Iris.ai — Agentic RAG Knowledge Foundation for Regulated R&D (Iris AI AS, Oslo)

Not shut down. Iris.ai (iris.ai, Iris AI AS, Oslo) founded 2015 by Anita Schjøll Abildgaard / Jacobo Elosua / Victor Botev; Series A May 2024 (€7.64M, Silverline Capital + EIC Accelerator, ~$21.6M total), 42 staff Jun 2026, active 2026-08, no acquire/rebrand/shutdown. Category ai-productivity-data; pipeline_stage none. Killer differentiator = Agentic RAG-as-a-Service over proprietary + scientific corpus with citation-grounded extraction — Axion (chaos→AI-ready) + Neuralith (enterprise knowledge engine) + RSpace (precision R&D intel) + visual RMap concept clustering + Focus filter (10k→100 papers) + autonomous table/text extraction with source links; not Elicit (self-serve $49/mo academic, no private RAG/on-prem), not Semantic Scholar (free search, no enterprise agent/deploy), not Causaly (biomed-only graph), not ChatGPT Deep Research (no private-corpus grounding/audit). 2026 trap: no public price/free tier/self-serve trial — sales-led only, Explorer/Researcher/Enterprise are quoted tiers not list prices, onboarding needs data-engineering (ERP/SharePoint/Veeva/SAP scoping), G2 listing pulled 2026, no MCP server, English-primary science NLP (weak humanities/law), consumer LLMs beat it on open-web breadth, 35%+ LLM-cost cut claim is enterprise-only.

1. One-line positioning

Iris.ai = the RAG foundation for regulated R&D: connect internal repos + PubMed/Scopus/patents → Axion normalizes to AI-ready chunks → Neuralith builds the retrieval engine → RSpace runs agentic literature reviews (problem→RMap→Focus 10k→100→Analyze extract tables→draft with citations) → agents monitor competitor patents / flag new filings / answer “what do our 200 papers agree on” with page-level traceability → deploy in private cloud/on-prem (Clean Room, no training on customer data) → evaluate + monitor via custom frameworks. Fits ai-productivity-data; pipeline_stage none (ingest→map→filter→extract→synthesize→deploy). For pharma/chemical/materials/agri/energy R&D, USDA/Mercedes/ArcelorMittal/L’Oréal tier procurement, it’s the defensible EU-sovereign seat; for a grad student doing one systematic review, Elicit+Semantic Scholar cost 1/200th.

2. Core features

Layer	Capability	Notes
Discovery	RMap visual concept clustering from 300-word problem desc, citation-count-agnostic	Surfaces non-keyword matches (“unknown unknowns”)
Screening	Focus tool: 10k papers → inclusion/exclusion logic → 50–100 set	Replaces title/abstract screen
Extraction	Autonomous table + prose data extraction (params, populations, doses) → structured DB	Source-linked, hallucination-detected
Agentic RAG	Multi-step workflows: patent monitoring, competitor scouting, cross-doc Q&A with traceability	Axion+Neuralith+RSpace stack
Workspace	Researcher Workspace, reading lists, team share, versioned reviews	Explorer=basic, Researcher=full RSpace
Deploy	Private cloud / on-prem Clean Room, model-agnostic (BYO GPT-4/Claude/open weights), ISO 27001, GDPR, EU-sovereign	Enterprise only
Integrate	Sales-led connectors: SAP/Oracle/SharePoint/Box/Veeva/PubMed/Scopus/WoS — no off-the-shelf MCP

3. Pricing (2026-08, USD — sales-led, no list)

Tier	Price	Who
No free tier	—	G2 pulled 2026; no trial/freemium
Explorer	Custom (quote)	Individual researchers, basic smart search, limited uploads, community support
Researcher	Custom (quote)	RSpace full, advanced filters, autonomous extraction, auto-summaries
Enterprise	Custom (annual, often 6-fig)	Private cloud/on-prem, Agentic Multi-RAG, Veeva/SAP/SharePoint, SSO, eval frameworks, priority support

pricing_type = enterprise-quote-only-no-self-serve; starting_price = "contact sales (no $0 tier)"; free_tier = "no"; pricing_note = "iris-ai-as-oslo-2015-active-2026-no-shutdown-no-acquire; explorer-researcher-enterprise-are-quoted-not-list; g2-listing-removed-2026-pivot-regulated-enterprise; onboarding-needs-data-eng-erp-sharepoint-veeva-scope; contracts-annual-auto-renew-impl-training-addons-hidden; 35pct-llm-cost-cut-enterprise-only-claim; no-mcp-server; english-primary-science-nlp"

4. Access Type

access_type: web-app-plus-api-no-self-serve-no-mobile

access_display: 🌐 iris.ai web workspace · 🔌 REST API (enterprise, scoping-led) · no MCP server · no desktop · no mobile app · no self-serve signup

5. Reviews (2026)

Source	Read
AI Agent Index	3.5/5 — “regulated-enterprise RAG done right; pricing clarity 2/5, setup 1/5, evidence 4/5; USDA/Mercedes/NATO NCIA logos”
CostBench	“3 quoted tiers, no free, 6 hidden costs (impl/training/addons), contracts auto-renew”
AIToolsAtlas	“purpose-built science NLP beats GPT on chem/patent text; enterprise-only excludes individuals”
DevOpsSchool	7.8/10 vs Elicit 8.2 / Semantic Scholar 8.7 — “discovery strong, deep-review workflows need Elicit pair”

Praise: citation-grounded extraction defensible for regulatory/IP, RMap finds non-keyword papers, on-prem Clean Room keeps pharma data in EU, Agentic RAG monitors patents autonomously, model-agnostic (no vendor lock), 35%+ LLM-cost cut at scale, NATO NCIA/USDA/L’Oréal deployments validate procurement.

Gripes: zero transparency — no price/page/trial, Explorer/Researcher/Enterprise are sales labels not SKUs, onboarding = data-engineering project (weeks), G2 delisted 2026 hints churn/pivot, no MCP/self-serve/dev API, English-primary science NLP weak on law/humanities/qualitative, consumer ChatGPT Deep Research broader on open web, learning curve steep, hidden impl+training+addon costs.

6. Best for / Not for

Best for	Not for
Pharma/chem/materials/agri/energy R&D needing private-corpus Agentic RAG	Grad students / solo academics (Elicit $49 + Semantic Scholar free)
Regulated orgs (USDA, Mercedes, ArcelorMittal, L’Oréal) with Veeva/SAP/SharePoint	Humanities/social-science/law reviewing (citation-graph tools fit better)
Patent scouting / competitor tech monitoring teams	Casual “summarize this PDF” users (Scholarcy/Scite/ChatGPT)
EU-sovereign AI buyers avoiding US-model dependency	Teams without data-eng capacity to scope connectors
Systematic-review labs wanting audit trail + extraction tables	Startups watching burn (sales-led 6-fig annuals)

7. Competitors

Tool	Lane	Starts
Elicit	Self-serve academic systematic review, structured extraction	Free / $49-mo
Semantic Scholar	Free 200M-paper discovery + TLDR + API	$0
Scite	Smart Citations (support/contradict), citation context	$20–50-mo
Consensus	Evidence-Q&A across papers	$0–45-mo
Causaly	Biomed knowledge graph, drug-discovery focus	Enterprise custom
ChatGPT Deep Research	Open-web agentic research, no private RAG	$20-mo
Gemini Deep Research	Google-stack research agent	$19.99-mo
Microsoft Copilot Studio	Broad enterprise agent build, not science-specific	Seat-based