Work that runsWork that runsWork that runsWork that runs itself.

Automation follows a fixed script: do this, then this, then this, every time, in the same order. An agent decides its steps as it goes, because the right move changes depending on what it finds. Automation is cheaper per outcome when the steps are fixed, and more predictable than an agent. If a person currently has to stop and think before acting, that is agentic — that lives at /agentic. Most businesses need some of both.

If you can already write every step down in order with almost no branching, it is a good candidate. Clear inputs, clear outputs, systems with APIs or webhooks, and a person who hates doing the same clicks every week. The $500 Automation Audit is the start — process map, go/no-go, topology, and a fixed-scope quote so you stop guessing before a bigger check.

Open-ended judgment with no clear success criteria. Jobs where being wrong once is expensive and hard to undo. Processes that change every week because nobody has written the real steps down. Tools with no API and no reliable export. If the hard part is deciding, not clicking, I will send you to Agentic instead of selling you a brittle workflow that pretends to think.

Yes — that is most of the job. Tier 1 usually touches two or three systems. Tier 2 and Tier 3 connect whatever the ops stack actually uses. If a tool has an API, a webhook, or even a clean export path, there is almost always a way in. The audit maps which handoffs are worth wiring and which should stay manual.

Either. I can stand up cloud n8n, use an instance you already have, or include self-host setup on Tier 2 and Tier 3. What matters is that the workflows end up as JSON on accounts you control. I am not going to trap your ops graph inside a black-box account you cannot leave.

Buy automation when the steps are fixed. Buy an agent when the steps change based on what it finds. Cheaper does not always mean a smaller invoice — a department-scale multi-workflow stack can still cost real money — but per outcome, fixed scripts beat agents when the path never branches. Unsure? Start the $500 Automation Audit. I will not upsell you into Agentic to inflate a build.

Three to five business days of process mapping: the bottleneck, systems touched, go/no-go on what is worth automating, a rough node topology, and a fixed-scope quote for the build. You keep the deliverables either way. The $500 comes off any project — automation or agentic — with no expiry. If the map says you need agents, I offer the $1,500 agentic pilot at /agentic; decline and the credit still applies to an agentic build.

I have built 500-plus automations. More than 600 have shipped across the book of work; 500-plus are still running in production. Those are separate counts on purpose — built is not the same as still running, and I will not blur them. Aggregate client busywork deleted across that work sits at 35,000-plus hours saved. I do not invent per-client metrics to dress that up.

Affordability tracks the pain of the bottleneck, not your headcount. A five-person shop drowning in the same weekly report is a better fit than a thousand-person company with no single painful job. Smallest real builds start around $2,500. Under that, buy a ready-made workflow and we adapt it to your stack. Start with the $500 Automation Audit if you are not sure the math works.

Start the $500 Automation Audit from the contact page. From there it is me — William Spurlock — not a sales floor handing you to a project manager. Bring the process, the tools it touches, and one decision-maker. I will tell you whether you need a ready-made workflow, a scoped build, or Agentic.

Because the invoice should track the bottleneck I actually see — systems touched, failure modes, self-host needs, and how messy the real data is — not a sticker I invent from a contact form. Builds are value-priced after the $500 audit. You get a fixed-scope quote before I write nodes. No printed ladder that pretends every CRM sync costs the same.

Smallest real builds start around $2,500. Under that, buy a ready-made workflow and we adapt it to your stack — fighting a custom build below that floor usually costs more in discovery than you save. The $500 Automation Audit tells you which path you are on before you buy.

Yes — the full $500, with no expiry. It comes off any project — automation or agentic — not just automation builds in a ninety-day window. You keep the process map and go/no-go either way. If the audit says agents and you decline the $1,500 agentic pilot, the credit still applies to an agentic build.

Watch is $3,000 a month for monitoring, dead-letter handling, and a small change budget. Keep Current is $6,500 a month for a larger change volume and new workflows on a cadence. Keep Building starts at $15,000-plus a month for department-scale stacks — priced on the size of the system. Commit to three months and take 25 percent off. Go annual and stack that same 25 percent with two months free.

Yes. Monthly is full price, cancel anytime. A three-month minimum takes 25 percent off. Annual takes that same 25 percent off and adds two months free — Watch lands at $1,875 a month billed annually ($22,500/yr), Keep Current at $4,063 ($48,750/yr), Keep Building at $9,375-plus ($112,500-plus/yr). Longer terms cost less per month once you know you are staying.

