AUTONOMOUS AI AGENTS

Multiply
yourself.

Train a Surogate agent on the way you already work, then offer it to your existing patients, clients, or students as a monthly subscription. You earn from every one of them - without adding a single hour to your day.

24/7
They work while you don't
0
Lines of code to build one
Deploy one or hundreds
Scroll

Who it's for

Professionals already doing this

Not generic use cases - real people, already running a Surogate agent for their own clients. Here's exactly what they built, and what it pays.

Cardiologist
Dr. Radu Pretorian

Checks in with his patients after every consultation, tracks symptoms and medication compliance, and flags anything unusual.

What it pays€15–30per patient / month

The agent follows Dr. Pretorian's own monitoring protocol. It doesn't diagnose, and it doesn't replace a consultation.

Law firm
Laza Lawyers

Handles the first conversation with every new client, answers common procedural questions, and routes anything complex straight to an attorney.

What it pays€25–50per client / month

The agent follows the firm's own intake protocol. It doesn't give legal advice in its own name.

Accountant
Rares

Tracks his clients' financial situation year-round and alerts them before small issues become expensive ones.

What it pays€30–60per client / month

The agent follows Rares' own review process. Financial decisions and sign-off stay with him.

Teacher
John Riley

Drills students for exams, tracks their progress, and sends parents a weekly update - using John's own method.

What it pays€20–40per student / month

The agent practices with students on John's curriculum. It doesn't replace his teaching or grading.

Each of them built one agent. Here's exactly how.

How it works

You describe the work. It does it for you.

Tell Surogate what you want done, in plain words. It builds an agent that handles it, runs it in the cloud around the clock, and gets sharper the longer it works - no code, no servers to manage.

Built from your workflow, no code

Define what your agent does in plain language - its goals, its tools, its rules. If you understand your work, you can build it.

Runs in the cloud, 24/7

Your agent lives in the cloud, not on your laptop. It works when you don't - no tab open, no machine running.

Gets better the more it runs

Turn your most common tasks into owned expert models trained on your own work - smarter and cheaper over time.

Why Surogate

Three ways it pays off

A way to turn your work into recurring income, agents that run themselves, and intelligence you own.

01
Rent it out, not just run it.

Build an agent once, deploy it to every client you already have, and earn a recurring monthly fee from each one. An agency without employees - a subscription business without code.

See how it pays
02
Agents that run themselves.

Agents that run in the cloud - executing tasks, coordinating with each other, and reporting back 24/7. You set the goal; they do the work.

See one work
03
Intelligence you own.

Distill your own work into small expert models you own - matching or beating frontier models on your tasks, at a fraction of the cost.

How the flywheel works

Build once, earn always

You build it once. It earns while you sleep.

The agent you ship keeps working for your patients, clients, or students around the clock - handling the routine, escalating only what matters. You're off the clock; the income isn't.

Once
You build and ship the agent.
you
24/7
It works with your patients, clients, or students.
deployed agent
Anytime
It escalates only what actually needs you.
deployed agent
Monthly
The subscription renews. You earn.
recurring income

You build once. Clients pay every month.

Proof - see it work

Watch one agent monitor one patient - then scale to all of them

Not a chat demo. This is Dr. Pretorian's own Surogate agent, checking in with a real patient after a real consultation.

Incoming · live
Hi doctor, I've had some chest tightness since yesterday evening - should I be worried?
Emma · patientvia Dr. Pretorian's monitoring agent
Agentpatient-monitoring · v1.4
Deployed forDr. Radu Pretorian, Cardiology
Runs onSurogate Cloud
He earns€15–30 / patient / month
Agent activity · live

The agent flags anything that needs Dr. Pretorian's attention immediately - it doesn't make the call itself.

That was one patient. Here's his full patient list.
0patients monitored right now
serving onboarding idle

Build an agent once, then it runs for every patient who subscribes. Each one runs on its own, around the clock - and each one is recurring revenue you don't have to be there for.

Capabilities

Far more than a chatbot - a capable digital worker

Pick a capability to see what it actually looks like in practice.

Works, not just talks
Knows your business
Uses your tools
Browses & operates the web
Handles long jobs
Knows when to ask
Works across your channels
Safe to put to work

The lifecycle

Build, run, review, improve - one continuous loop

Every agent moves through the same loop. Watch it cycle, or click a stage to dig in.

01
Build

Design an agent shaped to your business - its voice, the knowledge it draws on, the tools it uses, the limits it respects.

02
Run

Put the agent to work across the channels your teams and customers already use, doing real end-to-end tasks.

03
Review

Observe how every agent performs, flag what worked and what didn't, and turn real operations into clear signal.

04
Improve

Most changes apply instantly; for high-value tasks, train a specialist from your own data - a compounding advantage.

Every day you run it, your AI gets  better

For developers

Building this for someone else? Start here.

Surogate ships a ready-made kit for each professional category - medical monitoring, legal intake, accounting oversight, tutoring. Customize one for a real client instead of building agent infrastructure from zero.

Medical
Monitoring kit

Patient check-ins, symptom tracking, anomaly flags, weekly reports.

Legal
Intake kit

Client intake, procedural Q&A, contract red-flag review, escalation routing.

Accounting
Oversight kit

Financial monitoring, deadline tracking, proactive client alerts.

Education
Tutoring kit

Practice drills, progress tracking, parent/guardian reporting.

The agency case

Build a kit once, resell it to multiple businesses - like the booking agent one developer built for Bright Dental, now earning €450/month per clinic.

Why not just wire up APIs?

Most workflow builders connect to other people's APIs. Surogate trains, fine-tunes, deploys, and observes the agent itself - on infrastructure you control.

Expert models

Your own work, distilled into experts no one else can build

Fine-tune a small open-weights model on traces only you have. On a narrow task it matches a frontier model - at a fraction of the size, latency, and cost.

Compounding
The Flywheel
turns every day of work into owned intelligence
01
Capture production traces

Every session - prompts, tool calls, approvals, outcomes - is logged with full traces, versioned per agent.

02
Curate the dataset

Traces become training data: filtered for success, deduplicated, labeled, and split - with lineage tracked in the Hub.

03
Distil into an expert

Fine-tune a small base model on your data - SFT, then GRPO or DPO to reward the outcomes you actually want. LoRA or full.

04
Quantize, serve, repeat

Quantized to 4-bit and served behind the same gateway, the expert takes the high-volume tasks - and keeps logging traces for the next round.

Benchmarked on your task · support-ticket resolution
Qwen3-8B → SFT (12k traces) → GRPO → 4-bit AWQ
MetricFrontier APIYour expert
Task accuracy94.1%95.8%
p95 latency2,400 ms380 ms
Cost / 1k calls$9.20$0.40
Model size100B+ params8B params

The moat isn't access to the biggest model - it's the flywheel that turns your work into experts no one else can replicate.

The whole factory

One integrated platform: operate, library, train

Operate
  • Agents
  • Sessions- observe & evaluate
  • Compute- your clusters
Library
  • Models- bring your own
  • Skills & Tools
  • Knowledge Bases
Train
  • Datasets- versioned hub
  • Training- fine-tune & RL
  • Data Hub
Runs in the cloudNothing to host or maintain
No codeBuild in plain language
Bring your own computeOptional - scale on your GPUs
Versioned hubModels & datasets tracked

Stop scaling yourself.
Start multiplying.

Build your first agent in minutes - free, no code, no credit card. Deploy it to the cloud, put it to work, and come back to results.