Mikkel Andersen
AI & Digital Innovation Lead
AI is no longer confined to the desktop — it’s moving into the very engine room of digital business. It’s becoming an integral part of the architecture itself — an engine that connects data, processes, and customer experiences in real time. This shift unlocks new possibilities for efficiency, personalization, and innovation, but it also raises an important strategic question: Will you let third-party platforms define your framework — or will you build your own AI layer that gives you full control, governance, and ownership over your digital ecosystem?
In the beginning, LLMs were tools.
We used them to draft texts, summarize meetings, or generate (sometimes decent) images. But that definition is quickly becoming outdated. The line between using and building on AI is getting thinner by the month.
The major players — OpenAI and Anthropic — are no longer just developing tools; they’re building ecosystems. Their models are evolving from being the hammer to being the whole toolbox.
Recent announcements make this clear. OpenAI’s Instant Checkout, APC/AP2 protocols, and in-app experiences within ChatGPT show how tightly they’re weaving AI into digital workflows. Add ChatGPT Atlas, OpenAI’s new browser, and the picture becomes clearer: ChatGPT is positioning itself as more than a chatbot — it’s a platform, even an operating system.
Anthropic is following the same path. Claude’s integration with Microsoft 365 allows it to access and act on emails, documents, and files across apps. Claude Skills extend this further, enabling specialized modules and automated workflows tailored to specific business tasks.
Both companies are building AI that talks directly to the systems behind the interface — fetching CRM data, updating product descriptions, or following up with customers in real time.
The inspiration is clear: just as the web browser became the new operating system for many of us, AI assistants like ChatGPT and Claude are evolving into Conversational Operating Systems — places where most of our work and communication happen in collaboration with AI.
Still, one question remains:
Is this truly a functional transformation — or simply a matter of semantics?
A New Power Play: Platforms vs. Control
The latest moves from OpenAI and Anthropic reveal a familiar dynamic — two different paths, one shared destination.
Anthropic is doubling down on the enterprise world, focusing on secure integrations, productivity, and governance for organizations that demand stability and control. OpenAI, on the other hand, is expanding its reach into the consumer space — launching browsers, open app ecosystems, instant shopping features, and partnerships with brands like Spotify, Booking.com, and Shopify. Even initiatives such as age verification hint at a desire to make ChatGPT more personalized and less restrictive for adult users.
It’s reminiscent of an earlier rivalry — Microsoft vs. Apple. Both chased the same markets but ultimately carved out distinct territories: enterprise versus consumer, B2B versus B2C. The same pattern now seems to be unfolding in the AI landscape.
Despite their different routes, both OpenAI and Anthropic share the same endgame: becoming the foundational layer others build upon. When AI systems evolve from tools into infrastructure, they become integral to the digital ecosystem — influencing everything from business operations to everyday interactions.
For e-commerce and digital businesses, this shift means AI will not only enhance customer experiences but actively shape and direct them. And that raises critical questions about data ownership, responsibility, and long-term dependence on these platforms.
After all, the more tightly you integrate with a platform, the more control you hand over. Convenience and speed come at a cost — flexibility. Choosing your AI architecture, therefore, isn’t just a technical decision; it’s a strategic one about control, autonomy, and the future direction of your digital foundation.
Who Holds the Reins?
As AI platforms evolve toward becoming full-fledged operating systems, companies must treat them like any other core component in their tech stack. The challenge lies in striking the right balance — harnessing the immense potential of AI while maintaining control over data, security, and ownership.
At Vertica, we’re genuinely excited about what AI can do for e-commerce. While we remain cautious about whether language models can truly serve as end-to-end control systems, we embrace their ability to bring intelligence closer to business strategy, automation, and data-driven decision-making. Still, as with any transformative technology, integration must be done with intent and awareness.
That’s why we encourage organizations to define their own framework rather than relying entirely on external ecosystems. Tools like Anthropic’s Claude Skills and OpenAI’s Agent Builder offer impressive automation and orchestration capabilities — but they also create dependencies. When you build within their walls, you also accept their limits.
By contrast, designing your own Mixture of Experts architecture allows you to tailor AI to your business needs — optimizing for accuracy, scalability, cost, or compliance — without locking yourself into a single vendor. It’s not just a matter of flexibility, but of sovereignty.
Most importantly, this approach ensures that your data, your rules, and your values remain at the core. As AI embeds itself deeper into business infrastructure, data governance becomes more than a compliance requirement — it becomes a marker of brand integrity and trust.
Because in the end, while OpenAI and Anthropic may aspire to be the control systems, your competitive advantage lies in ensuring that you remain in control. The future won’t just reward those who use AI fastest — but those who build it wisely, with ownership, accountability, and long-term vision.
Discover where AI can truly transform your business — at AI Summit 2025.