Yes. Refer someone who signs a three-month minimum and you get two more months free on your own plan. It works on Watch, Keep Current, and Keep Building, and it stacks with annual. Same stacking language as the other services — you send business my way, I send months back yours.

After the quote is locked: 50/50 on smaller builds — half to start, half at launch — or milestones on larger stacks. The audit is one-time up front. Retainers bill monthly (or under the term you pick). I do not take a deposit against a mystery scope.

Yes, on your own accounts. Cloud n8n, self-hosted compute, and any third-party API costs sit on invoices you control. I do not mark those up or route them through Spurlock Studios. That is also why credentials and ownership stay with you from day one.

When the steps are fixed, automation is usually cheaper per outcome and easier to predict. That does not mean the invoice is always smaller — a Tier 3 department stack can still be a serious build. Agents cost more to design and to keep honest because judgment needs evaluation and oversight. I will push you toward automation when the script holds, and toward /agentic when it does not.

I tell you straight and offer an upgrade into the $1,500 agentic pilot at /agentic. If you decline, the $500 audit still credits toward an agentic build — same no-expiry credit that applies to any project. You keep the process map either way; the audit is not wasted if the lane shifts.

Buy a ready-made when the job already matches a shelf workflow and you just need it adapted to your stack — usually under the ~$2,500 custom-build floor. Start the $500 audit when you need a process map, go/no-go, topology, and a fixed-scope quote before you commit. If you already know the scope, book a build tier scoping call after the audit — not as a free start.

I hold all Make.com AI Automation certifications and I still ship a lot of work on Make when that is the right tool. For production graphs I own long-term, n8n wins on self-host, JSON ownership, and how the platform is moving — I collaborate directly with the n8n team. You get workflows as files on your infrastructure, not a seat tax you cannot leave. If Make is already your stack and the job fits, I will say so instead of forcing a migration for ego.

Cloud is fine for a single Tier 1 pipeline when you want speed. Self-host when the data should not leave your network, when you want full control of upgrades, or when the graph is big enough that residency matters — Tier 2 and Tier 3 include self-host setup for that reason. Either way, credentials stay on your accounts and the JSON stays yours.

You do, outright. Workflows are JSON files on your instance. No vendor lock-in, no per-seat tax to keep your own pipelines running, no hostage export screen. You could hand the files to another developer tomorrow and they would still run.

Every build ships with idempotency, a dead-letter queue, schema contracts, and human gates on anything with a real-world side effect. Tier 2 adds observability and replay tooling. Tier 3 adds department-scale monitoring and deeper training. Hardening is not an upsell — brittle happy-path demos are how automations earn a bad name.

Failed runs get parked in the dead-letter queue instead of silently dropping. You (or I, under a retainer) review the payload, fix the cause, and replay when it is safe. Observability on Tier 2 and above makes that visible without digging through raw logs. Watch and higher retainers include ongoing dead-letter handling so production does not rot.

Deterministic by default. Light AI nodes for extract or classify when they earn keep — for example pulling fields from messy email. I do not glue a model onto a fixed script for marketing copy. When the job needs real judgment mid-flight, that piece belongs in Agentic, not hidden inside an automation invoice.

Tier 1 lands in 1–2 weeks. Tier 2 in 2–4 weeks. Tier 3 in 4–8 weeks. The Automation Audit is three to five business days before any of that. Timelines assume credentials, a process owner, and fast feedback — stalled access stretches the calendar, not a surprise invoice.

Credentials, API access, a process owner who does the job today, sandbox or real sample data, one decision-maker, and feedback in days. Missing those does not always stop the work — it just means we fix access and clarity first. The $500 Automation Audit surfaces gaps before I start wiring nodes.

Yes. That is the point of JSON ownership and training. Tier 1 ships with a runbook and credentials setup. Tier 2 and Tier 3 include deeper training so your ops team can extend what shipped. Under Keep Current or Keep Building I can keep shipping changes with you — but you are never locked out of your own graph.

Three different numbers, kept separate on purpose. I have built 500-plus automations; more than 600 have shipped across the book of work; 500-plus are still running. Aggregate client busywork deleted sits at 35,000-plus hours saved — that compounds across clients, so it is higher than any one personal hours figure and I will not pretend it is a single-client case study. Ask for the distinction if a competitor blurs built, running, and saved into one vanity metric.