Independent Restaurant Operators Are About to Use AI to Perfect a Restaurant Built for the Past

Most operators will use AI to optimize a restaurant designed for a world that no longer exists. The question every operator has to answer before he adopts a single AI tool: which restaurant is the AI going to perfect — the one you have, or the one your Guests are going to want?

The Sales Pitch You’re Hearing

Every AI vendor selling into independent restaurants right now is selling the same thing. Faster scheduling. Cheaper prep. Automated inventory. Chatbots that answer reservation questions. Menu engineering that runs while you sleep. Marketing copy generated in twelve seconds instead of forty-five minutes.

The pitch is always about speed, cost, and productivity. And on those terms it usually delivers. You can genuinely cut a few hours out of your week, tighten a few line items, and clean up a few processes that were bleeding you.

The problem is not that any of that is wrong. The problem is that all of it starts with the same question, and it is the wrong question.

The question every one of those tools is answering is: how can AI help this restaurant do what it already does more efficiently?

That question sounds harmless. It is not. It is the exact question Brian Solis just spent a keynote at AI Week Milan warning enterprise leaders about, and independent operators are about to walk into the same trap at higher speed and with far less cushion to absorb it. His piece, From Automation to Imagination, is written for CIOs and Boards. The argument applies harder to you than it does to them.

What Solis Actually Said in Milan

Solis’s thesis in one sentence: the greatest risk in AI is not falling behind. It is using AI to perfect a business built for the past.

He argues that digital transformation, the last two decades of it, did not actually transform anything. It digitized. Paper forms became online forms. Meetings moved to video. Legacy apps moved to the cloud. What happened underneath the surface was almost nothing. The logic of the work stayed exactly the same. Everything got faster and cheaper and more accessible, and none of it got fundamentally different.

Solis’s claim is that AI is about to repeat that pattern at ten times the velocity unless leaders intervene. Companies will deploy AI across the enterprise, hit their ROI numbers, cut their costs, and quietly miss the future. He calls the result an AI-powered status quo. Faster work, lower costs, higher productivity, and a business model that is losing relevance while the dashboard says everything is fine.

That last sentence is the one you should read twice. A business model that is losing relevance while the dashboard says everything is fine.

If that description already sounds familiar in your restaurant, you are further into the problem than you think, and you have not even deployed AI yet.

Why This Hits Independent Operators Harder

Solis’s audience in Milan was global enterprise. Fortune 500 CIOs. Companies with balance sheets that can absorb five years of misdirected AI investment before the market punishes them. Independent operators do not have that cushion.

You have three things enterprises do not have.

First, you have less capital, which means the wrong AI investment is not a line item you write off in Q3 — it is a decision that shows up in your bank balance in six weeks.

Second, you have fewer degrees of freedom in your operating model. An enterprise can quietly restructure a business unit. You cannot quietly restructure your own restaurant. Every change you make is visible to your staff, your Guests, and your P&L within days.

Third, and this is the one nobody talks about, you have already spent a decade being sold every other operational shortcut on the market. Third-party delivery. Ghost kitchens. Loyalty apps. QR codes. Kiosks. Some of those were real tools used well. Most of them were [Hacksterism] — a shortcut posture applied to a problem the operator had not yet correctly diagnosed. AI is about to be the biggest [Hacksterism] wave the industry has ever seen, because it will feel more legitimate than any of the previous shortcuts. It has better packaging, more press, and vendors with real capital behind them.

None of that changes what it is. Applied to the wrong question, AI is the fastest, most expensive shortcut you will ever buy.

The Wrong Question Almost Every Operator Is Asking

The question every operator is being coached to ask right now goes something like this: where can I plug AI into my restaurant to save time or money?

That question is Solis’s exact warning in restaurant language. It begins with the assumptions, constraints, and operating structure of the restaurant you already have. It optimizes the process before questioning whether the process should still exist. It automates a task before considering whether the task should still be necessary. It improves a Guest journey without asking whether an entirely different Guest journey could now be designed.

This is [Hacksterism] wearing a machine-learning t-shirt. The posture has not changed. The tool got fancier.

Underneath [Hacksterism] is a deeper failure. It is the failure to have an [Operational Philosophy] before you touch the technology. If you do not have a governing idea about what your restaurant is, who the Guest is, what the Guest Experience is designed to produce, how your cast compounds, and where the economics come from, then AI does not amplify your operation. It amplifies your drift. Every AI decision you make defaults to the shape of the operation you already have. You end up with a faster, cheaper, more efficient version of a restaurant that was already losing relevance before you added AI to it.

