AI automation for manufacturers and freight forwarders · Built & Measured

Make it, ship it, stop typing it twice.

Factories and forwarders lose the same hours: orders, purchase requests and bookings copied out of email, keyed into systems, then chased. I find the few places AI genuinely pays off, build them with your team, and hand them over working. If nothing is worth building yet, I will tell you straight.

At 21:40 a buyer changes an order. At 21:42 a shipper asks for a rate on 2 x 40HC to Long Beach.

Your ERP and your TMS will run both beautifully. But only after somebody reads the email, works out who it affects, and types it all in.

So purchasing hears on Thursday, and the quote goes out Thursday. The shipper booked Tuesday, with whoever answered first.

The way out

Most AI never gets used.
Yours becomes your Ergo System.

Ergo, from the Greek for work: the system your work runs on. I find what is worth building, in the factory office or on the quote desk, build it, and make it part of how the work runs, end to end.

See how we do it

Where AI is worth it

I only build where it moves a number.

Hours back, cost per order, win rate. I start where the impact is biggest for the least effort, fix it part by part, then wire the wins into one system, and it all lands in the same place: your profit.

Win more of what you quote

Rate requests and buyer enquiries captured the moment they land, quotes drafted in your voice for a person to check and send, and follow-up until each one books or dies honestly.

Cut the re-typing

Purchase requests, order changes, pre-alerts and invoices read by the system and checked by a person. The same team moves more work.

See where work stalls

Orders, purchases and quotes tracked in plain language: what is stuck, what changed this week, which customers and lanes pay. You act on today, not last month.

It all lands in your numbers. Biggest impact, least effort. Part by part, then as one system.

Your system gets built in three moves…

  1. Identify

    I start with how the work actually runs, in the office or on the quote desk. What gets re-typed, where orders and quotes stall, what nobody can see. From there I pick what gets built first: biggest impact, least effort.

  2. Build

    The plan becomes working systems. Requests structured, approvals on screen, quotes drafted and chased, documents pre-filled, built to fit the ERP, TMS and inbox you already run, so it works from day one.

  3. Integrate

    Then I make it stick. I work side by side with your team until it is not a project anymore, it is just how work gets done. You own it, documented and handed over.

This is for you if

You run on orders, requests and follow-ups, and too much slips.

In the factory office and on the quote desk, the systems are tooled and the work between them runs out of the inbox. Changes reach people late, requests get re-keyed, and follow-up depends on memory. Pick a case and watch the mess go in, the system do its moves, and the outcome land.

The mess coming in
    The system
      The outcome

      See the full walkthrough →

      purchase requests to POsrate-request captureplanning packsquote follow-uporder-change alertspre-alert processingsupplier confirmationsdocument pre-fillhuman-sounding repliesstatus updatesemail-to-ERP entrycarrier-invoice checksask-your-data analysis

      Real work

      Built, shipped, and in use.

      No borrowed logos. Just work I can stand behind.

      Recruitment · Training

      Case study

      Candidate matching, with the reasoning shown

      Pulse scores every application for a lean training and recruitment startup against the role, the company, and the hiring manager's notes, checked across 1 000+ companies. Recruiters see the AI score next to their own read; a person makes the call.

      Travel · Operations

      Case study

      Messy enquiry email, turned into clean records

      For a travel and car-rental group I turned unstructured enquiry email into structured, usable records: who, what, dates and budget extracted, tracked and ready for fast follow-up. The same intake pattern a forwarder's rate desk or a factory's purchasing inbox needs.

      Fair questions

      Asked before, answered honestly.

      We already run an ERP, a TMS or both. Does this replace them?

      No. Your systems keep doing what they do. I build the layer that runs before and around them: requests captured, approvals on screen, quotes drafted and chased, updates in your voice, wired to the data you already hold.

      We've tried AI before and it didn't stick. Why would this be different?

      Because tools alone never stick. I start with how your team works, build only what earns its place, and stay until it is part of everyday work. Integration is the job, not the afterthought.

      Do you work outside Vietnam?

      Yes. I am based in Ho Chi Minh City, so I can work on-site in Vietnam, and remotely with businesses in the US and elsewhere.

      Our data is a mess. Is that a problem?

      It is normal, and it is usually where I start. Turning a shared inbox, emailed rate sheets and spreadsheets into structured, usable data is the foundation move, and the thing I do most.

      How much of our team’s time will this take?

      A few focused hours up front to understand how you work, then short check-ins while I build. I work around your team, not the other way round.

      Do we end up locked into you?

      No. You own everything I build. It is documented, your team is trained on it, and ongoing help is optional. If you never need me again, that is a good outcome.

      What does it cost?

      It depends on what is worth building, and sometimes the honest answer is that nothing is, yet. The first conversation is free and I will tell you straight.

      The next order change or rate request is already in the inbox.
      Answer it with a system.

      Tell me where the work leaks: purchase requests re-keyed, order changes that reach people late, quotes unchased. I will tell you honestly whether AI helps, and what I would build first.

      No hype. No stats I can't back. If you don't need it, I'll say so.

      Start here

      The first conversation is free.

      Tell me where the work leaks, in the factory office or on the quote desk. I will tell you honestly whether AI moves a number there, and what I would build first. If nothing is worth building yet, I will say so.