From Documents to Defensible Decisions
We deliver auditable, high-accuracy AI at highest scalability for all enterprise documents by replacing slow, high-cost LLM/RAG solutions with SLM-bootstrapped Claim-Graphs.
2026 | amini@deepentix.com
The Problem
Verifiable Decisions are Manual, Slow, and Unscalable
A real-world example: Our pharma pilot partner needs to research competitor drugs for their new development.
1.1M
Clinical Trial Summaries
Total corpus → filtered to 15,000 highly complex documents requiring expert extraction.
15
Trials per Day
The maximum throughput of expert researchers working manually with no path to scale.
100min
Per 1000 Docs via GPT
Standard LLM pipelines are slow, expensive, and must repeat from scratch for every new metric.
Similar cases across life science, insurance, banking, legal, and industries dealing with document-based decisions.
The Solution
We turn documents into Specialized Claim Graphs that LLMs can trust at enterprise scale.
01
Document Submission
Sample documents are submitted for semantic profiling.
02
Expert Verification
Results are verified and refined through iteration with expert input.
03
Pipeline Specialization
Graph pipeline is specialized using synthetic training data for accuracy.
04
Corpus Indexing
The full corpus of trials turns into a specialized claim graph.
05
Explainable Answers
Queries are submitted via API to the graph, providing full provenance for every answer.
06
Rapid Deployment & Scale
Saving 12-15X in processing time and 7-9X in cost compared to SOTA LLMs.
This solution can be deployed across a 1.1 million document corpus, effectively mitigating the inadequacies of traditional keyword filtering and providing deep, auditable insights.
The Technology
Hierarchical Semantic Claims (HSC)
We ground consuming LLMs with connected facts and multi-hop reasoning chains — eliminating hallucinations at the source, not after the fact.
Tuned SLMs
Specialized Small Language Models are 10x cheaper than generic 70B parameter models. They excel at granular extraction tasks and dramatically reduce error rates in domain-specific contexts.
Audit-Grade Provenance
Every answer traverses a fully verifiable chain: Answer → Claim → Paragraph → Document → Author. Built for regulators, not retrofitted for them.
Data Sovereignty: The Enterprise Moat
Unlike hyperscalers, our lightweight SLM architecture deploys on-premise or in private regional clouds. Zero data leakage of clinical or financial IP, non-negotiable in regulated industries.
Why Now
Regulatory Catalyst
EU AI Act + Solvency II make explainability mandatory. Claim‑level provenance turns compliance from manual to built‑in.
Economic Incentive
CER rework and claims appeals are costly. Verifiable answers shorten reviews, reduce rework, and unlock STP.
Technological Shift
95% of AI enterprise pilots have not led to any ROI (MIT, 2025). Graph x GenAI is mainstream. Big gap in specialized solutions remain open.
Market Size:
€37B
Total Addressable Market
AI Middleware & Knowledge Graphs (2025)
€1.8B
Serviceable Addressable Market
EU Pharma/MedTech & BFSI (2025)
€15M
Beachhead Market
DACH Pharma/MedTech + Insurers
Competitive Insight
Legend: ✅ native • ◑ partial/via partner/custom • ✕ not core
The Strategic Gap
Product-Led Scalability, Not Service-Led Dependency
We provide the platform that empowers YOUR team to build graphs.
Stop renting intelligence; start building assets.
Other Knowledge Graph solution providers earn >60% of their revenue from consultancy service fees.
Founders
Morteza Amini
CEO
CEO
- 3x tech startup founder
- Users from 150+ countries with 5-star reviews
- PhD Numerical BioMech, TU Wien (250+ citations)
Endri Deliu
CTO
CTO
- Ex-Salesforce ML Architect
- 15 years Silicon Valley enterprise AI experience
- AI Fellow, EuroCC Austria
Ali Khalili
CKO
CKO
- Sr. Mgr. Deloitte, Leading Knowledge-enriched AI
- Sr. Res. Fellow Knowledge Rep., VU Amsterdam
- PhD Knowledge Eng. (50+ publications), U Leipzig
Altan Nar
CFO
CFO
- CFA charterholder
- $10M+ raised across 3 companies
- 34% MoM growth track record
Milestones
Customers
- Pilot 1 (Pharma, Switzerland): Auditable protocol workflow automation.
- Pilot 2 (Pharma, Germany): New drug competitor analysis pipeline.
- MVP (Biomed., SciPub+): Knowledge gap-based hypothesis generator.
- 5 new pilots planned for 2026
Grants and Partners
- €270k aws DeepTech Grant.
- €2.5M grant pipeline
- TIB partnership: collaboration on the expert SLM project with Prof. Auer's team at the TIB with iver 2 decades of experience in Neurosymbolic Graphs
MVP
From Answer to Author — Gap Finder Demo
Business Model
Land, Prove, Expand
Usage‑based (Hosted GaaS)
€0.02 per API call/query
We host and operate the graph;
Free graph indexing with a 12‑month contract. Simple pricing; we will adjust as we learn.
On‑prem / Self‑hosted
€80k/year license
Customer runs their own graph;
Customer bears infrastructure/cloud costs
Free Trial (3 months):
Up to 1,000 documents / 20k claims ingested and 1,000 API calls included.
Enterprise add‑ons
Private SLMs, custom ontologies, fine‑tuned deep research tools, audit exports.
Go-to-Market Strategy
Who we target:
- Related solution vendors
- Consultancies/boutiques (e.g., Deloitte partner units)
Social Proof:
- CTO's 15+ yrs experience with SV Enterprise AI solutions
- CKO’s 10+ yrs as Senior Knowledge Solutions Manager @ Deloitte → partner pathways.
- Pilot success
Identify Key Enterprise Players
LinkedIn / Email Outreach
Discovery Call
Loop in Current Vendor
Data mgmt / automation / software
Provide Pain-Specific Content
Live Solution Demo
Pilot Program
RoadMap & Funding
2024
Project start
INiTS incubation
Core team onboarding
2025
Team of 4
MVP + PoC in Pharma
€270K AWS Pre-Seed Deepetech Grant
2026
Team of 7
Goal: Technology Validation with 5 pilots
Raising 1M (€300K + €700K)
2027
Team of 12
Goal: Product-Market Fit
Series A prep (Incl. FFG & AWS Seed)
Next Steps
