AgentUse Studioagentuse.io/studio

Before you hire the next ops
or content lead, let me absorb
the function instead.

I install the recurring work as an AI operating layer, not as another seat on the org chart. AgentUse-native by default, stack-aware when policy or legal requires it.

Book a 30-min fit callSee what we ship

Founder-led implementation engagements for B2B and B2C teams. The team stays small. Throughput compounds.

I install AI operating layers for founder-led B2B and B2C teams: closed-loop workflows that absorb recurring ops, content, or customer work so the team stays small and throughput compounds. Your founding moment is the moment to install the new shape, not bolt AI onto the old one.

~Leon Ho·operator, AgentUse
Sound familiar?

You decided AI belongs in your operations. It hasn't happened.

If any of these describe your week, the problem isn't conviction. It's shape.

1

The experiments never shipped.

A dozen AI pilots, a team ChatGPT plan nobody opens, maybe an automation contractor. Nothing runs a function end to end today.

2

You're still the coordination layer.

Every recurring workflow routes through you. Growth adds coordination faster than it adds throughput.

3

There's a req you don't quite believe in.

An ops or content hire where a gut feeling says it staffs work a system should own.

→

That's the old company shape hitting its limit.

The pick isn't hire vs tool. It's bolt AI onto the old shape, or install the new one. Studio exists for the second option.

What we ship

How we engage. Fixed scope, fixed price.

One path, three steps: prove it in a week, build the layer, keep it running. Fixed prices; no hourly billing, no time-and-materials, no RFPs.

Step 1 · Start here

Operating-Layer Pilot

Your first function running in one week. Not a slide deck: one real closed-loop agent, in your repo, on your data, with a human-review gate. The full operating-layer map ships with it.

$5K/ 1 week
paid upfront
  • Day 1: deep-dive interview; score every recurring function
  • Day 3: go / no-go gate. No credible install-first functions, full refund
  • Day 5: pilot agent running + the map (install first / augment / leave human) + review call
  • 100% credited toward your build if booked within 30 days
Worst case: one interview and your $5K back. Best case: you cancel a $120K hire.

Example first functions: inbound-lead research and enrichment · support triage and first response · recurring reporting · content repurposing and distribution. The Day-1 interview scores yours and picks the highest-leverage one.

Built for founder-led, post-PMF teams (usually 10-35 people) where recurring work has gone hire-shaped and internal AI attempts haven't reached production.

Founding terms: the first 3 Pilots run at $2,900 in exchange for a named case study and two warm intros. List returns to $5K as they fill.

Start with the Pilot
Step 2 · The build

AgentUse-native sprint

The full build: recurring functions absorbed one by one into an operating layer of agents on cron, webhooks, or CI. You own the setup the day it ships.

  • Each recurring function absorbed end to end (you own the agents that run them)
  • Fixed scope, fixed price; a full 3-function layer ships in ~30 days, single functions faster
  • You own the setup: .agentuse files, infra, dashboards
  • Built on AgentUse (open source, Apache 2.0)
From $12K per function, once · Pilot credits 100% in · the req it defers runs $120-180K/yr loaded
Book a fit callOr start with the Pilot
Step 3 · Keep it running

Retainer

I operate the layer and keep expanding it: roughly one new function a month, two working sessions a month with your team, async access with a one-business-day SLA. Month-to-month after a sprint completes, never a lock-in.

From $4K/mo · packaged, never hourly
Book a fit call to scope
How we work

A designed delivery system, not a calendar.

01 / 03
Leon
Architect, strategist, builder
  • Designs the system
  • Owns strategic decisions
  • Owns the dashboard
  • Makes the bets
02 / 03
AI layer
Analytical, generative
  • Most analytical work
  • Most generative work
  • Anything an agent does well
  • Routine recurring loops
03 / 03
Client
Decisions, accountability
  • Founder accountability
  • Internal politics
  • Final go / no-go
  • Domain knowledge

You are not hiring a generalist with a calendar. You are hiring a designed delivery system that absorbs the function instead of scheduling around it.

Why AgentUse

The reference architecture, not the required dependency.

AgentUse is the environment Leon moves fastest in. It is open source, in active production, and not a precondition for working together.

Open source · Apache 2.0

github.com/agentuse/agentuse

Markdown-defined agents, runnable via cron, webhook, CI, or Docker. The same runtime used in active client engagements ships under your own GitHub org the day a sprint starts.

