United Performance Metals MarginWise
01 / 08
Kickoff · AI Opportunity Assessment

United Performance Metals
× MarginWise

A six to eight week review of two parts of your business. At the end you get clear findings, a ranked set of recommendations, and working demos of what to build first.

Start  July 13, 2026 Duration  6–8 weeks Format  Focused diagnostic
02 / 08
Your team

Who you'll be working with

MarginWise is Davis and Ryan. You will work directly with both of us for the full project.

Davis Handler

Davis Handler

Founder, MarginWise

Leads the project. Former Technical Architect at Accenture and Capgemini. Davis runs the interviews, builds the business case, and presents the findings.

Ryan Esposto

Ryan Esposto

Partner, MarginWise

Leads the technical side. Ryan maps the systems behind both areas, checks the data, and builds the demos.

03 / 08
The engagement

What you get

The scope is fixed. Here is what is included and what comes after.

2
Flows mapped end to end, in depth.
6–8 wk
From kickoff to live readout. Timing depends on interview scheduling.
8–10
Recommendations, ranked by value and risk.
1–2
Working demos of what the tools could look like.

In scope

  • A close look at the two areas below.
  • Written findings you can act on.
  • 8–10 prioritized recommendations.
  • 1–2 working demos.
  • A live walkthrough of the findings with your team.

Follow-on

  • Implementation and system integration.
  • Production deployment.
  • A formal proof of concept.

We scope these after the readout, once we agree on priorities. This project tells us what is worth building first.

04 / 08
Where we are focused

The two areas we are focused on

Both areas were picked with your team before kickoff. Below is our starting picture of each. We will update it as we learn your business.

A

Planning & scheduling optimization

The larger build, with the bigger payoff.

A production scheduling tool that combines three inputs to turn orders into a schedule that works.

Inventory · what is on hand and where OEE across FIRSTCUT+® lines Labor · matching skills to the schedule
B

Sales order processing agent

The faster build, with savings that are easy to measure.

A tool that removes the re-typing from order processing. It works whether orders arrive electronically or via PDF, fax, mail, etc.

Hours saved Fewer entry errors Low disruption to the team
A The scheduling flow Illustrative
Inventoryon hand, by location
OEEFIRSTCUT+® lines
Laborpeople & skills
→
Scheduling
optimization
→
Production scheduleorders sequenced across the floor
B The order flow Illustrative
Order arrivesemail, EDI, PDF, or scan
→
Agent drafts
the entry
→
Your team reviewsreview and approve
→
Order in Invera

These are illustrative examples of the artifacts we will build as we dive deeper into your business. They will be redrawn with your team as we learn.

Order of work. We recommend starting with the order tool because it is faster to build and easy to measure. Scheduling comes second because it is larger and depends more on data quality.
05 / 08
How we work

The plan, week by week

Interviews come first and matter most. Flow mapping happens in the middle. Demos and the readout come at the end.

Wk 1
Wk 2
Wk 3
Wk 4
Wk 5
Wk 6–8
Kickoff & interviews
Interviews (plus a few targeted surveys)
Flow mapping
Scheduling + order flows
Findings & recommendations
Analysis
Demo build
1–2 demos
Live readout
Readout

Shown as six weeks. Scheduling interviews and plant availability can stretch this to eight. We would rather add time than skip important conversations. We believe Invera sits at the center of both flows. Confirming that is part of the work.

The readout includes: Interview impact summaries MarginWise commentary Proposed solutions for both flows The business case behind each What the tools could look like
06 / 08
Discovery

The interviews

Most of what we learn comes from these calls, so how they are run matters. Here is the format.

A Planning & scheduling · about 6 calls
Scheduling & planningmanagers + schedulers
Operations leadership
Shop-floor supervisors
Inventory & materials
IT · ERPInvera
IT · MES & labor systems
B Order processing · about 4 calls
Inside sales
Customer service
Sales & CS managers
IT · order systems & CRM

Around 10 calls total. The exact list gets confirmed with your team in week one.

How each call runs
Questions sent aheadat least 24 hours before
→
The call4–5 people max, managers and doers
→
Written notesshared in Teams
→
Feeds the findingsand the business case

Some calls may only include one or two people. We keep every call under an hour, and the questions arrive ahead of time so people can prepare.

Questions that build the business case
Average cost per employee.
Known problem areas and where they cost you most.
If this system were redesigned today, what would it look like?
What is still manual, and how much of it.
Do costs and procedures differ by location and country?
Where do orders arrive today, and in what formats?
What other projects, technology or otherwise, are underway in these two areas that could overlap with this work?
Where does Invera fall short today, and how much do you trust the data in it?

This is a sample of the questions we will ask. Each call gets its own full list at least 24 hours ahead of time.

07 / 08
Foundations

Is the data ready?

Most AI tools depend on clean, connected data. When the data is not ready, the usual fix is one shared data layer, sometimes called a "Company Brain." If we find that gap, we will say so and give it its own track.

Where the foundation sits Illustrative
ERPInvera
MESshop floor
Labor systems
→
Unified data layerthe "Company Brain"
→
A · Scheduling optimization
B · Order agent

This is an illustrative example of what a shared data layer could look like. Part of this project is finding out whether UPM needs one, and how much work it would take.

Where it matters most

Scheduling needs live inventory, OEE and labor data in one place. That means connecting systems that may not talk to each other today.

The order tool needs less data. Orders can arrive electronically or on paper. What matters is a consistent place where they land, so the tool can pick them up.

How we handle it

If this comes up during the project, we flag it in the readout as its own track with its own path forward.

We will not promise a scheduling tool if the data underneath is not ready. If foundation work is needed, it goes in the plan first.

08 / 08
Working together

How we stay in touch

We work inside your tools so questions get answered the same day.

T

Embedded in your Teams

A channel in your Microsoft Teams just for this project. Scheduling, quick questions, and day-to-day coordination happen where your team already works.

W

Weekly checkpoint

A 30-minute session each week covering what we found, what is next, and what we need from your side.

D

Direct line

You work directly with Davis and Ryan throughout the project.

  • We set up the Teams channel and confirm the interview list this week.
  • Interviews for both areas start in week one, with a few short surveys alongside.
  • Flow mapping starts as the first interviews finish.
  • Live readout in six to eight weeks: findings, recommendations, demos.
Davis Handler
Davis Handler
Founder, MarginWise AI
Davis Handler
davis@marginwise.ai