Foundational AI Consulting Book a 30-minute callBook a call
A physical cutaway model of a small commercial building standing on a slab, with the earth beneath cut away to show concrete piles driven through topsoil, clay and gravel into bedrock.

Built into
the systems
under your
business.

ERP first, and the shop-floor and customer systems around it. We choose and run what fits, rescue what stalled, build AI into it, and run it until the results show up in your numbers.

Book a 30-minute call Free. No deck, no pitch.

Three things we do. In that order.

Most ERP projects fail the business that paid for them, and most AI projects stall between the strategy deck and the first working system. We take on the whole thing, all the way down to something that holds.

  1. Put the right ERP under your business.

    An assessment of how you take orders, make, ship and close the books, ending in the ERP that fits and a go or no-go on each AI opportunity inside it. We run the candidates side by side on your kind of data before we recommend one: Odoo, ERPNext, Aureus, Dolibarr, Acumatica and more. If yours has stalled, we move you off it; if it is right, we keep it and work on what sits around it.

    You get a fit report, a business case for every opportunity, and a plan you can run with us or without us.

  2. Build AI into it, and into what runs beside it.

    Plain-English answers across your ERP, shop-floor and customer systems, supplier invoices turned into draft bills, a month-end close that finds what is missing, dashboards and guided flows your own people describe in a sentence, and planning models that read the ERP and the manufacturing execution system together: delivered inside the tools your people use every day.

    You get a working system, tested against goals agreed before the work started, that shows its sources for every figure.

  3. Run it until the numbers move.

    We host, operate, monitor and tune the ERP, the systems we connected to it and what we built into them, and we report against the outcome we agreed, not against activity.

    You get a monthly review, a quarterly value report, and a team of your own that can take it over.

Close up of one concrete pile of the model passing down through ochre clay and grey gravel into pale bedrock.

What holds is what you can't see.

The systems that keep paying are the ones built into how the business already runs: the order desk, the shop floor, the monthly close. Usually that is the ERP and the systems beside it on the floor, and it is where the AI has to live. Not a pilot on the side, not a tool nobody opens. The work is in the footings.

What an engagement looks like.

A manufacturer's production schedule, and the AI built into its ERP and shop-floor system.

An illustrative example. A composite of the kind of work we take on, written to show how an engagement runs; not a named client.

  1. The situation

    A mid-sized manufacturer ran orders, inventory and costing in its ERP and the shop floor in a separate manufacturing execution system. The two synced every few minutes, but neither said what should run next, what it would cost to change over between products, or whether a promised date would hold. Planners kept a two-week buffer, a spreadsheet, and a habit of walking the floor to find out.

  2. What we found

    The assessment ranked the opportunities. Changeover cost and due-date confidence were the clear first case: the history was already in the two systems, the rules were the planners' own, and every hour of changeover was an hour of lost output. Two other ideas got a no-go: one because the quality records were not reliable enough yet, one because the payoff did not justify the change. The client got a business case for each and a plan they could run with us or without us.

  3. What we built

    A planning view inside the systems the team already used. It reads six months of run history from both, works out what a changeover between any two products really costs, shows the capacity picture with a confidence behind every due date, and lets a planner try moving an order and see the knock-on before committing. Every figure shows the records it came from; a planner confirms every change. Nothing was ripped out. The goals were agreed before work started: hours lost to changeovers, and orders promised with confidence.

  4. Where it is now

    It runs as a managed service. The planners review the model's suggestions monthly with us, the goals are reported against every quarter, and the client's own people are learning to run it.

If that sounds like your operation, that is what the first call is for.

How we work.

Four commitments, agreed in writing before the first project starts. They hold for every project after it.

Ways to work with us.

Each one stands on its own. Most engagements start with the assessment and go deeper from there, one step at a time.

Assessment Start here
A fixed-scope look at your ERP, the shop-floor and customer systems around it, and where AI pays off across them, with read access and your recent history, ending in what fits, ranked opportunities and a pilot recommendation.
Pilot build
One well-scoped system, built and running: a rescue migration, an AI feature on your live data, a connection between your ERP and the floor, or a first planning model, measured against the goals agreed before it started.
Managed service
We host, run, monitor and improve the ERP, what we connected to it and what we built into it, with a monthly review and a quarterly value report.
Advisory retainer
Monthly strategy sessions, roadmap governance and executive briefings for teams building their own AI fluency.
A model-maker's bench in morning light: a steel scale rule, a pencil, a folded drawing, three glass jars of cork, sand and gravel, a plaster block and a green notebook.

Who you work with.

A technical partner who architects and builds. A business partner who scopes, prices and keeps the work honest. Both roles are on every engagement, from the first call to the last report, working from Seattle and Denver.

Book a 30‑minute call.

Tell us what your systems make hard, or what you would want AI to take off your team's plate. No deck, no pitch.

Pick a slot that suits you; the invite arrives with a Meet link. Prefer to write first? hello@foundationalaiconsulting.com

How the free call works

  1. 30 minutes on the phone about your problem, with a partner, not a salesperson.
  2. Within a few days, a one-page note: what we would assess, what it would cost, or why we are not the right fit.
  3. Paid work starts only if you say yes. Nothing is billed for the call or the note.

Pick a time for your call.

Choose a slot; Google asks for your name and email and confirms with a Meet link. Signed in to Google? It fills your details in.

Open the booking page