Bro, Do You Even Claude?

I joked with a friend yesterday that I need to get a T-shirt made that says, "Bro, do you even Claude?"

The original meme, "Bro, do you even lift?", has already been recycled for everything, so why not this? But applying it to Claude, or to LLMs more broadly, seems to poke fun at something way more serious, and that is how AI is quickly becoming a dividing line in how people think about building software.

For the purposes of this article, I want to focus less on the broad societal implications and more on what it feels like AI is doing to software in the restaurant industry, as someone who talks to a lot of restaurant tech companies, startups, and technology leaders at restaurant brands.

I'd put it this way:

AI has lowered the perceived cost of building software faster than it has lowered the real burden of owning software.

AI can help teams write code, prototype products, document requirements, test ideas, and move faster. Google's 2025 DORA research found that more than 80% of software professionals surveyed said AI enhanced their productivity, while McKinsey found that top-performing AI-enabled software organizations saw meaningful improvements in productivity, customer experience, time to market, and software quality. But both reports also make clear that AI adoption alone is not enough; teams still need the right workflows, operating models, governance, and engineering discipline to turn speed into reliable outcomes.

That matches what I'm seeing in restaurant tech right now. AI has made the first 70% of software feel a lot more accessible. The hard question is whether brands, startups, and established vendors are underestimating the last 30%: integrations, reliability, edge cases, data quality, support, security, governance, change management, and long-term ownership.

Among larger multi-unit restaurant brands, I'm hearing more leaders say some version of: "Yeah, we can build that."

This is not a blanket statement about whether building is smart or not. Even before the AI wave, restaurants had plenty of build-versus-buy debates. Some brands have always believed that certain technology should be proprietary. Others have preferred to stay focused on operations and buy software from vendors. What feels different now is the level of conviction.

The public conversation does not yet adequately reflect how much some brands have shown interest in developing technology they previously would have bought off the shelf. And I would include in this category not only brands building fully in-house, but also brands using agencies, consultants, or development partners to build and maintain custom software.

A lot of what gets described as "building internally" is really "custom-building with an outside partner." That can solve some vendor-fit problems, but it creates a different set of ownership and maintenance questions. You may reduce dependency on a product vendor, but you still have dependency risk: documentation risk, support risk, key-person risk, roadmap risk, architecture risk, and the question of who owns the product when the original builders move on.

Over the past month, I've talked with three or four technology leaders at brands with more than 200 locations that are developing their own decision intelligence solutions. By decision intelligence, I mean tools that help operators make better decisions around areas like forecasting, labor, inventory, pricing, promotions, site selection, purchasing, benchmarking, and operational recommendations.

The interest in those use cases is not surprising. Deloitte's 2025 study of 375 restaurant executives across 11 countries found that more than half of restaurants already use AI to forecast product needs or track inventory. But interest is not the same as readiness. The same study found that many operators feel they still have work to do before leaning on AI more heavily, citing concerns about risk and governance, gaps in talent, and tech infrastructure that is not yet where it needs to be.

That combination explains a lot of what I'm hearing. The demand for better decision-making tools is real, the dissatisfaction with generic software is real, and AI makes the lift feel more manageable.

The brands I spoke with were not casually brainstorming, and they were not on the fence in the way I expected. They were talking with confidence. They believed they could get closer to exactly what they wanted, that the build was manageable, and that they could avoid the risk of having a core operating workflow subject to the unknowns of working with a vendor.

There are also high-profile examples that make this mindset feel more plausible. Yum! Brands introduced Byte by Yum!, a proprietary AI-driven SaaS platform across KFC, Taco Bell, Pizza Hut, and Habit Burger & Grill, covering areas like online ordering, POS, kitchen and delivery optimization, menu management, inventory, labor, and team member tools. Yum said 25,000 restaurants globally were using at least one Byte by Yum! product at the time of launch. McDonald's has also been explicit about using Google Cloud's hardware, data, edge computing, and AI technologies across restaurants, including applying generative AI to business priorities for crew and customers.

Of course, Yum and McDonald's are not typical restaurant brands. Their scale gives them options most brands do not have. But their moves still matter because they change the imagination of the market. They make the question "Could we build this?" feel more normal.

The future is probably not build or buy. It is more likely build, buy, partner, and agency-build, often inside the same brand. A restaurant company might buy the POS, build a decision layer, use a partner for integrations, license models from OpenAI, Anthropic, Google, or others, and still depend on incumbent vendors for core workflows.

That is a much more complicated world than the old build-versus-buy debate.

Startups continue to spring up everywhere.

AI has enabled builders to build faster, cheaper, and with less experience than before. That is exciting, and it also makes the market exhausting to sift through. The barrier to creating a demo has collapsed; the barrier to creating a durable company has not.

This is where I'd qualify the claim that there are "more startups than ever." The funding picture is more nuanced. WIPO reported that AI captured 53% of global venture capital deal value in Q3 2025, but also that overall deal numbers fell year over year and that a handful of large AI megadeals were driving a lot of the headline growth. So the better way to say it may be: the market feels noisier than ever, even if capital is increasingly concentrated around a smaller number of perceived winners.

That noise creates a problem for restaurant brands. There are more products, more pitches, more AI wrappers, more demos, and more founders saying they can transform operations. Some of them are right. Most will not make it.

