Of course, there are still differences between models and applications. But access to good AI is becoming less scarce. That leads to a more interesting question: if more and more organizations have access to similar AI, where will the competitive advantage come from?
In my view, it won't come from the model itself. The real difference lies in what AI knows about your organization, how well your processes are designed, and how effectively AI can actually contribute within those processes.
From Access to Application
That shift is already becoming clear. In its recent Enterprise Signals research, OpenAI found that organizations leading the way don't necessarily have access to different AI models. The difference lies mainly in how they use AI. Leaders give AI more context and tools, moving from general assistance toward actually carrying out work. (openai.com)
From a marketing perspective, I find that perhaps more interesting than the next model release. Technology only becomes a real differentiator when an organization uses it to do something a competitor can't simply copy.
A general-purpose AI assistant can write an email, summarize a report, or come up with campaign ideas for almost any company. But it doesn't automatically know which customers are strategically important to your organization, why certain deals are won or lost, which exceptions are common in an order process, or what agreements are in place with suppliers.
That's business context. And every organization has built its own.
Years of Knowledge Are Scattered Everywhere
Over the years, organizations have accumulated enormous amounts of proprietary knowledge. Some of it is neatly stored in CRM, ERP, and other business applications. Another part lives in documents, dashboards, work instructions, project environments, and emails. And a surprisingly large amount probably still exists in employees' heads.
That means an AI model, by itself, is only a small part of the solution.
Take a sales AI assistant. A general-purpose model understands sales perfectly well and can suggest a follow-up email or help prepare for a customer conversation. But to genuinely support an account manager, it also needs to know something about the customer, previous interactions, open proposals, products, margins, and commitments that have been made.
The same applies to customer service. Generating a friendly response is no longer particularly difficult. It becomes much more interesting when AI also understands which product the customer uses, what issue they reported previously, what service agreements are in place, and what the logical next step is.
The same principle applies to logistics. The differentiator isn't an AI model that understands what inventory management means. The value comes when AI can take into account actual inventory levels, delivery schedules, customers, exceptions, and operational agreements specific to that organization. That's when AI stops being a general-purpose tool and starts becoming technology that genuinely understands the business.
Context Is Becoming a Business Asset
Major technology companies are clearly moving in this direction. In September, Salesforce introduced a new Enterprise AI Harness focused on exactly this challenge. The AI model itself is only one part of the picture. The approach also brings together business data, processes, business rules, and the ability to take action based on that context. (salesforce.com)
Salesforce and Google Cloud have also announced further integration that allows AI agents to work with the same data, workflows, and business logic across different platforms. (salesforce.com) Google Cloud and Accenture are also building a joint team of 1,000 specialized engineers to help companies move from experimentation to real-world applications across the organization. (googlecloudpresscorner.com)
These developments point to where the market is heading. The challenge is becoming less about whether organizations can use AI. The more interesting question is how to make sure AI understands your organization well enough to actually do something useful.
So Don't Start With the Model
AI conversations often start with the technology. Should we use ChatGPT, Gemini, Claude, or something else? Which model performs best? Which solution should we buy? Those are reasonable questions, but for many organizations, I don't think they're the best place to start.
Start with a process instead. Look at where people are losing time, where employees repeatedly have to gather information from different systems, or where the same decision has to be made over and over again. Look for bottlenecks caused by knowledge being difficult to access, and situations where someone would benefit not just from finding information, but from having it placed in the right context immediately.
Only then does the technology become interesting. Maybe a simple AI assistant is enough. Maybe information needs to be brought together from multiple systems. And sometimes you'll find that you don't need AI at all, and that the smarter move is to improve the underlying process first.
That's something that can easily get lost in all the excitement around AI. A bad process doesn't automatically become a good one just because you add AI to it.
Your Way of Working Is Harder to Copy
That's where I see the biggest strategic opportunity. A competitor can buy the same AI model and probably use much of the same software. Features that seem unique today can become standard within a matter of months.
What a competitor can't simply copy is the context your organization has built over years. Customer knowledge, processes, data, experience, exceptions, and decision logic are all specific to the way your organization operates. When AI gets access to that exact context and can use it in the right way, you create something that is much harder to reproduce.
Organizations that get this right are ultimately building more than a more efficient chatbot. They're making their existing knowledge more accessible and connecting technology more closely to the way the business actually works.
That doesn't mean you need to automate entire business processes from day one. A clearly defined process with a specific problem and a measurable outcome can be a much better starting point. The question then shouldn't just be whether the AI works technically, but whether it actually improves something. Do employees get the help they need faster? Do customers get answers sooner? Do manual steps disappear? Does it create more time for work where human attention adds greater value?
AI Becomes the Foundation. Context Makes the Difference.
We're probably heading into a period where AI will increasingly stop being seen as a separate technology and become part of the software and processes organizations use every day. As that happens, some of the differentiation that comes simply from having access to powerful AI will disappear.
Not everyone will literally have the exact same AI, but the differences in access will continue to shrink. That changes the playing field. The quality of the model will still matter, but so will the quality of the context you can provide it with and how effectively you use that context across your organization.
Organizations thinking about this now may have a more interesting starting point than those simply waiting for the next model or AI feature. Not because they necessarily have better technology, but because they have a clearer understanding of where their own value lies and how technology can support it.
The question is becoming less about how smart your AI is. A much more relevant question is how much your AI understands about your organization, what information it has access to, and what role it can play within your processes.
