Survey
Fixed fee · 2 hours
Map your business. Leave with a plan.
- Two-hour discovery call with the operator
- Workflow, stack, and risk mapping
- A written plan of work, yours to keep
- No pressure to continue
An AI operations studio · Melbourne
We build AI operations for Australian businesses.
OpsDock designs, builds, and runs the AI side of your business. Installed in fourteen days. Supported for ninety. We work with a small number of clients each quarter.
Recent work by the operator
4
clients per quarter
14
days to live
90
days of joint run
29
skills in the index
The stack
Built on the AI you trust. Wired into the
channels you live in.
We build on the models that fit your data policy, then connect them to the tools and channels your team already uses every day.
Models we build on
OpenClawNemoclaw
Claude Code
ChatGPT
GeminiChannels we integrate with
Telegram
WhatsAppiMessage
SlackPlus the tools you already use
Xero
HubSpot
Shopify
Google Workspace
Google Ads
Meta
LinkedIn
Mailchimp
KlaviyoOkendo
Microsoft 365
Notion
Airtable
Monday.com
Fireflies
Stripe
n8n
Zapier
Supabase
VercelServices
Three ways AI ops gets into your business.
Most operators start with the middle path. The survey gives you a plan; the install gives you a working system; the run gives you a team that owns it.
Most chosen
Install
Fixed fee · 14 days to live
We build the agents and wire them in.
- Agents, integrations, dashboards, runbooks
- Deployed inside your environment
- Human-in-the-loop with full audit trail
- Live by day fourteen, not theoretical
Run
Monthly retainer · 3-month minimum
We run it with you for ninety days.
- Joint operation with your team
- Weekly written ops reports
- Tuning, hardening, escalation coverage
- Your team takes the helm by day ninety
How it works
Four phases. Three months. One operator.
01
Survey
Two-hour discovery. We map your business, your stack, and the work you would most like to give to a machine.
02
Build
Fourteen days. We design the agents, wire the integrations, set up dashboards, and document escalation paths.
03
Install
On day fourteen your team gets credentials, runbooks, and a private channel with the operator who built it.
04
Run
Ninety days of joint operation. We monitor, tune, and harden. By day ninety, your team owns it.
Patterns from past engagements
What it tends to look like.
Indicative ranges drawn from past engagements. Your numbers fall out of the survey call, not the website.
DTC brand
8 figuresReturns triage, supplier follow-ups, post-purchase comms.
15 to 25 hours of founder time recovered each week.
Brokerage
30 to 80 staffInbound lead triage, document workflows, inspection coordination.
30 to 50% faster routing on new enquiries.
Agency
10 to 25 staffClient reporting, intake forms, internal QA against playbooks.
8 to 12 hours per team member recovered each week.
Trades and construction
Family-ownedQuoting, supplier comms, job tracking, after-sales follow-up.
Owners back to a 6pm finish inside two months.
What we install
The index.
We've installed most of the items here at least once. Some are off-the-shelf, some are bespoke. None are theoretical.
Talk to us about a specific oneCustomer-facing
- ·Inbound email triage and routing
- ·Lead qualification and enrichment
- ·Post-purchase comms sequencing
- ·Returns and refunds workflows
- ·Bookings, reschedules, no-show recovery
- ·Review requests and follow-up
- ·Support ticket classification
- ·Knowledge-base and FAQ agent
Sales and CRM
- ·CRM hygiene and deduplication
- ·Account research and brief generation
- ·Outbound sequencing
- ·Proposal and quote drafting
- ·Pipeline health reports
- ·Loss-reason capture
Internal ops
- ·Weekly written ops reports
- ·Daily dashboard generation
- ·Meeting note capture and routing
- ·Executive inbox triage
- ·Calendar coordination
- ·Document drafting from templates
- ·Internal QA against playbooks
Vendor and supplier
- ·Supplier email follow-up
- ·PO generation and tracking
- ·Invoice extraction and routing
- ·Quoting from supplier feeds
Bespoke tooling
When off-the-shelf falls short,
we build the tool.
MCP servers, connectors, dashboards, agent skills, payment gateways — whatever the job actually needs. Shipped with security and engineering best practices baked in, not bolted on.
Waydock
Operator dashboardA daily cockpit for the metrics, queues, and agent runs an operator actually needs to see before their first coffee. Built when the off-the-shelf dashboards drifted out of sync with the workflows behind them.
Chest Gate
Pay-per-call gatewayx402 payment gates for APIs, agents, and MCP servers. Wallets are managed by Privy so keys never touch our infrastructure, and every request is metered, signed, and auditable.
Principle
Security by default
Least-privilege credentials, scoped tokens, signed requests, audit trails. We assume the agent will be wrong sometimes — and design the blast radius accordingly.
Principle
Built on standards
MCP for tool servers, OAuth and OIDC for identity, x402 for paid endpoints, Tailscale for private networking. Boring choices that age well.
Principle
Yours to keep
Source in your repos, deployed inside your environment, documented for the team that will run it. We are not the dependency.
Daniel Montoya · Melbourne
The operator
Daniel Montoya.
Engineer and founder, twenty years building production systems for businesses that needed to keep running while they were being rebuilt. OpsDock is what I'd hire if I were running your business today.
I take on four clients a quarter. If we're a fit, we'll both know by the end of the first call.
Clients
4 / quarter
Base
Melbourne
Stack
Your own
FAQ
Questions,
answered.
The things operators ask before the first call. Anything not here gets answered on it.
Ask us directlyWhat does OpsDock do?
OpsDock is an AI operations studio. We design, build, and run the AI layer of your business — the agents and integrations that handle repetitive operational work like inbound triage, follow-ups, reporting, and document workflows. We install a working system in fourteen days and run it with your team for ninety.
How much does it cost?
There are three engagements: a fixed-fee Survey (a two-hour discovery call with a written plan of work), a fixed-fee Install (a live system in fourteen days), and a monthly Run retainer with a three-month minimum. Exact figures come out of the scoping call, sized to your workflows — not the website.
How long does it take to get live?
Fourteen days from the start of an Install. In that window we design the agents, wire the integrations, set up dashboards, and document escalation paths. On day fourteen your team gets credentials, runbooks, and a private channel with the operator who built it. Then we run it jointly for ninety days.
Do you only work with Australian businesses?
OpsDock is built in Melbourne and focused on Australian owner-operators, but the work is delivered remotely and we work with clients globally. Everything is deployed inside your own environment, so location is not a constraint.
Who runs OpsDock?
Daniel Montoya — an engineer and founder with twenty years building production systems for businesses that had to keep running while they were being rebuilt. He takes on four clients a quarter and is the operator on every engagement, so you work directly with the person who builds your system.
How do you handle security and my data?
Security by default: least-privilege credentials, scoped tokens, signed requests, and full audit trails. Agents run human-in-the-loop inside your own environment on the AI models that fit your data policy. The source lives in your repositories — you keep everything, and we are not a dependency.
What happens after the ninety days?
Your team owns the system. Across the ninety-day Run we monitor, tune, and harden while handing over control, so by day ninety your people are operating it themselves with runbooks and documentation. There is no lock-in — the source and deployment are already yours.
Ready when you are
Ready to get AI ops into your business?
We build it. We install it. We run it with you for ninety days. Every engagement starts with a scoping call. No templates, no canned decks.
Six short fields: the two workflows you would most like to give to a machine, the stack you run them on, and what good looks like in ninety days. We reply within a business day.