter, map, chart trends and compare — and ask questions answered against a defined evidence collection. Summaries stay connected to their sources; structured analytical rules and expert review govern interpretation.
Examples
- Evidence maps and dashboards
- Questions on a defined collection
- Briefings, evidence tables, reports and multilingual outputs
Assess economic and operational impact
In development
Extend from evidence to costs, resource requirements, implementation scenarios, potential benefits and investment decisions. The current economic demonstration starts with Thailand and one violence-prevention intervention; the intended expansion covers all six pilot countries, with regional and international inputs where methodologically justified. Observed evidence, published findings, assumptions and calculated results stay distinct, and missing inputs remain visible.
Examples
- Implementation costs
- Operational outcomes
- Investment cases
Maintain traceability and human oversight
Demonstrated in the pilot · configured per engagement
Keep citations, source locations, publication status, versions, retrieval dates, validation and review visible. For organisational deployments, access controls and confidential-document workflows are requirements configured for the engagement. Please do not upload confidential material to a public demonstration.
Examples
- Citations and source locations
- Versions and retrieval dates
- Review and validation records
Health and research
From research to policy and practice
Envisaged, configurable workflows for health and life-sciences evidence. No PubMed integration is currently implemented.
Sources
- Official WHO and national evidence
- Publication records and abstracts discoverable through PubMed
- Full texts where access and reuse rights permit
- Preprints, authorised unpublished manuscripts and pre-publication material
- Guidance, evidence reviews, intervention studies and institutional reports
Potential uses
- Organising a literature collection
- Extracting study characteristics
- Supporting screening and appraisal
- Comparing findings
- Tracking versions
- Identifying gaps
- Preparing evidence briefings
Publication status stays explicit: an unpublished manuscript, a preprint and a peer-reviewed publication are distinct, and being indexed is not a quality verdict. Human review remains central.
Relevant to public-health institutions, WHO-facing policy work, pharmaceutical and life-sciences teams, and industry associations such as IFPMA. These are examples of institutional contexts, not clients, partners or endorsers.
Languages
Interface language, source-document language and report or answer language are separate layers. Evidence values and source identity stay consistent across translations. Additional languages can be configured and reviewed for an implementation.

