SteelWorks Intelligence · an autonomous operating system for business
SteelWorks designs, deploys and operates AI agent systems that carry the work a business depends on: operations, being found by AI search, the web and app layer, and the security around all of it. Every engagement starts with an assessment that says exactly where you stand — evidence first, then the fix list.
Built for any business that runs on the internet — local service firms, professional practices, agencies, e-commerce, software. Based in Acworth, Georgia; delivered remotely.
Four pillars, one operating system
Agent systems that run scheduled work — intake, follow-up, reporting, monitoring — on hardware you own, routed to free and local models, with approval gates where a person should decide. Built from the operating system that runs SteelWorks.
Operations build & careWhen a customer asks ChatGPT, Claude or Perplexity who to call, the answer is read from your pages, not from Google's ranking. We audit what assistants can read about you, fix what stops them, and keep the score holding month after month.
Audit, fix sprint, retainerThe layer your operation runs on: sites that pass the same readiness gate we sell, internal dashboards that read live state, and workflows wired to the tools you already use. Nothing ships until it clears the gate.
See a demo buildAlready running agents or automations? We score who can do what, where approvals sit, and how secrets and inbound mail can be abused — against the NIST AI Risk Management Framework, with every finding owned and dated. Evidence, not a certificate.
AI Security & Risk AuditStart here
Assessment 1
Your live site scored on five legs: whether AI crawlers can read it, what the engines already believe about you, who gets cited instead of you, what the rest of the web says, and the fixes written for your site. Re-check included after you apply them.
Assessment 2
For a business already running agents, copilots or automations. A 25-control scorecard against the NIST AI RMF: permissions, approval gates, secret handling, inbound-content abuse, logging. Every finding is owned, dated and comes with the fix.
Assessment 3
The wide view across all four pillars: how the business runs day to day, where it is invisible to AI search, what the web and app layer costs you, and where automation would pay for itself first. Scoped from a three-line intake, priced before any work starts.
Not sure which one? Compare the three, or run the free AI answer readiness check on your own URL first — it is the same ten checks we run on our pages.
After the assessment
Fixed scope, fixed price, measured before and after. The Visibility Fix Sprint applies every audit finding and re-scores your site. The AI Operations Build puts a governed slice of our operating system on your hardware: three to five automations, routing to free and local models, reporting to your phone, approval gates.
Ongoing
The system keeps running and keeps reporting. A fresh visibility audit every month, a content engine built from your real proof, a lead engine that sources and follows up at a human pace, and care for the operations we built. Cancel any month.
Run on ourselves first
AI agents, each with a written role and a tool scope
scheduled jobs running the company around the clock
skills available to the agents, from research to publishing
pillars: operations, visibility, web & apps, security — the same four we deploy for you
Real audits
Names withheld until the owners say yes; the scores and the findings are exactly what each report said. Rows and the scorecard beside them are redrawn from the audit log on every deploy.
The most common miss is the simplest: a business whose name, address and phone are not stated identically on the page and in its schema. An assistant that cannot confirm who you are will not recommend you, whatever your Google ranking.
Resource Library
Free instruments, paid assessments, working automation systems, growth playbooks and prompt packs — every one pulled from the fleet that runs SteelWorks. Checkout is on Gumroad; the library is where each resource is explained and supported.
Checkout opens Gumroad in a new tab and delivers the files the moment payment completes. Every resource page here explains what is inside, who it is for, and which service deploys it if you would rather have it run for you.
Straight answers
Any business whose revenue depends on being found, answering fast and running reliably online: local service firms, professional practices, agencies, e-commerce, software. The assessments are sized for a single business; builds and retainers scale from there.
With an assessment. The AI Search Readiness Audit is $97 and takes 72 hours. The AI Security & Risk Audit is $1,500 to $3,000. A broader Digital Intelligence Assessment is scoped through the intake form. Each produces a scored report you keep whether or not you go further.
No. Every fix is written for the site and stack you already have. Operations builds run on hardware you own, with approval gates where a person should decide, and are handed back to you with documentation.
What we will not do
No. It scores your AI operation against the NIST AI Risk Management Framework and hands you owned, dated fixes. It is evidence for your own governance, not a certificate, and it does not guarantee that nothing can go wrong.
Invent reviews, fake numbers, or claim a result that was not measured. If a check could not be run, the report says so and why. Retainers can be cancelled any month.
SteelWorks Intelligence, Acworth GA. The audit engines, the agent fleet and the gates are ours; the judgement, the access and the accountability are a person's.
Not ready to buy
Leave your email and your website. We run the technical check and send you the top three findings — the same ones the four businesses above received — within two business days.