AI strategy and systems for operations teams

We close the gaps in
how work gets done.

Your software holds the data. Your people fill the gaps: the spreadsheet beside the system, the manual cross-check, the thing everyone knows to check on Fridays.

Balmer Works helps business leaders decide where AI can reduce manual work, improve accuracy, or speed up decisions. We turn those priorities into tools your team can use in its daily work.

Let's talk See what's running
Strategy · Design · ImplementationBalmer WorksDenver, Colorado
What we do

From the right question to a working system.

You may be considering where AI belongs in your business, or already have a problem in mind. Balmer Works helps you work out whether AI is a good fit, define the outcome and a practical first step, and design and build a system around the people doing the work.

Find the right starting point
Look at how the work gets done, the available information, and the constraints. Decide where AI could help and what would make a first investment worthwhile.
Design and build the system
Build around your data, your workflows, and your exceptions. Build a working version for the team to try.
Make it work in practice
Test routine and unusual cases, then refine the system with the team. Make clear what it does, where its limits are, and when a person needs to step in.
Our work

Selected work.

We build tools for staying informed, tracking work, reviewing ideas, and handling orders. Each project starts with a specific need. We build the tool around the people using it.

Built around the way work arrives
Fabric invoice with handwritten order details and a circled request
Marked up by hand
A fabric order acknowledgement opened as a PDF, lines picked out in highlighter
Annotated PDF
An order spreadsheet color-coded by status, kept alongside the system of record
Spreadsheet tracking
Live since July 2026 Client production system

Order intake at Peter Fasano

It reads. It validates. Your team decides.
Read the case study

Orders arrive by email as typed documents, scans, and handwritten notes. The system extracts order details, checks them against the catalog, and brings discrepancies and source documents together for review. It also flags anything it could not check. Every order requires your team's approval. An ERP handoff is being developed to place approved orders in a staging queue for release in the ERP.

Checks the inbox every hour

First working version in two days

Demo order review queue showing which orders passed checks and which need attention

Demo screen using a fictional company. All names and numbers are invented.

More from the workshop
Live · Personalized daily AI briefing

Dialed In

A daily AI briefing you can configure around your own or your company’s priorities. It assesses the news in that context, explains what it could mean for you, and recommends next actions with links to the sources.

Live · Job-search management

Next

A workspace for job seekers to track their own applications and keep their search organized.

Live · Shared work and discussion

Collaborative project review

A shared review space where teams comment directly on designs, drafts, and plans, with an AI assistant available to help discuss and revise the work.

How we work

Built to be used. Clear enough to check.

Your team can check the results, correct errors, and review anything that needs judgment. We build those checks into the work from the start.

  1. 01
    Start with the gap
    We look at the manual work your existing software leaves to the team: the checks, the retyping, and the exceptions. That is where we start.
  2. 02
    Put real work in it early
    Start with one workflow and put a working version in front of the people who will use it. Use what they learn to shape the next version.
  3. 03
    Find where the tool needs improving
    Then fix the tool, never the person. A recurring workaround tells us where the design needs attention.
  4. 04
    Say what it can't do
    Show what could not be checked, what information is missing, and when the data was last updated. Make the limits visible where the work happens.
  5. 05
    Earn control, never take it
    Decide what the system should handle and where a person needs to stay involved. Build around those boundaries, and make them clear to your team.
  6. 06
    Make ownership clear
    Before work starts, we agree who owns the finished system, where it runs and processes your data, and who will maintain it.
About

A designer who builds the system, too.

I'm Greg. My background is in product design, and my method is simple: build something, put it in someone's hands, watch where people get stuck, and make it better. It has to work for the people using it, without me standing beside them.

I've applied that approach at EY for Fortune 100 clients and at Amazon Ads. At Balmer Works, I help leaders identify useful applications for AI and turn them into systems designed around how people actually work: what they need to see, what they need to question, and what they need to decide.

Greg Balmer
Greg Balmer
Founder and Principal

In 2023, I designed a real-time logistics planner at EY as part of the team preparing it for production. My recent work includes live tools for personalized AI news, application tracking, and collaborative review with AI participation, alongside a production order intake system for Peter Fasano. I'm based in Denver, and Balmer Works continues the independent practice I founded in 2015.

Greg on LinkedIn →

Start with a conversation.

Send me a note. I'll help you work out where to start.

Say hello greg@balmerworks.com