CARLILE × EOTECH

Intro Call Prep Brief · Confidential

EOTECH & Dan Agosta

Everything you need to walk into Wednesday's call sounding like you already understand their world — and to leave with one clean first win.

Wed, July 1 · 10:00 AM ET Video call (you send invite) Intro by Joe Caradonna Prospect: Dan Agosta, CIO

The 60-Second Version

What this call really is

A warm intro from a family friend to a pragmatic CIO. Not a pitch. Your job is to listen, show you respect the constraints, and find one small, measurable problem worth solving.

105

People, not thousands

EOTECH is a mid-size manufacturer, not a giant. You're not out of your league.

1

One win, not the whole company

Aim to leave with a single, low-risk first project — then expand from there.

70%

Listen more than you talk

Dan's an ops CIO. Good questions beat a slick pitch every time on call one.

The Company

EOTECH, Inc.

Plymouth, Michigan. They design and build holographic weapon sights, magnifiers, and thermal/night-vision optics for military, law enforcement, and civilian markets.

What they make
Holographic weapon sights & advanced optics. The red-reticle sights you've seen on rifles are their signature product.
Markets
Military, law enforcement, and commercial/civilian shooters.
Size
~105 employees. Mid-size, single main site in Plymouth, MI.
History
Founded 1995 (spun out of ERIM). Spent years as "EOTech, an L3 company" before going independent.
Ownership
Acquired by American Holoptics in 2020. Independent and building out its own systems since.
Their system
Just migrated to Odoo ERP (sales, inventory, manufacturing, accounting). Still customizing it and cleaning the migrated data — this is the live project.
The constraint
ITAR / export-controlled. Product technical data is legally restricted. Respect it, but it's not the whole company — most back-office/ops data isn't controlled.

Why "independent since 2020" matters: a company that recently left a big parent is usually still building out its own IT and processes. That's fertile ground for someone who can make a small, smart improvement.

The Person You're Meeting

Daniele "Dan" Agosta — CIO

Read him right and the call gets easy. He is a manufacturing-IT operations leader, not a software hobbyist. He buys outcomes, time saved, and clean metrics.

Headline
"CIO · CMMC Level 2 Certified Leader · Advanced Analytics & Data Strategy." His own bio literally says "I like data." Talk metrics and he lights up.
Career
Project Lead at General Motors; before that multiple roles at Stellantis / FCA (outbound logistics, BI, in-transit repair). Deep auto-manufacturing IT & logistics background.
Education
MBA, Walsh College (2017–2022); BA Finance, Tiffin.
Strengths he lists
IT Strategy, IT Operations, Key Metrics, Data Warehousing, Business Intelligence, CMMC 2.0 / NIST 800-171.
Based
Utica, MI. Dad of two, husband. Culture he posts about: "no red tape, just a great team doing cool shit that moves the needle."
The tell — read his open roles

They just migrated to Odoo, and he's hiring to clean it up

Dan has two roles posted: a Full-Stack Odoo Developer (custom modules, RESTful integrations, Power BI connectors, multi-company sync) and a Data Integrity Specialist whose posting says outright: "EOTECH has recently transitioned to Odoo ERP and we are refining our system configurations, permissions, and data accuracy."

Translation: fresh ERP migration, data still dirty, and he's paying a human to eyeball it. That is the most AI-shaped pain a company can have — and it's your wedge. He's proud of the CMMC/security posture, so respect the boundary before he raises it.

How to talk to him

Speak operations and ROI — not model names

Dan came up running projects in giant auto plants. He thinks in process, throughput, metrics, and dashboards. He wants someone who can take a boring, expensive task off his team's plate and prove it with a number.

Do: talk about time saved, cleaner data, live dashboards, read-first. Don't: name-drop GPT-5 vs Claude, get abstract, or oversell. Calm and practical wins him.

Your Sharpest Angle

The Odoo data-integrity layer

Don't walk in with "AI strategy." Walk in with the thing he's already spending payroll on. His open roles tell you exactly where it hurts.

The wedge

A read-only intelligence layer that sits on top of Odoo

They just migrated to Odoo and the data's still messy — duplicate contacts, inconsistent product/BOM records, missing fields, pricelist errors. Their current fix is a human intern eyeballing it. AI is built for exactly this: continuously read Odoo, flag the inconsistencies, and hand the team a prioritized cleanup list — plus natural-language reporting and auto-drafted exec summaries on top.

