Founded 2016 - Chicago, Illinois $16M raised - $10M Series B led by Baird Capital ModelOp Center: the control tower for enterprise AI 50+ out-of-the-box integrations Customers include Fidelity, FINRA & Bristol Myers Squibb Governs ML, generative AI, agentic AI & 3rd-party models ~2x faster AI deployment reported Founded 2016 - Chicago, Illinois $16M raised - $10M Series B led by Baird Capital ModelOp Center: the control tower for enterprise AI 50+ out-of-the-box integrations Customers include Fidelity, FINRA & Bristol Myers Squibb Governs ML, generative AI, agentic AI & 3rd-party models ~2x faster AI deployment reported
ModelOp logo
The wordmark of a company that doesn't build a single model - it keeps tabs on everyone else's.
Company Profile - AI Governance

ModelOp

The Chicago software company that asks every enterprise an uncomfortable question: can you name every AI model running in your business right now?

HQ Chicago, IL Founded 2016 Stage Series B Team ~44 Category AI Governance
Share X / Twitter LinkedIn Facebook Instagram
The control tower

Somewhere in a Fortune 500 bank, a model just made a decision

It approved a loan, flagged a transaction, or drafted a customer email. Multiply that by a few thousand models scattered across data-science teams, cloud accounts, and third-party tools, and you have the modern enterprise: enormously productive, and not entirely sure what it has running. ModelOp exists for the executive who has to answer for all of it.

The company sells software called ModelOp Center. Think of it less as another place to build AI and more as the room where someone finally turns on the lights. Every model - traditional machine learning, generative AI, the newer agentic systems, and anything bought from a vendor - gets registered, risk-tiered, tested, approved, and watched. One inventory. One paper trail. One answer when the regulator calls.

"We don't build the models. We govern the thousands of models everyone else builds." - The ModelOp pitch, paraphrased

It is not the most glamorous corner of artificial intelligence. There is no chatbot to demo, no image to generate. ModelOp's product is closer to air-traffic control: invisible when it works, very loud when it doesn't. The bet is that, eventually, every serious enterprise needs a tower.

The problem they saw

AI was easy to start and nearly impossible to account for

By the late 2010s, building a model had become a weekend project. Governing one had not. Data scientists could spin up a predictive model in an afternoon; nobody could say, six months later, whether it had drifted, who approved it, or what would happen if an auditor asked to see the math. The gap between "we have AI" and "we know what our AI is doing" widened with every successful experiment.

The uncomfortable truth

In most large enterprises, the people responsible for AI risk cannot produce a complete list of the AI in production. The models outnumber the governance.

This is the tension ModelOp has lived inside since day one. Speed and control usually trade against each other - the faster you ship, the less you document; the more you document, the slower you ship. Regulated industries felt it worst. A bank cannot deploy a credit model on vibes. A hospital cannot let an algorithm influence care without a record. For them, ungoverned AI isn't a productivity win. It's a liability waiting for a date.

"The models outnumber the governance. ModelOp's whole job is to flip that ratio." - Why the category exists
The founders' bet

Three builders who thought governance was the real product

ModelOp was founded in 2016 by Pete Foley, Stuart Bailey, and Robert Grossman - a combination of enterprise-software operator and deep data-science research. Foley had already run and exited several enterprise companies before this one, including a stint as CEO of Infoblox and leadership of two firms acquired by Citrix and Websense. Bailey and Grossman brought the open-science and machine-learning heritage. Between them, they made an unfashionable wager.

The fashionable thing in 2016 was to build models. The ModelOp founders bet that the durable business was in operationalizing and governing them - a discipline they helped name "ModelOps." It sounded like plumbing. It was plumbing. But plumbing, as anyone who has had a leak knows, becomes the most important thing in the house at exactly the wrong moment.

The wager

Everyone will build AI. Far fewer will be able to prove what it's doing. Sell the proof, not the model.

It took years for the rest of the world to agree. The arrival of generative AI - and the regulation chasing it - turned a niche concern into a board-level one. The founders had been early. Being early, of course, is just another word for being wrong for a while.

