Why We Built RAC Projects AI as an AIOS Firm

We built RAC Projects AI as the firm that installs AI Operating Systems for mid-market operators. Here's the gap we saw and the posture we took.

This post is about the firm itself, not a specific methodology piece. If you want the methodology side, our posts on how we work and what an AIOS Blueprint measures cover that. What we want to do here is explain why this firm exists at all, in the shape it exists, and what we are and are not trying to prove with it.

The gap in the mid-market lane

When a CEO of a $5M or $20M or $40M company decides they need to do something serious with AI, they usually get two options from the market.

Option one is strategy consulting. Partners in nice shoes, a discovery phase, a workshop or two, and eventually a deck. The deck is often good. It names the right themes, points at the right risks, maybe even scores your readiness on a maturity model. Then the engagement ends and the deck sits in a shared drive. Nothing installs. Nothing compounds. The client is more informed and exactly as operationally constrained as before.

Option two is an automation agency. Smart builders, real tools, often good pricing. They ship a narrow win. An email workflow. A scraper. A chatbot. A lead-scoring script. It works on day one and quietly decays by month four because nobody owns it, it isn't connected to the rest of the business, and no one ever said what "done" looked like. We've written about this failure mode in why AI pilots die in month 4.

Neither of those is an operating layer. A deck doesn't run. A script doesn't think. What mid-market operators actually need is a coherent system that teaches itself the business, centralizes the data, synthesizes signal into a leadership view, automates repeatable work behind approval gates, and gets custom-built modules only where the off-the-shelf ones don't fit. That's the five-layer AIOS, and no one was installing it as a product for companies our size.

So we built the firm that does.

The posture we took

We made four deliberate choices about what kind of firm we wanted to be, and every one of them is a reaction to something we thought the market was getting wrong.

Advisor-first, builder-second. Every install starts with a diagnosis. A Fit Check to decide whether the business is even ready, then a Blueprint that produces the actual install spec. Strategy drives the technical decisions, not the other way around. The shape of the automation follows the shape of the business, not the shape of whatever tool is on sale that quarter. Ed runs this side of every engagement.

Named partners, not a freelancer marketplace. When a client signs with us, they know exactly who runs their engagement. Ed leads advisory. Jeremy leads technical delivery. There is no account-manager layer between you and the people doing the work. There is no junior we hand the real execution off to once the contract is signed. This is a deliberate choice about accountability. Harvard Business Review's coverage of leadership and accountability in professional services keeps coming back to the same finding: the firms that compound client trust are the ones where the partner who sold the engagement is the partner who delivers it. We agreed, and we structured accordingly.

Public method. Every phase of our engagement is documented publicly on this site before you ever talk to us. Fit Check. Blueprint. Build. Run. What each one produces, what the acceptance bar looks like, how the fee structure works, when and how engagements end. No black-box pitch deck. No "we'll walk you through our proprietary framework" on a sales call. You can read the method and decide whether we're the firm to run it for you, before we ever get on a call.

Installable, not rentable. Engagements end. They don't become indefinite retainers. The Run phase has a defined arc, and it ends in graduation, not churn. The client owns the AIOS when we're done. We're not trying to become the permanent external brain for your company. We're trying to install an operating layer your team runs.

Taken together, that's a different posture than either strategy consulting or an automation agency. We've written about the specific differences in AIOS partner vs consultant vs agency, but the short version is: strategy consulting stops at the deck, agencies stop at the script, and we stop at a working layer your team can run without us.

Why mid-market specifically

We don't take every engagement that shows up. We focus on companies in the $1M to $50M revenue band, and that's also a deliberate choice.

Below $1M, AIOS is overkill. A company that size probably has one or two good automations worth installing and a founder who can still hold the whole business in their head. They don't need an operating layer yet. They need focus, a few tool decisions, and some breathing room.