This is [Static Decline] with a productivity dashboard on top. The restaurant looks fine on its own metrics. Prep times drop. Ticket times drop. Labor as a percentage of sales holds. And meanwhile the Guest is quietly deciding, week by week, that there is no reason to come back next month that they could not get from three other places faster and cheaper.

Speed is not transformation. Efficiency is not relevance. That is the whole argument, and it is the argument almost no operator is having with themselves right now.

The Two Curves You Have to Run at the Same Time

The single most useful frame in Solis’s piece, and the one that maps hardest onto restaurant work, is his two-curve model.

He splits AI use into two curves that must run simultaneously.

Iterative AI. This is the curve you are being sold. It supercharges what already exists. It removes drudgery, complexity, and friction. It makes existing work faster, more efficient, and cheaper. Scheduling, prep sheets, inventory reconciliation, invoice matching, review responses, basic menu copy, first-draft marketing, staff communications. Real value, real ROI, entirely legitimate.

Innovative AI. This is the curve almost nobody in restaurants is running. It uses AI to create capabilities, experiences, products, services, and business models that did not exist before. Not “reservation confirmation but faster.” A fundamentally different way for the Guest to encounter the restaurant, the Team to develop, the operator to make decisions, or the economics to work.

Solis’s warning, and the one you need to hear clearly, is that organizations that only run the iterative curve produce short-term gains that quietly consume the long-term future. Every dollar saved through automation gets absorbed by cost reduction instead of reinvested into invention. The organization becomes highly efficient at operating a model that is losing relevance. The gap between what iterative AI can produce and what innovative AI can produce is what Solis literally labels disruption in his own graph. The distance between the two curves is not a technical gap. It is a strategic one, and it widens every quarter.

This is a [By Design or By Default] fork, straight through the center of your operation. By Default, you run only the iterative curve. You save some money, you buy some time, you feel modern, and you fall further behind. By Design, you run both — you use iterative AI to free up capacity, and then you reinvest that capacity into innovative AI experiments that ask what a fundamentally different restaurant could look like given today’s tools. The reinvestment is the entire discipline. Without it, the iterative curve does not free you. It quietly finishes you.

The two curves have to spin against each other continuously. That is the [Operating Helix] applied to AI. Read your operation, design the changes on both curves, execute, read again, redesign, execute again. If you only spin one axis, it is not a Helix. It is a treadmill.

Automation, Augmentation, Agents — And Why the Order Matters

Solis lays out a three-stage progression of how AI participates in work. In restaurant terms, it is worth naming clearly, because operators are already skipping stages and paying for it.

Automation. AI takes existing work and makes it more efficient, scalable, and consistent. Scheduling, invoice reconciliation, review triage. The catch, which Solis names precisely, is that automating a fragmented process does not repair it. It can simply help dysfunction fail faster. If your inventory process is broken and you automate it, you now have a broken inventory process running at machine speed. If your onboarding is inconsistent and you automate it, you now scale inconsistency. The pre-work of automation is architectural, not technical. Fix the process by design first. Then automate.

Augmentation. AI becomes what Solis calls a cognitive exoskeleton for your Team. Not replacement. Expansion. Your kitchen manager makes faster and better decisions about prep. Your GM sees patterns in Guest feedback they would have missed. Your host understands reservation flow across a week instead of a shift. The operator sees three weeks ahead instead of three days. This is where the real leverage is for most independents, and it is the stage almost nobody is investing in because vendors do not sell it — they sell automation because it has cleaner ROI math.

Agentic AI. Systems that interpret context, coordinate across tasks, adapt to changing conditions, and act within defined parameters. In restaurant terms, this is the always-on layer that watches your operation continuously, recalibrates, and moves in event-time instead of shift-time. This layer will exist inside independent restaurants sooner than the industry expects, and the operators who arrive there without having done the automation and augmentation work first will have expensive, uncontrolled systems running on top of unresolved architecture.

The order matters. You cannot skip augmentation to get to agents. You cannot skip architecture to get to automation. Every stage requires a different governing discipline, and that discipline is not technical. It is [Discipline Architecture] — Product, People, Performance — applied to how intelligence participates in the work. If those three disciplines are not clean at the human level, the AI layer on top of them will not clean them up. It will amplify whatever is already there.

WWAID as a Pre-Prompt Discipline

The single most portable move in Solis’s piece is a small one, and it is worth adopting directly. He calls it WWAID — What Would AI Do? — and it is not what it sounds like.