The ModelOp timeline

A decade of betting on the boring, necessary part
2016
Founded in Chicago by Pete Foley, Stuart Bailey, and Robert Grossman, around the then-novel idea of "ModelOps."
2020
Early institutional funding begins as enterprises start formalizing model risk and operations.
2023
Generative AI breaks loose. Governance jumps from niche to mandatory; ModelOp extends Center to LLMs.
2024
$10M Series B led by Baird Capital; joins the AWS Global Startup Program and lists on AWS Marketplace.
2025
Agentic AI governance capabilities added as autonomous agents enter the enterprise.
2026
Leadership handoff (reported): Dave Trier, formerly SVP of Product, named CEO as co-founder Pete Foley steps back.
The product

One system of record that sits above everything else

ModelOp Center does not try to replace the tools companies already use. It sits on top of them. Below it run the MLOps platforms, the cloud accounts, the ticketing systems, the GRC software. ModelOp orchestrates across all of them - more than fifty integrations out of the box - so governance happens once, centrally, instead of being reinvented by every team.

AI System of Record

A single inventory of every AI use case and model, no matter where it was built or what it runs on.

Automation & Orchestration

Standardized intake, auto-generated risk tiers, automated bias/drift/performance tests, and enforced approval workflows.

Continuous Controls

Live monitoring of production AI from one dashboard, with cost, risk, and ROI tracking baked in.

Audit-Ready by Default

Auto-generated model cards, documentation, and compliance reports - the paperwork writes itself.

"It's air-traffic control for AI: invisible when it works, very loud when it doesn't." - On what ModelOp Center actually feels like

The irony ModelOp has had to sell against is that governance has a reputation for slowing things down. Their counterargument is data: when approvals, testing, and documentation stop being manual, the bottleneck disappears. Customers report roughly twice the deployment speed. Control, it turns out, can be the thing that lets you go faster - which is a harder sentence to put on a billboard, but a truer one.

The proof

Banks, regulators, and a drugmaker walk into a platform

The customer list is the argument. Publicly referenced names include Fidelity Investments, FINRA, and Bristol Myers Squibb - a money manager, a financial regulator, and a pharmaceutical company. What they share is a low tolerance for unexplained algorithms and a legal obligation to prove their work. These are the hardest customers to win and the least likely to leave.

$16M
Total raised
50+
Integrations
~2x
Faster deploy
2016
Founded

Funding to date

Cumulative capital raised (USD, approximate)
Early rounds
~$6M
2024 Series B
$10M
Total
$16M
Modest by AI-hype standards - which is rather the point. The 2024 round was led by Baird Capital, a B2B-software specialist that liked the blue-chip customer base more than the buzzwords.

There are partnerships to match. ModelOp joined the AWS Global Startup Program and put Center on the AWS Marketplace, smoothing procurement for enterprises that already buy through Amazon. The platform plugs into Databricks, Snowflake, MLflow, ServiceNow, Jira, OpenAI, PowerBI, and Tableau, among others - the unglamorous connective tissue that makes a governance layer actually usable.

"We've been extremely impressed by ModelOp's robust software capabilities, strong leadership and blue chip customer base." - Jim Pavlik, Partner, Baird Capital
The mission

Responsible AI, but with a ledger

ModelOp frames its mission around a single phrase: safeguarding every AI initiative while enabling responsible innovation at scale. Stripped of the slide-deck gloss, it means letting risk-averse organizations say yes to AI without losing the thread of what they've deployed. "Responsible AI" is a popular term; ModelOp's contribution is to make it auditable rather than aspirational.

ai governance modelops model monitoring agentic ai ai compliance responsible ai model inventory audit trail

The mission has a convenient tailwind: regulation. As governments draft rules for AI, the documentation and controls ModelOp sells stop being best practice and start being required. A company built on making AI accountable is, almost by accident, perfectly positioned for a world that's about to demand accountability in writing.

Why it matters tomorrow

The agents are coming, and someone has to watch them

The next wave is agentic AI - systems that don't just predict but act, chaining steps and making calls with less human in the loop. That is exactly the moment a control tower earns its keep. ModelOp has already extended Center to govern these agents, betting that autonomy without oversight is a headline nobody wants to be in.

The next decade

More AI, more autonomy, more rules. Each trend pushes in ModelOp's direction. The harder AI is to trust, the more valuable the company that documents it.

Go back to that bank, and that model making a decision. A year ago, no one could fully account for it. Now it sits in an inventory, risk-tiered the day it was registered, tested on a schedule, approved by name, monitored for drift, and documented automatically. The decision still happens in a fraction of a second. The difference is that someone, somewhere, can now explain exactly why - and prove it. That, quietly, is the whole company.

"The lights are on. The tower is staffed. The models, for once, are being watched." - ModelOp, 2026