Above $50M, enterprise consulting with its governance overhead, procurement cycles, and multi-vendor change management is probably the right shape. Not cheaper, not faster, but structurally correct for the size. A firm of our size can't credibly run a 200-stakeholder install, and we're not going to pretend we can.

The $1M to $50M band is where AIOS shows up as an asymmetric bet. The company is big enough to have real operational complexity. Multiple departments. Real data across multiple systems. Decisions that are happening in too many places. A CEO who can feel themselves becoming the bottleneck. And small enough to install a coherent operating layer in three to six months rather than three to six years. That's the band where an AIOS install compounds the hardest, and it's the band where we think we can be uniquely useful.

We've written about the timing shape of this bet in the 24-month window for AIOS installs and about what it looks like inside a specific size of company in what AI-first means in a 50-person company.

Ed's read and Jeremy's read on why we built this

We share the frustration that led to this firm, but we came at it from different angles.

Ed's read: I spent years watching AI consulting calcify into PowerPoint. Smart partners, good frameworks, real diagnostic insight, and then at the end of the engagement the client was handed a deck and a set of recommendations and was expected to figure out the rest. The gap between "here's what you should probably do" and "here's a thing that runs in your business" is enormous, and it's the gap where most of the value leaks out. I got tired of leaving that gap for someone else to close. I wanted to run engagements that end with something installed, not something concluded. The advisory work I do now, Fit Check and Blueprint, is still strategy. But it's strategy that lives upstream of a real install, not strategy that lives upstream of an unopened PDF.

Jeremy's read: I spent a decade running operations inside a mortgage origination business. Small company, real revenue, real operational mess. I saw exactly where the pain was: work that lived in too many spreadsheets, handoffs that happened over Slack, a leadership team trying to make decisions from numbers that got argued over more than they got trusted. When I started watching what the automation-agency world was shipping to companies that size, it was almost all narrow wins. A cool script that moved one piece of the problem and didn't touch the other nine. I got tired of watching $20K automation projects decay into shelfware by month four. Build is where the real work is. Not building one clever thing, but installing a coherent stack that a team can actually run. That's what I want to do for a living.

Both reads point at the same thing. The market was producing good individual pieces and no integrated layer. So we built the firm that installs the layer.

What Audity is, and what it isn't

We should be clear about something because it comes up in conversations with prospective clients.

Audity is a separate product. It's a SaaS platform for running AI transformation audits, and it has its own team, its own customers, and its own business model. It is not part of RAC Projects AI. It's not the tool we use inside our engagements. It is not an upsell inside a Blueprint.

RAC Projects AI is a services firm. We install AIOS. We charge for Blueprints and Build engagements and Run retainers. The deliverable is an operating layer inside your company, not a login to a platform.

We mention Audity here only because people reasonably ask. The honest answer is: Audity is a separate venture with separate economics, and what you're reading about on this site is the firm, not the product.

Where we are now and what we're trying to prove

RAC Projects AI is a newer firm building on a newer model. We want to be honest about that.

The AIOS framework is new enough that the five-to-ten-year case studies don't exist yet. The posture we're taking, advisor-first, named partners, public method, installable, is one we believe is right, but the proof is in whether engagements in year two and year three and year five keep producing operating layers that compound rather than decay. Harvard Business Review's ongoing work on strategy execution is clear that the firms that hold up long-term are the ones whose stated method survives contact with real client outcomes. We're early enough that the full story isn't yet written.

What we can say with confidence right now is the shape of the gap and the shape of the posture. The mid-market gap between strategy consulting and automation agencies is real, and it's wide, and we think an AIOS Partner is the right shape to fill it. The next three to five years of engagement outcomes will tell the full story, and we intend to keep writing honestly about what we see as it unfolds. If you want the firm-identity context, our about page covers the same ground from a different angle.

If you're a mid-market operator wondering whether any of this applies to your company, the lowest-friction place to start is a Fit Check. Five minutes, self-serve, no call required. If you come out the other side wanting to talk through what an engagement would look like, we'll take it from there.

-Ed and Jeremy