It is not “use AI more.” It is a pre-prompt discipline. Before you ask AI to help with anything, you ask a different question first: if I were designing this from a blank sheet, given the tools available today, would I even structure the problem this way?

Applied to your restaurant:

Before you use AI to write faster staff schedules, WWAID asks whether the schedule needs to look anything like the schedule you have been building for fifteen years, or whether an entirely different shape of coverage would produce a better Guest Experience and better economics.

Before you use AI to answer reservation questions faster, WWAID asks whether the Guest journey to the reservation should look anything like the one you have now, or whether the entire pre-arrival experience could be redesigned.

Before you use AI to generate menu descriptions, WWAID asks whether the menu itself, as a document and as a Guest touchpoint, is even structured the way it should be structured in a world where a Guest can hold a conversation with the restaurant before they arrive.

Before you use AI to draft marketing copy, WWAID asks whether the entire premise of the restaurant’s positioning is still true, or whether you are about to write beautifully-generated copy that reinforces a market position that is quietly eroding.

WWAID is a By Design intervention wearing a machine-learning label. It forces you to interrogate the underlying architecture before you optimize the surface. That is the whole discipline.

The operator who runs WWAID before every AI decision produces a fundamentally different restaurant than the operator who runs AI over the top of their existing decisions.

Cognitive Exoskeleton, Not Labor Replacement

The default framing of AI in restaurants, industry-wide, is labor. Cut labor. Replace labor. Reduce labor cost as a percentage of sales. Every vendor pitch bends toward that framing because it is the simplest ROI math in the industry.

Solis’s better metaphor, and the one you should adopt, is cognitive exoskeleton.

AI does not replace your Team. It expands what your Team can do. Your kitchen manager, augmented by AI, sees inventory patterns and prep variance in ways he would not have caught unassisted. Your GM, augmented by AI, sees the Guest Experience across a rolling window instead of a single shift. Your host, augmented, spots reservation friction three weeks out. You, augmented, see the economics of the restaurant with a level of clarity you have never had.

None of that requires cutting a single hour of labor. In fact, done correctly, it does the opposite. It makes your Team more valuable, more effective, and harder to replace. It compounds the [Real Team Work] rather than eroding it. It builds a Cast that becomes more capable over time instead of a headcount that gets thinner.

This is a Road 2 move on AI. Road 1 uses AI to reduce cost. Road 2 uses AI to expand capacity, deepen the Guest relationship, and produce outcomes AI alone could not achieve and your Team alone could not achieve. Road 2 is where the real economic upside lives. Road 1 is where the AI-powered status quo lives.

Choose which road you are running before you buy the tool. Not after.

What This Costs You if You Get It Wrong

Solis frames the enterprise version of this failure cleanly. The organization becomes highly efficient at operating a model that is losing relevance.

The restaurant version is worse.

You spend six to eighteen months rolling out iterative AI tools that legitimately produce savings. Your dashboards look good. Your margins tighten. You feel modern. You tell yourself you are ahead of the curve. Meanwhile, the actual Guest — the one you built the restaurant for a decade ago — is quietly deciding that your restaurant does not do anything for them that three other places do not do faster, cheaper, or more interestingly. Your traffic softens. You blame the economy. You blame the neighborhood. You blame delivery apps. You cut a little more cost, add another AI tool, tighten another line item.

That is [Static Decline] on the AI-powered status quo curve. It is invisible from inside the restaurant until it is not. And by the time it is visible on the P&L, you have burned the capacity you would have needed to run the innovative curve, because you spent it all on efficiency gains that never got reinvested.

The independent operators who come out of this decade with real businesses will be the ones who ran both curves. Iterative AI to free up capacity. Innovative AI to invent the restaurant they need to become. And the discipline to reinvest every hour saved on the first curve into an experiment on the second.

Nobody is going to sell you that discipline. It is not a product. It is a leadership posture. Which is exactly what Solis said in Milan, and exactly what the restaurant industry is least prepared to hear.

The Question That Belongs on Your Whiteboard

Solis closes his piece with a single question he wants every leadership team to bring to the table. Translated for the independent operator, it belongs on your whiteboard until you have an answer:

Are you using AI to improve the restaurant you were, or to invent the restaurant you need to become?

If you cannot answer that question in a sentence, you are not ready to sign the vendor contract in front of you. If you can answer it, and the answer is honest, you now have a filter for every AI decision you make from this point forward.

That is the whole discipline. Everything else is [Hacksterism] with a better logo.