Why it's safe: read-only, scoped, audit-logged. You never let AI write into a defense ERP on day one. You surface issues for humans to approve. That framing is what makes a CMMC-certified CIO relax.

Your proof — say this almost verbatim

"I did nearly this exact thing for a diesel-parts company running NetSuite — same idea as Odoo, just a different ERP. I layered an operating system and live dashboards on top of it: pulled their scattered tools into one Supabase-backed brain, built warehouse and ops dashboards, and let AI reason across the data. Read-first. We replaced nothing we didn't have to. That's the pattern I'd bring here."

The ITAR boundary — respect it, don't oversell it

A short, credible nod. You're not their compliance lawyer — you're the person who obviously gets the constraint before they have to explain it.

Lane 1 — Non-controlled (start here)

Internal ops, Odoo back-office data, email/HR/IT helpdesk, SOPs, commercial/LE quoting. Not controlled tech data, so normal cloud AI is fine. Your first win lives here.

Lane 2 — Controlled (later, if ever)

Anything touching design specs / manufacturing tech data needs an on-prem / air-gapped model. Doable (~$10–15k hardware + open models), but that's a phase-two conversation, not call one.

If Dan says "we can't put our data in the cloud"

"Totally — controlled technical data stays in your environment, on-prem if it ever needs a model at all. But your Odoo back-office and commercial-side data mostly isn't controlled, and that's where the fast wins are. Cleanly separating the two is exactly the kind of thing I help with."

Where AI Creates Value

Ideas to have in your back pocket

Don't pitch these as a plan. Surface the right one based on what Dan tells you hurts. Each is ITAR-safe (Lane 1) unless noted, and each maps to something you've already built.

1

Inbound email / request triage agent

Your strongest — you've literally built this
The pain
Sales, vendor, and support email piles up. Humans sort, summarize, and route it by hand. Things slip.
The fix
An agent watches a shared inbox 24/7 — classifies, summarizes, flags urgency, drafts a first reply, routes to the right person. Never touches controlled data.
Your proof
Guardian Security already ingests email 24/7 and returns an AI verdict. Same pattern.
The payoff
Hours back per day; nothing falls through the cracks; faster commercial/LE response.
2

Quote / RFQ intake & prep

Biggest, most CIO-friendly ROI number
The pain
Quoting is slow and expert-dependent; specs get read by hand. Expertise walks out when people retire.
The fix
AI extracts requirements from incoming RFQs, pre-fills the quote, reuses proven pricing logic. Keep it to commercial/LE specs, not controlled data.
The numbers
Industry: RFQ processing ~2.5 hrs → ~25 min; one aerospace maker hit +85% quote volume at the same headcount.
Your proof
Warsaw listing tracker — monitor → triage → draft. Same shape.
3

Internal SOP / knowledge "second brain"

Speaks to his data-warehouse instincts
The pain
Staff hunt through SharePoint for SOPs, policies, IT runbooks, HR answers. Senior people get interrupted constantly.
The fix
A private assistant that answers only from EOTECH's own validated, non-controlled documents. Faster onboarding; knowledge preserved.
The proof
Joe's Kids & Freemyer — autonomous agents that monitor sources and pull from documents.
The payoff
SOP assistants (AODocs-style) cite only the latest validated version — exactly what a quality/compliance shop needs.
4

BI dashboards from manual reports

Directly in Dan's wheelhouse
The pain
Someone rebuilds the same weekly/monthly report by hand from ERP/MES exports.
The fix
Turn those manual reports into live dashboards. Low-risk, high-visibility, and it speaks his exact language (BI, data warehousing).
Context
Agentic AI is showing 30–50% cycle-time cuts vs. 10–15% from old rule-based automation. Audi stood up an internal AI assistant in ~2 weeks — these are weeks-not-years projects.

Why You're Credible

You've already built the patterns

You don't need defense experience. You need to show you build small autonomous agents that watch an input, pull from documents, and draft the next action — without touching anything sensitive.

Your one-liner

"I help businesses find where AI actually saves time, then either advise your team or build it — a second brain over your own documents, agents that handle repetitive work, or just practical tips you can run with. Everything stays in non-sensitive internal operations, which matters in an ITAR shop like yours."

