AWS Summit London and Oracle AI World Tour Underscore a Dual-Track Strategy for Enterprise AI

As the spring season is traditionally very rich in industry events, it is time to look back at Oracle’s AI World Tour in March and the AWS EMEA Summit in April to get an understanding of the state of the market and understand where the industry is heading. At the end of March, Oracle launched 22 Fusion Agentic Applications embedded directly into Oracle Fusion Cloud, spanning Human Resources (HR), supply chain, payroll, and customer experience, alongside an expansion of its AI Agent Studio for customers to orchestrate and configure their agentic workforce.

At its EMEA summit at the end of April, AWS expanded the existing Connect offering into four dedicated Agentic Artificial Intelligence (AI) solutions covering supply chain, hiring, customer experience, and healthcare, while also advancing its Amazon Bedrock AgentCore platform, deepening its partnership with Anthropic by making Claude available within the Bedrock environment, and launching Amazon Quick, an AI assistant designed to streamline work processes by connecting with various tools, learning user preferences, and automating tasks accordingly.

What is striking about both events is less the technology on display and more the commercial logic driving it. Rather than articulating a long-term vision, both Oracle and AWS were focused on near-term revenue opportunity, converting enterprises trapped in legacy systems into paying customers for agent-led modernization. The pitch is essentially transactional, and the deeper questions about how Agentic AI reshapes business models and competitive advantage over the longer term were largely left unanswered.

This is in stark contrast to recent activities of Graphics Processing Unit (GPU) vendors NVIDIA and AMD or neocloud providers (e.g., Vultr, Nscale and Nebius) that are pushing technology innovation to drive the adoption of enterprise AI. So-called AI Factories and other reference designs provide an infrastructure and software package for innovative enterprise AI use cases (drug discovery & molecular simulation, materials science & compound optimization, supply chain resilience & demand forecasting, financial fraud detection & real-time risk scoring and others). The uptake of these offerings, however convincing the value proposition may sound, remains focused on particularly tech-savvy enterprises and innovative startups.

These observations show, once again, that the enterprise vertical landscape is incredibly fragmented. What we have seen with other technologies (connectivity, most importantly) is even more the case for more disruptive technologies, like enterprise AI. On one hand, there are tech-savvy enterprises that keep abreast with new innovation and are on the forefront of enterprise AI adoption with important implications for technology providers. Not only are these enterprises more receptive to disruptive value propositions, but they are also open to engage in new partnerships and willing to adapt their current supply chain to providers of new technologies.

On the other hand, there is a formidable body of enterprises (across all different industry verticals, incidentally) with a sizable body of legacy operations. These are not redesigning and developing new processes and workflows, but instead using technology innovation to improve the efficiency of existing legacy processes. Large industrial conglomerates are running decades-old Enterprise Resource Planning (ERP) systems that bolt onto AI dashboards without restructuring underlying data pipelines. Energy & utilities companies are applying predictive maintenance AI on top of fragmented Supervisory Control and Data Acquisition (SCADA) systems without consolidating operational data. And hospital networks are deploying AI scheduling tools on top of siloed Electronic Medical Records (EMR) systems. These enterprises continue to rely on their incumbent digitization partners, like System Integrators (SIs) for any production infrastructure upgrade and hyperscalers for data storage and management.

So far, the efforts to advance enterprise AI adoption have been focused on the few lighthouse projects of particularly tech-savvy enterprises that are ready to throw their existing processes overboard (if indeed they have had any) to embark on a journey of embracing new disruptive technologies, like enterprise AI. While this should remain part of the strategy to bring technology innovation to enterprise verticals, technology vendors and innovative service providers will also need to develop a strategy on how to address and what to offer to large enterprises with a multitude of existing legacy workflows.

The magnitude and attendance levels of recent hyperscaler events (AWS Summit and Oracle AI World Tour) highlighted that these efforts should be combined with a dedicated strategy that addresses enterprises with a considerable amount of legacy workflows and processes. Consequently, technology disruptors should target these legacy-heavy enterprises to increase their Total Addressable Market (TAM) (in the short term) and drive these enterprises toward a lasting partnership with the potential to upsell in the long term. At its core, this requires component vendors and Communication Service Providers (CSPs) to leave the realms of technology innovation and meet enterprises where they are. This should follow these distinct steps:

1. Understand the enterprise's actual pain points and current workflows. To be able to meet traditional enterprises at their current state on the enterprise digitization/AI journey, cloud service providers and computing hardware vendors first need to gain a thorough understanding of existing enterprise workflows and bottlenecks. Identifying where manual effort, delays, or data fragmentation are creating the most friction allows providers to ground their AI strategy in operational reality and gives them the right tools to develop solutions for the most pressing enterprise use case cases/pain points.
2. Identify where enterprise AI can run within the existing stack without replacing it. Based on the insights provided, hardware vendors and service providers should identify clear integration points that allow AI to sit alongside existing systems and identify opportunities for AI to be embedded into current applications, data pipelines, and business processes without major transformation/disruption. Not only does this approach reduce friction (by working within the enterprise’s established architecture), but it also enables faster deployment by leveraging infrastructure and systems already in place.
3. Build up channel partnerships that align with existing enterprise supply chains. Previous waves of enterprise technology (most recently, private cellular) assumed that enterprises would abandon their existing supply chain and transformation partners to switch to new specialist vendors. In reality, however, this shift has rarely happened at scale. These enterprises tend to adopt innovation through existing procurement, integration, and service relationships. Consequently, hardware vendors and service providers should look at established (industry-specific) enterprise SIs and Independent Software Vendors (ISVs). These also have relevant expertise in enterprise workflows and pain points, which makes developing appealing solutions easier.