As 2025 wraps up, I’m writing this from a place of deep gratitude. Last year, Daniel, Jason and I had an idea on a whiteboard. Today, we’re processing real revenue, automating real workflows, and giving real people their time back.

I am grateful to my co-founders for going on this journey with me, to my investors for trusting us to shepherd their investments, and our customers for trusting us with their business. Those who know me know that service is my number one priority, and I couldn’t have asked for a better group of partners on this journey.

Buckle up, it’s a long one!

TL;DR:

1. First network deployment exceeding expectations, 100+ sales reps, 16 sales agencies & $570k+ in revenue flowing through Deviceflow in just 3 weeks
2. AI Purchase Order automation in production — handling bill-only and stocking POs daily
3. 171 pull requests shipped in December despite the holidays
4. New website launched focusing on our automation capabilities
5. 2025 Year in Review — from zero to deployed across a manufacturer’s entire distribution network
6. Looking ahead to 2026 - let’s grow this thing!

How You Can Help

Same asks as last month, now with more urgency as we gear up for 2026:

1. Spine Manufacturer Intros — The billing complexity in spine is exactly what our PO automation was built for. If you know manufacturers struggling with price lists & manual PO processing, we’d love to connect.
2. International MedTech Companies — Still looking for European/Asian companies planning US market entry. Our quick deployment model is perfect for their tight FDA approval windows.
3. Spread the Word — Our new website is live! If you know anyone in medical device operations, point them to ​deviceflow.com​.

First Network Deployment — The Numbers Speak

Remember last month when we went live with our first manufacturer distribution network? In December we onboarded more than 100 sales reps across nearly every state in the US, representing 16 distributorships. The early results are in, and they’re exceeding expectations.

$570k+ in revenue flowed through Deviceflow in less than 3 weeks.

That’s real purchase orders, real inventory movements, real billing — all orchestrated through our platform. Much of that revenue was facilitated by our purchase order automation features, which brings me to…

AI Purchase Order Automation — In Production, Working Daily

Here’s the workflow we’re replacing:

A manufacturer has a central email address where purchase orders land. Staff sits on that inbox, processing each PO manually. When a bill-only doesn’t have matching utilization data, they email the responsible rep to enter it — or do it themselves. At 30+ procedures per day, this becomes untenable for a single person. So headcount grows linearly with procedure volume. That’s a massive drag on margins.

And POs are just one of dozens of email-initiated workflows their staff handles daily.

Enter Deviceflow.

Our AI agents sit inside their email inbox and triage these tasks automatically. Here’s what happens when a PO arrives:

1. Deviceflow reads the incoming email, recognizes an attachment
2. Parses the attachment to understand the data inside
3. Classifies the PO type (bill-only for a procedure? stocking order for inventory replenishment?)
4. Orchestrates the appropriate business logic for that PO type
5. Transforms the PO into structured data
6. Routes it through our system to the ERP, QuickBooks, wherever it needs to go

Our PO matching logic and data extraction has been trained on real-world hospital and ASC purchase orders. We’re improving it daily based on production feedback, making this functionality increasingly bulletproof.

With Deviceflow, staff focuses on exceptions and relationships, not data entry. Procedure volume can grow without proportional headcount growth. Margins improve.

For Deviceflow, these email-initiated workflows are our product roadmap.

Development Velocity — 171 PRs in December

Even with the holidays, December was our second-highest productivity month ever. 171 pull requests delivered major advances in:

• Bill-only automation — transformed from manual workflow to intelligent automation
• Enterprise analytics — giving operations teams visibility they’ve never had
• Production reliability — the observability infrastructure needed for enterprise-scale operations

In December we established the foundation for the scale we’re building toward.

New Website Launch

Late December we launched our new website, refocusing our messaging on what’s resonating most: automation capabilities, starting with purchase order automation.

The site tells the story of where we’re headed — not just another inventory system, but an AI-native operations platform that meets people where they work.

2025 Year in Review

One year ago, Deviceflow didn’t exist. Looking back at what we’ve built, I’m proud of how much ground our team covered this year

Q1 2025: Foundation

• Interviewed 100s of folks to understand the real-world problems they’re experiencing in Medical Devices
• Signed first design partners
• Worked on our first audit & visual machine learning tools

Q2 2025: Learning

• Deployed our solution to a large distributorship
• Processed 20,000+ shipments
• Tracked 100,000+ devices
• Proved the “text your AI ops assistant” concept works

Q3 2025: Validation

• Built core platform architecture
• Selected for Plug and Play Health Batch 21
• Completed first major audit — $1.5M in inventory cost recovered, 80% reduction in variance
• Participated in Teamworthy Ventures Fellowship

Q4 2025: Acceleration

• Closed our $1.5M pre-seed
• First manufacturer network rollout — 100 sellers across 16 agencies
• AI PO automation in production
• $570k+ revenue through platform in December alone
• 142 PRs in November, 171 in December

By the numbers:

• $1.5M+ in inventory cost recovered for customers
• $570k+ in revenue processed through platform (December alone)
• 500+ pull requests shipped in Q4
• 80% variance reduction in audits
• 3 weeks from deployment to live revenue processing

What we learned:

1. Meet people where they work. Every successful feature we’ve built follows the inbox pattern — sitting inside existing workflows rather than forcing new ones.
2. Pattern-driven development compounds. Our “enterprise Legos” approach means each new capability builds faster than the last.
3. The network effect is real. Distributors want their manufacturers on Deviceflow. Manufacturers want their distributors on it. The platform becomes more valuable with each participant.
4. AI-native beats AI-bolted. Building with AI from day one lets us do things competitors can’t retrofit, and our approach is completely different than making reps use yet another app to do their work.

Looking Ahead to 2026

2026 is about scaling what works:

• Expand network deployments — more manufacturers, more sales agencies
• Deepen automation — PO automation is just the beginning; case scheduling, rep onboarding, and more are in the pipeline
• Commercial launch — moving from design partners to general availability
• Team growth — still hiring founding engineers to accelerate everything

The medical device industry is ready for this change. 2025 proved the model works. 2026 is when we scale it.

Final Thought

A year ago, my pitch to potential investors was that we were going to build an AI-native operations platform that would transform how medical devices move through the world. That we’d replace spreadsheets and text messages with intelligent automation. That we’d create a network where every participant makes everyone else’s job easier.

Twelve months later, we have revenue flowing through the platform, AI agents processing purchase orders, and a manufacturer’s entire distribution network on our system.

We’re just getting started.

Thank you for being part of this journey. Your introductions, advice, and belief in what we’re building have made all the difference. Here’s to an incredible 2026.

Happy New Year!

Best,

Anatoly, Dan & Jason Team Deviceflow