Connectionary
Category: AI Infrastructure
South Korean AI decision infrastructure startup building structured intelligence for humans and AI agents Connectionary was founded in 2025. The company is led by Jiwon Lee (이지원). Based in Seoul & Busan, South Korea. Team size: 11-50. Latest round: Pre-Seed. Key investors include Bluepoint Partners (블루포인트파트너스).
- Founded
- 2025
- Headquarters
- Seoul & Busan, South Korea
- Team size
- 11-50
Value proposition
Connectionary builds "Decision Infrastructure" — technology that structures unstructured information into decision-ready form before AI systems reason over it, lowering token and computation costs while preserving accuracy. Its core platform AID (Atomic Intelligence Decomposition) converts unstructured information into structured intelligence that can be reused across queries.
Products and solutions
AID (Atomic Intelligence Decomposition) platform — converts unstructured information into structured, computable intelligence, Human-facing intelligence views (company, industry, policy views), Machine-readable intelligence data feeds (in development), Living Spreadsheet for decision-making
Unique value
Structures information before inference (not after), preserving conflicting signals rather than collapsing them, enabling reusable structured intelligence that cuts token/compute costs while maintaining accuracy — solving the fundamental cost-accuracy trade-off in generative AI information processing
Target customer
Financial institutions, enterprises, government agencies, policy analysts, researchers, and security/defense organizations requiring high-precision decision intelligence
Industries served
Financial services, semiconductors, aerospace/space, defense/security, cybersecurity, policy & regulatory analysis, supply chain
Technology advantage
AID (Atomic Intelligence Decomposition) engine that decomposes unstructured sources (PDFs, articles, reports, web pages) into atomic claims at ingestion time, preserving conflicting signals in a shared intelligence structure that can be recomposed under any analytical lens (competitive, policy, financial, security). Team combines deep AI/security expertise from S2W with national intelligence experience.
How they differentiate
Unlike competitors focused on survey/interview research or market reports, Connectionary builds a structured intelligence layer that processes unstructured information at ingestion into reusable atomic claims, enabling multi-lens analysis (competitive, policy, financial, security) from a single shared structure. The team's deep expertise in AI security intelligence (S2W) and national intelligence (NIS) provides unique credibility for high-stakes decision environments.
Main competitors
BlackWho's (블랙후즈) - CURI AI, Intellicia (인텔리시아) - TheSurvey, DBDBLab (디비디랩) - UserSpoon, Counterpoint Research (카운터포인트리서치)
Key partnerships
Bluepoint Partners (investor), Shinhan Financial Group - Shinhan Future's Lab (startup accelerator program), Completed PoC with a major domestic financial institution
Notable customers
Major domestic financial institution (PoC completed, name undisclosed), Shinhan Financial Group (Shinhan Future's Lab selection)
Major milestones
Founded June 2025, AID alpha test launched late 2025, Beta test early 2026, PoC with major Korean financial institution completed, Selected for Shinhan Future's Lab accelerator, Exhibited at INTEROP Tokyo 2026 (June 2026), Raised pre-seed from Bluepoint Partners (July 2026)
Market positioning
Early-stage AI decision infrastructure startup targeting the structural cost-accuracy trade-off in enterprise AI information processing, starting with financial services and expanding into semiconductors, aerospace, defense, and policy analysis
Geographic focus
South Korea (primary), Japan (exhibited at INTEROP Tokyo 2026), global expansion planned
About Jiwon Lee (이지원)
Ex-Deputy CEO (부대표) at S2W (2019-2024); CSO at Dinotia (2024-2025); Chief Strategy Officer at S2W; Senior Researcher at Lotte Strategy Center (2016-2019); Strategy Planning Director at OB Beer (2014-2015); Consultant at Arthur D. Little (2010-2014); Researcher at KT (2006-2010)
Official website: https://connectionary.io