Why I am the one installing this

  • 20 yrsLifeHack, a B2C media business: 200K subscribers, billions of pageviews, RAG over 30,000 articles, native iOS app.
  • 12 moTransformed LifeHack to a solo operation. Same revenue, same output, different shape.
  • NowTwo active engagements installing this pattern: a B2B sourcing-operations team and a B2C consumer-health brand (named writeups publish as each client approves).
  • PriorRed Hat: led 12 engineers across 4 countries shipping enterprise Linux.
Case study

Operating layers, in progress.

Early engagements, anonymized while live. Client names and quantified outcomes publish as each one closes and the client approves.

Anonymized client · reference available on request
B2C · early-stage consumer brand
Founder-led consumer startup
solo
founder operation, ops absorbed by AI
~40
past users being re-engaged for the first cohort
in progress
first validated cohort pending

Installing the operating layer for a solo founder: an AI source-of-truth system that keeps the company's canonical docs consistent and flags contradictions before they ship, plus an insight engine re-engaging past users for the first validated cohort. The goal is to keep the business a one-person operation as it scales, with AI absorbing the recurring governance, content, and research work.

  • Source-of-truth governancekeeps canonical docs consistent, flags contradictions
  • Customer-insight enginere-engages past users, structures the findings
  • Brand & content opspositioning, SEO, founder dashboard

Anonymized for now. A reference is available on request; the named writeup and quantified outcomes publish as the first cohort lands.

Things buyers ask first

The three questions worth answering on the page.

?

Can AI actually replace a whole role?

No, and that's not the offer. Agents absorb functions: the recurring, rule-based majority of a role, behind a human review gate. Judgment, relationships, and accountability stay human, and the Pilot's map says explicitly which work should stay that way. What happens to the req is simpler: deferred a year, shrunk to fractional, or dissolved, because the volume that justified it is gone. When whole seats disappear, as they did at LifeHack, it's because the company's shape changed, not because an agent impersonated an employee.

?

What if AgentUse is not a fit for us?

AgentUse is the default because it is the environment I built and move fastest in. If you have an existing stack (LangChain, CrewAI, raw SDKs, internal tools), a policy constraint, or a legal constraint, Studio applies the same operating-layer pattern on your stack as a scoped build, from $40K, 5 to 7 weeks. The methodology is what you are hiring, not the tooling choice.

?

How small is the team really?

Studio is founder-led. Leon is the architect and accountable lead on every engagement. AI does most analytical and generative work. We deliberately stay small to keep quality and judgment density high.

FAQ

Operational details.

What engagement shapes do you offer?
Three: an Operating-Layer Pilot (1 week, $5K: your first function running plus the full map), a fixed-scope build sprint (from $12K per function; a full 3-function layer ~$30-36K, typically 30 days), and a monthly retainer (from $4K/mo) that starts only after a sprint has shipped. No hourly billing, no time-and-materials.
How long does a typical engagement take?
The default sprint is 30 days. The Pilot is 1 week, ending with a working pilot agent and a review call. Retainers run month-to-month after a sprint completes. Custom-stack sprints run longer because they require additional discovery and integration.
How does payment work?
The Pilot is $5K paid upfront ($2,900 on a founding slot, in exchange for a named case study and two warm intros). Sprints are 50% on start, 50% on completion. Retainers are billed monthly, from $4K/mo. The Pilot fee credits 100% toward a sprint if you book within 30 days.
Who owns the code and infrastructure?
You own all of it. Code, .agentuse files, dashboards, and infra ship to your GitHub org and your accounts. Leon retains rights to anonymized patterns and architecture for case studies, never to your data or code.
Do you take enterprise engagements?
No. Studio sells to founder-led teams with founder-discretion budgets. Procurement loops, RFPs, and security reviews longer than two weeks are a polite mismatch. If you need enterprise, you should hire an enterprise integrator.
Why founder-led? Is that a scale risk?
Founder-led describes how we deliver, not the size of company we work with. The bet is that one operator plus an AI layer ships better operating layers than a 12-person agency does, because judgment density and architectural taste live in one head, and you get that head directly, not an account manager. If we scale, it will be by cloning the model, not by adding intermediaries.
Next step

Hire the seat,
or install the layer.

A 30-minute fit call. You describe the recurring work that has become a hiring-shaped bottleneck. I tell you whether an operating layer fits, and whether Studio is the right team to install it. Worst case, you leave knowing which of your open reqs a system could own.

Book a 30-min fit callEmail instead

Technical buyer? Read the runtime first at github.com/agentuse/agentuse.