The AI-native
operating layer
Accelerating high-intensity knowledge work with agentic systems.
Built by an investor, for investment firms.
- Employee #1 of a public equity fund grown from scratch to $11B: ~20%+ annual returns over 8 years, $3B overseen at peak
- Led a research team in software and semiconductors
- Delivered production-ready agentic systems to investment and advisory firms
- Continuous test users at QIA, Ashburton, Lombard Odier, Federated Hermes
- Event-driven fund, US Midwest
- Sustainability fund, Australia
- M&A advisory, Istanbul (next slides)
- Dutch Climate Transition PE
ONE OPERATING LAYER FOR AI-NATIVE KNOWLEDGE WORK: PE · VC · M&A ADVISORY · PUBLIC MARKETS
The scale ceiling is shattering for knowledge work.
Headcount and coordination capped growth and quality. With agents, the caps are gone and market shares will follow.
10× SPEED · 10× APERTURE · ~100% ROBUSTNESS VS 50–60% IN HUMAN ORGS
Examples: Block targets $2M gross profit per employee (4× pre-COVID). Kirkland & Ellis invests $500M into its own platform. Starbucks builds AI software to replace Microsoft and IBM tools. Sierra’s internal agent, Pinecone, writes 70% of its pull requests. Similar systems run at Shopify, Box, Deel and Browserbase.
Better eyes and ears are deployed. Nothing moved.
Brilliant AI tools bolted onto the same way of working. The work still leaks at every handover, humans remain the integration layer, and the org moves at human copy-paste speed.
BCG: AI value is 70% operating model, 10% tools. Most invest in the 10%. McKinsey: high performers are 2.8× more likely to have fundamentally redesigned workflows (55% vs 20%).
Eyes and ears can't coordinate. A spine can.
Every tool and model is a socket on the spine. Tuck in the best one today, tuck it out tomorrow. No vendor lock, no re-platforming. One place to easily maintain and extend.
Spine encodes your workflows, domain knowledge, standards, edge cases, judgment, in a way that evolves in time with every process cycle.
37% of enterprises now run 5+ models in production (a16z, 2025). OpenAI fell from ~50% of enterprise share to ~25%, Anthropic now #1 at ~32% (Menlo). One CIO: "all the prompts have been tuned for OpenAI". Re-tuning is engineering time you pay at every switch.
AI you own. AI you command.
Memory is your prime property. Every decision and artefact stored in full context, as time series.
Audit and improve every stage. Agentic runtimes are traceable, explainable and editable.
One edit updates the whole firm. A new screening rule, risk limit or valuation method applies immediately.
Every process run is a learning cycle that compounds. Two-weekly process, 1% better each cycle: +30% in a year. 5%: +250%.
Sentinels scan the world 24/7. Filings, deals, news, emails, calendars: processes are triggered just in time.
Govern the process instead of doing the job. Every run observable step by step, every instruction editable in minutes.
Work flows down the spine. Learning flows back up.
Complete sub-sector sweep, under one hour.
Industry associations and their boards, targeted M&A news, trade fairs, buyer-record adjacencies, recognition lists, industrial-zone tenants, foreign chambers, enterprise-software adoption.
Generational transition, sponsor exit pressure, carve-outs, growth, customer quality and concentration, regulatory tailwinds and headwinds, distress, innovation, cross-border arbitrage.
Finished work, not tools.
Delivered per qualified prospect: thesis, why-now signals, named buyers.
The universe lives in your Sheets and your CRM, synchronized for convenience.
A whole firm as an instrumented graph.
Three stages, measurable outcomes.
Map the workings
We create a full-fidelity map of your workflows, systems, data and institutional knowledge to design your agentic spine, targeting:
— Same output, fewer people
— More from the same team, in quality and output
— Doing things you couldn’t before
Live in 30 days
Within thirty days your first system runs in production, doing real work, not sitting in a pilot. That's faster than onboarding a new hire. We measure the impact against the baseline we mapped, hand the controls, and tune it on live outcomes until it earns your trust.
AI-Native in 6 Months
Weekly check-ins. Test and improve the first system. Strategize a roadmap to expand the spine with new workflows until your core operations run entirely on agentic systems. Your team operates at a new altitude, and the business scales far beyond what headcount alone allows.