Several agents work the same event at once.
Intake, research, and capacity read the same inquiry simultaneously. Three agents cost one agent's worth of waiting, and each produces its own checkable output.
10 templates · 50 agents · installed, not rented
Agents that run
your business.
In parallel.
Not a chat window. A system: agents that work the same event at once, chain in order once they report, and fire themselves on triggers and timers. You keep one gate, on your phone, for anything that cannot be undone.
$2,500 install $400/month after 30 days Live in 72 hours Cancel anytime
Jamie / 5-agent systemInbound Lead Responder
Live Trigger
Website form, 8:47 PM: "Need a quote for a water heater replacement, 50 gal gas, Folsom. How soon can someone come out?"
Output
Hi Dana, thanks for reaching out. A 50-gallon gas water heater replacement in Folsom typically runs $1,900 to $2,600 installed, depending on venting and code updates. We can have a tech out tomorrow at 9:00 AM or 1:30 PM to confirm the exact price on site. Which works? — Mike, Northside Plumbing
Trigger to send: 71 seconds. Three agents ran in parallel, two in series, policy agent cleared it for auto-send because job type, area, and price band were all inside the rules.
Every template page shows the whole topology: what fires it, what runs at the same time, what runs in order, and exactly where it stops to ask you.
Three primitives
How the systems actually run.
Same three ideas in every template. What changes is the job, the triggers, and which agents do the work.
The next agent cannot start until the last one reports.
A writer merges what the parallel stage found. A separate policy agent audits the result. Chaining is what makes the output specific instead of plausible.
Triggers and schedules, not someone remembering.
A form webhook at 8:47 PM. A day-10 timer on an unpaid invoice. A 6 AM sweep. The system starts its own work and arms its own follow-ups.
10 templates
Pick the job. We install the graph.
Every template has a fixed price, a fixed install, and a page showing the full agent topology working on a real case.
Inbound Lead Responder
Four agents wake on every inquiry, answer it in under two minutes, and arm the follow-up.
- 5 agents
- 4 triggers
- Home services
Quote Chaser
A scheduler that wakes on its own clock and runs a three-stage chase until every quote is won or lost.
- 5 agents
- 3 triggers
- Contractors
Win-Back Bot
A monthly cadence that segments your own customer history and builds approval-ready batches while you sleep.
- 5 agents
- 3 triggers
- Cleaning companies
Review Engine
A watcher on your listings plus a timed asker on your job list, running on two independent clocks.
- 5 agents
- 3 triggers
- Home services
Invoice Chaser
A ledger watcher with its own escalation clock, and a classifier that knows a dispute from a delay.
- 4 agents
- 4 triggers
- Agencies
Proposal Drafter
Four agents take a call transcript apart in parallel and reassemble it as a priced proposal in your template.
- 5 agents
- 3 triggers
- Agencies
Support Desk
A retrieval agent, a drafter, and a confidence judge, with escalation as a first-class outcome.
- 4 agents
- 3 triggers
- Software companies
Founder Chief of Staff
Four agents sweep your calendar, inbox, commitments, and pipeline overnight and converge on five decisions.
- 6 agents
- 4 triggers
- Founders and owners of 5 to 50 person companies
Outbound Researcher
A daily trigger scan across your target accounts, then parallel research and one approval-ready draft each.
- 5 agents
- 3 triggers
- B2B services
Search Revenue Engine
A weekly loop that mines your own search data with four agents and ships approved fixes.
- 6 agents
- 4 triggers
- Local services
Wire. Tune. Go hot.
A 90-minute call. We wire the triggers to your forms, inbox, and calendar, capture your voice, and write the approval rules: what each agent may do alone and what waits for you.
Every agent in the graph is tuned on your real history, then the whole system runs against past cases. You watch it work before it touches a customer.
Triggers go hot and an approval desk lands on your phone. Thirty days of tuning follow before the first monthly charge.
What you getOne install.
One install.
One managed desk.
The install is real work: wiring triggers, configuring every agent in the graph, and dry-running the whole system on your history. The desk keeps it running, tuned, and honest every month after.
- 90-minute kickoff: accounts, voice, rules, and approval boundaries
- Triggers wired to your forms, inbox, calendar, and tools
- Every agent in the topology configured on your real history
- A dry run on past cases before anything reaches a customer
- An approval desk you can run from your phone
- Live within 72 hours of kickoff
- Hosting, scheduling, and the trigger watchers, 24/7
- Claude usage up to 1,000 agent runs a month
- Rule, prompt, and routing changes whenever you ask
- A 30-minute tuning call every month
- Weekly report: what ran, what it decided, what it escalated
- Cancel anytime; the bots run to the end of the paid month
Install a systemBuy the graph.
Buy the graph.
Keep the desk.
One install for one business. The managed desk starts after 30 days, so the first month has one job: prove the system deserves to stay.
Today$2,500installation
thenAfter 30 days$400per month · cancel anytime
- Live within 72 hours of kickoff
- Every agent tuned on your history
- Triggers wired to your tools
- Approval desk on your phone
- Monthly tuning call
- Direct founder support
Fleet of 3$6,000 install, then $900 a monthThree systems wired to hand off to each other. Saves $1,500 up front and $300 a month against three separate installs.
Prefer to talk first? Book 25 minutes Runs on Claude
Built on Anthropic's Claude. Installed by a human who answers email.
Judgment-heavy agents run on Claude Fable 5.1. Routine agents use smaller Claude models so your monthly cost stays predictable. Your data stays in your accounts; the agents work through the connectors you authorize and nothing else. Jamie is an independent service. Claude and Anthropic are trademarks of Anthropic, PBC; no affiliation or endorsement is implied.
Plain answers.
What is a "bot" here, exactly?
A system of agents with a defined graph: a trigger that fires without a human, several agents that run in parallel, a chain that runs in order once they report, and an approval gate. Each template page shows the whole graph.
Why not just one big prompt?
Because one prompt that reads your calendar, recalls the customer, and writes the reply will quietly get one of the three wrong, and you will not know which. Separate agents produce separate checkable outputs, and a separate policy agent audits the result before anything sends.
Why a template instead of custom?
Each graph has already been built and broken in. Installation means fitting it to your voice, rules, tools, and history, not designing a topology from a blank page.
What is in the $400 a month?
Hosting, the trigger watchers and schedulers, Claude usage up to 1,000 agent runs a month, rule and routing changes whenever you ask, and a monthly tuning call. Usage above fair use is billed at cost, and you will see it coming.
Can the agents act without me?
Only inside the policy you approve on day one, and only at the stages marked auto-send on the template page. Every graph has a gate, and you can tighten it to "approve everything" at any time.
Which model does this run on?
Claude, by Anthropic. Judgment-heavy agents run on Claude Fable 5.1; routine agents use smaller Claude models to keep your monthly cost predictable. Jamie is an independent service and is not affiliated with Anthropic.
What if it does not work for us?
You have 30 days before the first monthly charge, and you can cancel the managed desk anytime after. The install fee covers real work and is non-refundable once kickoff begins, which is why we start with a call.
The honest contract
No promised revenue. No invented proof. No autonomous spam. You own your accounts, your data, and every decision that cannot be undone. We build the graph and keep it honest.