Doc's Diesel · operating layer + dashboards on NetSuite ERP Guardian Security · 24/7 email verdicts Joe's Kids · grant-scout agent Freemyer · regulatory monitoring WTP Media · internal AI assistant + workflows Gulker · CRM + ops automation Warsaw · monitor → match → draft

You've delivered value both ways — as quick advisory tips (Brandon Noll, Jim Lancaster, Matt Shatto) and as full builds (WTP, Gulker). Lead with flexibility: "I can advise or build, whatever's right." That lowers the commitment and fits a first call.

The Call Playbook

Questions that make you sound sharp

Ask these, listen hard, and let the right use case surface on its own. The goal is to find the one task worth taking off their plate.

Say this first — your opener

"Dan, good to meet you. Joe's awesome — I met him this weekend and told him a little about what I do, so I appreciate the intro.

Quick background on me: I started a couple years ago helping people understand prompting and AI tools, then that turned into building websites and internal tools, then CRMs and workflow systems. Now most of my work is auditing how a business actually runs — where the data lives, how the tools connect, where the team is doing manual work, and where AI can create a real operational win.

I did a little homework and saw EOTECH recently moved onto Odoo, and that you had both a full-stack Odoo developer role and a data-integrity role. So I'm guessing a lot of the current focus is customizing Odoo, getting clean data, permissions, reporting, and integrations right. Is that about right?"

  1. I saw you posted for a Full-Stack Odoo Developer — what's the work that person is going to own? Flattering, shows homework, and his answer maps where the company's headed.
  2. How much of EOTECH actually runs through Odoo today — sales, inventory, manufacturing, accounting?
  3. You migrated fairly recently — where's Odoo working well, and where's the data still messy or people still working around it? Straight at the data-integrity pain. This is your wedge.
  4. Have you built against the Odoo API yet, or connected it to Power BI / shipping / analytics?
  5. Where does Odoo stop being enough and someone drops out to a spreadsheet to get the job done? That gap is almost always the first project.
  6. What data do people trust today, and what still needs cleanup before it's report-ready?
  7. Have you experimented with AI at all yet — what are you and the team using today? Claude, ChatGPT, Copilot, anything built into Odoo or Power BI? Learn his real appetite and maturity. What stuck, what flopped, who actually uses it vs. who's curious.
  8. And given CMMC / ITAR, what are the hard boundaries on AI touching ERP or controlled data? Shows you respect the constraint before he raises it. Don't over-lawyer it.
  9. If I could take one boring, high-volume task off your team in the next 30 days without any security risk, where would you point me first? The closer. His answer is basically your pilot scope.

Flow & Next Steps

How to run it

1

Warm open (2 min). Thank Joe, keep it human. "Mostly wanted to meet you — let's make it useful."

2

Discovery (15–20 min). Work through the questions. Listen 70% of the time. Take notes on the pain points.

3

Reflect back (3 min). "Here's the cleanest first win I'm hearing..." Name one specific, small thing.

4

Offer the next step. "Want me to put together a one-page, fixed-scope pilot on just that? Small, low-risk, measurable." Then send it within a day.

Posture reminder

You're not biting off EOTECH. You're finding one small, measurable win and pricing it so it's easy to say yes. Land it, prove the ROI, expand into a retainer. That's the whole game on call one.

What to charge — and the next move

The Odoo data-integrity build

Paid discovery / audit first: $1,500–$3,000 fixed. One-to-two weeks: they give you read-only access (or a data export), you profile the Odoo data, quantify the mess, and hand back a scored report + a build plan. This is the exact move you ran with WTP ($1,500 audit → retainer). It de-risks them and gets you paid to learn their world.

The build: read-only data-integrity layer that continuously flags duplicates, missing fields, BOM/product inconsistencies, pricelist errors — scoped as a $6k–$12k fixed pilot, or a retainer at $150–175/hr (your Midwest AI-consulting band). Frame the number against what he's already spending — a human intern eyeballing data is ~$40k/yr fully loaded, and it never scales.

The ask at the end — get the next meeting: "This was great. If it's useful, the natural next step is a short paid discovery — read-only, I dig into your actual Odoo data, and come back with exactly what's dirty, what it's costing you, and what I'd build. Want me to put a one-pager together and grab 30 minutes next week to walk you and whoever owns the data through it?"

Sources: EOTech (Wikipedia) · Dan Agosta (LinkedIn, ZoomInfo) · StartProto, Microsoft & AODocs on manufacturing AI ROI · Greypike, Concentric & OutcomeOps on ITAR / air-gapped AI. Prepared by Carlile Advisors, June 29 2026.