Setanta

Where to start

Where do we even start with AI in a mid-size manufacturing company?

Most AI plans stall because they start with the tool. Setanta is an Anthropic Partner that helps US manufacturers and steel companies of 50 to 600 people start with their best person instead: find the hardest part of a frontline job, like the estimator's or the dispatcher's, and make every seat work the way that person does.

Gary O'Neill, founder and CEO of Setanta

Gary O'Neill, Founder & CEO, Setanta.

The short answer

How should a mid-size manufacturer start with AI?

Start with a person, not a platform. Pick the job your business leans on most, find who does it best, and watch where their day goes. Then put Claude to work on the reading, matching and retyping around their judgment, with a person approving anything that changes price, inventory, quality or a customer commitment.

For a steel service center that is often the inside sales desk. For a fabricator, the estimator. For a carrier, the dispatcher. The answer comes from your own floor, not from a list of use cases, which is why we look before we recommend anything.

Why it matters

The board wants a plan. The floor wants its day back.

The board keeps asking what the AI plan is. A vendor offers a platform. A consultancy offers a roadmap. Meanwhile, the people who would gain most, the estimator, the inside sales rep, the dispatcher, are the busiest people in the building and have no time to work out how AI fits their job.

That is why so many companies buy seats that nobody opens. The tool was never the hard part. The hard part is the frontline person's day: the RFQ that needs retyping, the order that needs checking, the system with no connector that still holds the answer.

Setanta makes your whole company as good as your best person. A good AI plan for a company your size is a short list of jobs, the person who does each best, and the number that will tell you it worked.

How we do it

From the board's question to a number that moved.

  1. 01

    Operating Diagnostic

    Free, and the first step

    We look at how your team works today, where the hours go and which numbers matter to you. We find the job where your best person's way of working would move the most, and check what your systems can share, including the ones with no ready-made connector.

    Done when You leave knowing where to start and which package fits, or that it is not time yet.

  2. 02

    Launch

    Claude goes to the whole team, trained on the work they already do. It is the fastest way to show the board progress and to find out who will own the next step.

    Done when The team would push back if you took it away.

  3. 03

    Build

    We rebuild that one job end to end, run the way your best person runs it, with a person approving the steps that matter.

    Done when A number you agreed on has moved against where it started.

  4. 04

    Scale

    What worked spreads to three or four processes across up to three departments, including one customer-facing workflow.

    Done when Someone in-house keeps it going.

  5. 05

    Run

    A retainer for monitoring, fixes and the next improvement once the work is live.

How long each step takes is scoped in the Operating Diagnostic.

What you get

What you can take to the board.

  • The job to start with, who does it best, and why it comes first.
  • The number that will show it worked, measured on your own data.
  • A team using Claude on real work, not a pilot nobody remembers.
  • Everything in your own accounts, with Claude licences billed to you by Anthropic.
  • A straight answer if AI is not the right next step.

Where this fits

What we can show you, and what we cannot yet.

A quote-heavy business we work with took a client proposal from 5 days to 5 minutes. It is not a manufacturer, and we have no manufacturing or steel client yet.

If you already have someone inside who knows both the work and the tools, with time to own it, you may not need us for this step. Our guide on how to choose an AI implementation partner is written so you could use it to turn us down.

Questions

Asked before every first call.

Do we need an AI readiness assessment before we start?

You need to know where the hours go and whether your data can be reached. You do not need a long report to find that out. Our free Operating Diagnostic covers it and ends with where to start, or an honest not yet.

What should I tell the board about our AI plan?

Tell them which job you will start with, who does it best today, what number will move and how you will measure it. Boards respond better to one result on your own numbers than to a roadmap. Then tell them what comes next.

Are manufacturers getting real results from AI, or is it mostly hype?

Both are true. Results come where AI is built into a real job, with the people who do it, and measured on one number. Hype comes from seats bought with no job in mind. We would rather show you one moved number than promise many.

Should we add AI on top of our ERP or wait for our ERP vendor's AI features?

Turn on what your ERP offers if it fits a job you have. Most ERP features cover the data inside the ERP, while much of the work sits in email, PDFs and spreadsheets around it. We work on top of the ERP you run, so you do not have to wait.

What is the free Operating Diagnostic?

A free call (book a time) that looks at how your team works today. It finds where Claude would move your numbers first and the process worth fixing first, and says which package fits. If it is not the right time, we say so.

See the Operating Diagnostic

Board asking about AI? Tell us how the work runs today.

We will say where we would start, or that it is not time yet.

Book a call