Investor Presentation · March 2026
Turning team knowledge into AI execution
Pre-Seed
$500K · SAFE
Singapore · SF
Andrey Leskov · CEO & Co-founder · andrey@illumi.one · illumi.one
Turning team knowledge into AI execution
Andrey Leskov · CEO & Co-founder · andrey@illumi.one · illumi.one
Model quality, protocols, and agent infrastructure will keep improving and become broadly available.
The durable moat is how a company thinks, decides, and operates. That is what AI cannot invent on its own.
illumi gives every company its own operational memory layer and makes that intelligence executable for teams and agents.
If the last generation of software digitized workflows, illumi digitizes intelligence and makes it executable. It is the foundation for AI-native work.
illumi is a brain for AI-native company to build skills for its AI tools and agents
Inconsistent AI output
Rebuilt context every session
Repeated decisions, same dead ends
Stalled agents, no shared memory
MIT 2025
Atlan 2025
Asana 2025
The missing system is shared context.
Andrey Leskov
CEO & Co-founder
Scaled distributed engineering teams from 10 to 150 across 30 nationalities and 20 countries. Former CTO — watched institutional knowledge evaporate across every handoff. Knows exactly where business intent dies before it reaches engineering. CoverGo (prev. employer) raised $15M Series A.
Lingyi Chang
CPO & Co-founder
15 years in tech & AI. 8 years at Microsoft & AWS leading AI/ML adoption across Asia-Pacific. Ran enterprise AI programs that failed — not from bad models, but missing context. Saw how successful projects made it by collaboration between business and tech. Building the layer she needed and couldn't find.
Drop existing docs:
Team reviews on canvas:
Illumi builds for:
Use them seamlessly with any agents & models:
| Feature | ✦ illumi | Miro / FigJam | Notion / Confluence | ChatGPT / Cursor |
|---|---|---|---|---|
| Multiplayer | ✓ | ✓ | partial | — |
| Visual | ✓ | ✓ | — | — |
| Doc depth | ✓ | — | ✓ | partial |
| AI Context | shared | — | — | not shared |
| Executable | ✓ | partial | partial | ✓ |
Visual like Miro + Executable like Cursor
Real-time collaboration like Figma
Version control without Git trauma
Cross-platform: one source → all agents
Network effects: more context = stickier
Proprietary compiler: visual → multi-skill formats
Enterprise moat: compliance, SSO, audit logs
ChatGPT, Copilot, Midjourney
Focus: Personal productivity
Claude Computer Use, GPT-4 Agents, MCP
Focus: Autonomous workflows
Team agents, multi-step workflows
Focus: Governance at scale
The market has moved past "Can AI work?" to "How do we govern AI at scale?" We're building the governance layer before it becomes a compliance crisis.
Usage depth — professionals vs. casual
Professionals avg 3.8 boards vs 2.3 · 68 cards vs 28. Users come to illumi for complex work
Named outcome
Sprint realignment meetings: 4×/week → 1x. Collaborative workflow for speaker sourcing and event orgainising
No revenue yet. Strongest proof: deep retention and measurable outcome in the wedge.
Math
~1.25M businesses (US 590K + EU 306K + UK 354K)
× avg 30 seats × $29/seat × 12 mo = $10.4K/team
Expansion SAM
$150B+
~15M knowledge-worker orgs globally × $10K — as MCP / agent adoption scales to all enterprise.
43 teams × $3.5K ACV = $150K ARR
500 teams × $5K ACV = $2.5M ARR
Poor collaboration costs $10K–$30K per employee per year. illumi is cheaper than the problem.
Every board a team builds makes agent output more accurate for that domain. The moat is the growing knowledge graph, not a feature competitors can ship.
Only platform where business stakeholders co-own agent skills without code or YAML. Context shifts from engineering bottleneck to shared asset.
Works with any agent stack via MCP. Distribution unlock that didn't exist before Nov 2024. Customers own the context, not the model.
Business approval before remote agents run. A governance layer large platforms are unlikely to prioritize.
1.5 years on collaboration-native architecture — before agents needed it.
Free tier → PM adopts → team joins → org buys. ICP: PMs, chiefs of staff, innovation leads at tech-forward SMEs.
Convert LITEON, CoverGo network. 6-mo pilot at reduced rate — case study rights in return. Entry: VP Ops, Head of AI, Chief of Staff.
Ops-heavy orgs with active AI programs. Governance layer pitch. 2–6 months sales cycle — no IT procurement at pilot stage.
This round proves Motion 1 + 2. Design partner → paid conversion is the single commercial milestone this capital needs to deliver.
Individuals and small teams. Build and test agent skills. Learn and integrate into workflows.
Shared workspaces. Collaborative skill development. Higher limits and productivity features.
RBAC and governance. Security and compliance readiness. Internal skill distribution.
Usage-based and credit pricing. Skill marketplace. Deploy skills into internal tools and IDEs.
We don't just monetize seats. We monetize the agent skills companies depend on.
Known risks
· Market education — context gap not yet named by buyers
· Activation gap — broad self-serve onboarding is weak
· Competitors copying surface features
· No repeatable monetisation yet
Still needs proof
· Design partner to paid conversion
· One-team to multi-team expansion
· Efficient acquisition cost at scale
This round is to prove repeatable paid pull in the wedge. That is the only question this capital needs to answer.
18-month runway
Milestones this capital buys
Visual skill builder + full MCP integrations shipped (Q3 2026)
10 enterprise design partners with active knowledge context (Q4 2026)
5 design partners converted to paid ($3.5K–$12K ACV) by Q1 2027
43 paying teams → $150K ARR by mid-2027
✦ The problem is real
✦ The product is real
✦ The wedge is real
→ Now we prove repeatability
illumi.one · andrey@illumi.one
Turning team knowledge into AI execution