As a result, I've found myself using quick filters to decide what is worth paying attention to: domain expertise, founder-market fit, customer love, technical velocity, early traction, and a credible answer to enterprise durability risk.

I used to think mostly in terms of product quality: Does this solve a real problem? Is the workflow good? Does the team understand restaurants? Those questions are still important, but the bigger question I now find myself asking — and the question I think brands are asking, too — is whether, by circumstance or by choice, a vendor might pull the rug out from under its customers.

That question deserves to be central in restaurant tech.

For a brand, vendor risk is not abstract. Will the product still exist in two years, and will anyone still support it if it does? Will the company get acquired, change priorities, or raise prices overnight? If integrations break, will the data even be portable? Will investors hijack the roadmap or chase a bigger market that leaves restaurants behind, and will the underlying AI model provider change its terms, pricing, or performance? Underneath all of it sits the most basic question of all: will the company survive?

This is part of why more brands are considering building their own solutions. It is also why many brands continue to choose trusted established companies even when those companies are not offering the most cutting-edge functionality. You cannot risk an enterprise workflow on a vendor that could surprise you in a not-so-good way.

For companies I believe in, I will even try to nudge them, help them, or bring people to help them with the things that make for a stronger answer to that rug-pull question: clearer contracts, better implementation support, stronger data portability, more transparent roadmaps, robust partnerships, better documentation, or a more credible plan for long-term customer success.

The product matters. But in enterprise restaurant tech, trust matters just as much.

It is tricky to talk about established tech providers as one group, because their situations vary widely. Some are modern technology companies; others are payments companies with little R&D. Some are legacy platforms with deep customer relationships but aging architecture, some are service businesses with software attached, and some are software companies in name but not yet in operating model.

The established companies that incorporate AI the right way i.e. thoughtfully, flexibly, safely, and fully integrated into real workflows, will have big advantages. And they probably have to do it. If they do not, someone else will.

But I would sharpen the common claim that incumbents have an advantage because they "have the data." Data is only an advantage if it is clean, permissioned, normalized, connected to outcomes, and usable inside the product experience.

Restaurant data is messy. POS data, labor data, inventory data, ordering data, loyalty data, delivery data, franchisee data, accounting data, guest feedback, and operational data often live in different systems with different definitions. Having access to data is not the same thing as being able to turn it into intelligence.

The real incumbent advantage is broader than data. It is distribution, workflow context, integrations, trust, implementation capacity, historical operating knowledge, and the ability to embed AI into places where work is already happening.

That is why I think the winners among established providers will not be the ones that simply add an AI feature or put "copilot" in the roadmap. The winners will be the ones that use AI to make the product meaningfully better: fewer clicks, better decisions, faster onboarding, stronger support, more accurate forecasting, smarter automation, and fewer operational surprises.

The danger is that investor pressure pushes companies to "do AI" quickly in a way that creates more mess than value. AI cannot just be a press release, a chatbot, or a feature checklist. It has to improve the customer experience, the operator experience, or the economics of the business. If it does not, customers will eventually see through it.

There is another group worth watching: high-end consulting firms, agencies, and custom development partners.

I'm bullish on this layer becoming more important. Many consulting firms and agencies develop deep domain expertise. They understand specific brands and their operational constraints, and they know the difference between what sounds good in a boardroom and what a general manager will actually use on a Tuesday night.

In some cases, these firms may be better positioned than a traditional SaaS vendor to help a brand create exactly what it needs. They can combine restaurant knowledge, process work, change management, data strategy, AI tooling, and development talent. For complex enterprise brands, that can be extremely valuable.

But custom software is not a free lunch. A consulting firm can become a strategic partner, or it can become a different kind of vendor lock-in. The same questions still apply: Who owns the code? Who maintains and supports it? Who documents it? And what happens when the original team rolls off, leaving the brand with a one-off tool that becomes expensive to change and hard to integrate?

This is why I think the most interesting future is not pure SaaS versus pure internal build. It is a hybrid market where brands, vendors, agencies, consultants, and model providers all play a role. That will create opportunity, and it will also create confusion.

The obvious counterargument is that most restaurant brands should not become software companies.

Restaurants are hard enough. The hidden cost of custom software is rarely the first build. It is the second year of maintenance, the edge cases, the integrations, the support tickets, the security reviews, the data cleanup, the franchisee rollout, the training, and the product decisions no one budgeted for.

In franchised systems, the question is even more complicated than whether corporate can build something. As many industry colleagues note, it is whether franchisees will trust it, pay for it, use it, and see enough restaurant-level value to change behavior. That is a high bar.

But the fact that building may be risky does not mean brands will stop considering it. AI is changing the psychology of the decision. It is making more leaders ask, "Why not us?" And once that question enters the chat, vendors have to answer it.

I do not think AI means restaurant brands will build everything themselves. I also do not think it means startups will wipe out incumbents, or that incumbents will automatically win because they have data and distribution.

What I do think is that AI is changing the perceived economics of restaurant software. It is making brands more willing to build or custom-build. It is making startups easier to launch but harder to trust. It is forcing established providers to prove that their data, integrations, and customer relationships can become better products, not just better sales decks. And it is creating a bigger role for agencies and consulting firms that can translate domain expertise into usable software.

The new enterprise question goes beyond whether the product “works”, by all of the definitions of the term, It now includes whether a brand can trust this company, this architecture, this data model, and this roadmap with a core operating workflow. That is the question every restaurant technology provider should be preparing to answer.