Questions
Everything people ask before the call.
Costs, timelines, who owns what, and what happens when something breaks at two in the morning. If yours isn't here, ask it directly and we'll answer in a sentence rather than a brochure.
Getting started
We like to move fast! A Starter play is usually live in one to two weeks, and a Workflow build across two or three systems takes two to four. Orchestration work spanning approvals, reconciliation, and dashboards runs four to eight weeks, and a full Platform build eight to twelve. What moves those numbers is rarely the code. It's how quickly we can get credentials, and how long your team takes to agree on what the process actually is.
No. We build around the stack you already run, including Salesforce, HubSpot, Stripe, QuickBooks, Jira, Slack, and most things with a workable API. If a tool in your stack has a genuinely bad API, that surfaces during scoping and we tell you before you commit rather than after. We recommend a switch only when the current tool is the actual problem, and then we say why.
Thirty minutes. We ask what systems hold your data, where the process breaks, who picks up the manual work, and roughly what it costs you each week. You get a realistic scope, a price range, and usually a couple of alternative approaches. Most calls end with a written scoping note in your inbox within a day, which is yours to keep whether or not you hire us.
Mostly B2B firms between 20 and 100 people: IT services, marketing agencies, consulting and professional services. Big enough that manual handoffs genuinely hurt, small enough that nobody has an internal automation team. We take work outside that range, but that band is where we're useful fastest.
Then buy it, and we'll say so on the call. A lot of what gets asked for is already a feature of something you pay for: approval routing in an AP tool, lifecycle management in your identity provider, dunning in Stripe. Building makes sense when the workflow crosses systems no vendor covers, or when your rules don't fit a product's assumptions. We'd rather lose the project than sell you a worse version of a tool you could buy.
That's the usual path and the one we recommend. Pick one workflow that's costing real hours, ship it, and see how we work before signing anything larger. A Workflow-tier build is the normal starting point at $2,400 to $6,000. If it goes well, the next build or a retainer is an easier conversation for both of us.
Pricing and terms
Our rate is $150 an hour and it's published, so every estimate on the site is hours multiplied by it. Starter plays run $900 to $1,800, Workflow builds $2,400 to $6,000, Orchestration $6,000 to $13,500, and Platform work $15,000 to $30,000. Project minimum is $2,500. You get a written fixed-fee quote after scoping, and it doesn't move mid-build unless you add scope.
Starter items price below $2,500, and a single one isn't worth the contracting, scoping, and handover overhead for either of us. In practice they ship two or three at a time, or as an add-on to a larger build. That's usually the better outcome anyway, since small automations tend to be worth more together than apart.
For fixed-fee projects, a 25% deposit at kickoff reserves a build slot. It's calculated from the low end of the quoted range, fully refundable until development starts, and credited against your invoice when it does. The balance is billed at go-live, or at agreed milestones on longer projects.
Fixed-fee scope locks when you sign, which is the point of fixed fee. New asks get quoted as a change order with their own price and timeline. We don't quietly compress the original scope to absorb them, because that's how a fixed-fee project turns into an hourly one nobody agreed to.
You pre-purchase a monthly bucket of hours and we work the highest-value items on your backlog. They start at 10 hours a month on a six-month engagement, then continue month to month with 30 days' notice. Unused hours roll forward one month and then expire, because true hour banking rewards hoarding rather than shipping.
Yes. Mutual NDAs before any commercial conversation if you'd rather start there, and DPAs where the work touches customer data. We don't sign one-sided agreements that bind us indefinitely, and we'll flag anything in your paper that would stop us doing the work properly.
How we build
The founder scopes, architects, and builds it. There are no account managers between you and the person writing the automation, and nothing gets handed to a junior to learn on. That's the trade with a deliberately small firm: you don't get a fifty-person agency, and you never explain your business twice.
n8n by default. It's self-hostable, so execution data can stay inside your own infrastructure, and it's priced per execution rather than per task, which matters at volume. We also build in Make and Zapier when a team already lives there, and natively inside Salesforce or HubSpot when that's the right home for the logic.
Credentials or a sandbox, one person who can answer questions about how the process really works rather than how the handbook says it works, and someone to sign off at the end. Usually a couple of hours a week. The most common cause of a slipped timeline is waiting on access, so sorting that early is the single most useful thing you can do.
Self-hosted n8n inside your own environment, so payloads never transit our infrastructure. Every run writes an audit log, high-risk steps sit behind approval gates, and credentials live in your secrets store. We design against HIPAA, SOC 2, and FINRA control requirements. To be clear about what that is: it's how we build, not a certification we hold, and your auditor still signs off on the result.
Often yes, and it's a common way engagements start. Usually it's an inventory first: what's running, what's silently stopped, what's firing on nothing. Plenty of what we find is fine and should stay. The parts worth rebuilding are the ones with no error handling, no alerting, and no record of what they did.
Yes, and it's often the right shape. We build the first version and set the patterns, your team maintains and extends it. When that's the plan we write for the handover rather than for ourselves, which changes how much gets documented and how clever the code is allowed to be.
After launch
Expected, and designed for. Every build includes 30 days of post-launch tuning, which is when the edge cases nobody thought of turn up. After that you can run it in-house from the documentation, or put us on a retainer. Either way the workflows are yours to change.
Every build ships with retries, error handling, and alerting, so a failure pages a human instead of sitting quietly until month-end. The failure mode that actually costs money is the sync that stopped three weeks ago and nobody noticed, which is why jobs report success as well as failure. Response is next business day, with faster windows by agreement.
You do, completely. Source, workflows, platform configuration, and documentation, delivered into your repository and your accounts. You can extend it, change it, or hand it to another vendor without involving us. There's no proprietary framework holding anything hostage.
It depends entirely on what the manual version costs you now, which is why we'd rather you model it than take a number from us. The calculator runs the same math we use in scoping: hours saved multiplied by a fully loaded hourly rate, with a cumulative view over time. Bring the output to the call and we'll pressure-test the assumptions.
Still wondering?
Ask on a 30-minute call. You'll leave with a scope and a price range whether or not you hire us, and the catalogue already lists what most of this costs.