Series A & B · Lead investorVertical AI · Infrastructure · Security · Dual-useSan Francisco, CA
Manor X
Our thesis

Software is moving from tools to work.

The next large software companies will replace entire workflows rather than add AI features. We lead Series A and B rounds in the companies doing it first, and doing it best.

The argument

Why now, and why it’s different.

For forty years, enterprise software has been sold by the seat. It recorded the work, routed the work and reported on the work, but people still did it. The budget for software was a rounding error next to the budget for labor.

That boundary is dissolving. Models can now read, reason, call tools and act across systems with enough reliability to own a task from start to finish. That opens the much larger services and labor budget, not just the IT budget.

The prize isn’t a smarter spreadsheet. It’s the job the spreadsheet was built to support.

Incumbents will respond by adding AI features to products priced per seat, built around human workflows. Some will do it well. But a feature that makes a person 20% faster is a different business from a product that does the work and is paid for the outcome. The second kind of company rebuilds the workflow around the agent, owns the result and captures far more of the value.

Not every AI product will last. Foundation models get more capable every quarter, and anything a general-purpose model can do out of the box will become a commodity. The companies that last will compound advantages a general model can’t download: proprietary data from doing the work, integrations into systems of record, earned trust in regulated settings, and feedback loops that make the product better every time it runs.

That is what we look for, in four areas where we think the shift is biggest.

Focus areas

Where the work is moving.

01

Vertical AI agents

Industry-specific agents that complete skilled, high-volume work: processing claims, preparing filings, coding medical charts, reconciling ledgers, dispatching field teams. They win by knowing one domain better than anyone else, and they are often priced on tasks completed rather than seats sold.

What we look for

  • Measurable outcomes customers already pay humans or outsourcers for
  • Deep integration with the industry’s systems of record
  • Founders with real operating experience in the domain
  • Gross margins that improve as automation rates rise
02

Enterprise AI infrastructure

Moving an agent from demo to production is still the hardest part of enterprise AI. We back the layer that makes it dependable: data infrastructure, orchestration, evaluation, observability, and the tools that keep inference cost and latency under control at scale.

What we look for

  • Products that become critical to running production AI
  • Model-agnostic architecture that benefits from model progress
  • Bottoms-up adoption that converts to enterprise contracts
  • Clear value at the CIO, CISO or CFO level
03

AI security

Agents act with real permissions on real systems, which makes them a new and fast-growing attack surface. We invest in two directions: securing AI (agent identity, model and data protection, runtime guardrails) and using AI to secure enterprises (autonomous detection, triage and response).

What we look for

  • Teams with deep offensive or defensive security backgrounds
  • Products that fit the security team’s existing workflow
  • Measurable reductions in exposure or analyst workload
  • A wedge into an emerging, budgeted category
04

Dual-use technology

Commercial AI-native companies whose products also strengthen national capability: industrial autonomy, geospatial intelligence, resilient supply chains, energy and public-sector operations. We want businesses with commercial revenue and a credible, deliberate path into government.

What we look for

  • Commercial traction that doesn’t depend on a single program
  • Teams who understand procurement and compliance
  • Software-led businesses with software-like margins
  • Work that serves both enterprise and mission customers
Discipline

What we don’t do.

  • ×

    Pre-revenue bets

    We invest once customers are paying. Seed-stage founders are welcome to build a relationship with us early.

  • ×

    Thin wrappers

    If the core value is a prompt on top of someone else’s model, the next model release will compete with it for free.

  • ×

    Consumer apps

    We focus on enterprise, mid-market and government buyers, where workflows are complex and switching costs are real.

  • ×

    Party rounds

    We lead or co-lead, take a board seat and own the work that comes with it.

Our process

A decision in about three weeks.

We respect your time. You’ll know where you stand at every step, and you’ll meet the full partnership before we issue a term sheet.

Week 0

First meeting

A 45-minute conversation with a partner about the problem, the product and the customer. No deck required.

Week 1

Deep dive

Product walkthrough, a working session on metrics, and a few customer conversations that you choose.

Week 2

Partner meeting

You meet the full partnership. We share our open questions in advance, so there are no surprises.

Week 3

Term sheet

A clear yes with a term sheet, or a clear no with honest reasons. We never leave founders waiting.

After we invest

Built for the A-to-C gap.

Between Series A and Series C, most companies have to go from founder-led sales to a repeatable enterprise engine. Our platform focuses on that move.

01

Enterprise go-to-market

Introductions to buyers at Fortune 500 companies, help designing pilots that convert, and pricing support for outcome-based models.

02

Senior hiring

An in-house talent partner for VP and C-level searches, plus a bench of operators who have scaled AI companies before.

03

Security & compliance

Fast paths through SOC 2, HIPAA, FedRAMP and the security reviews that stall enterprise deals.

04

Follow-on capital

We reserve to support our companies in later rounds and work closely with growth investors when it’s time to raise.

Fits the thesis?

Show us the workflow you’re replacing.

Pitch us →