The rapid evolution of AI models—specifically the recent releases of GPT-5.3 Codex and Claude 4.6—has sparked a significant shift in how digital leaders think about work. A recent essay by Matt Shumer, Something Big Is Happening, has been making the rounds because it highlights a transition from AI that assists to a system that finishes. Shumer notes that these models are moving past simple execution toward a level of judgment and taste that was previously reserved for humans.
For health system marketers and digital leaders, this isn’t just about a smarter chatbot. It is a fundamental change in what your Digital Experience Platform (DXP) can actually do. We are paying close attention to how this shift bridges the gaps in transactional experiences where patients typically encounter friction.
From Task-Based to Goal-Oriented AI
Most AI in healthcare today is reactive: you prompt it, and it generates a result. What’s cool about the current shift is the move toward Agentic AI. This is goal-oriented. Instead of waiting for a command, an agent can plan and execute multi-step workflows.
In a DXP context, this means moving toward systems that don’t just host content but coordinate the patient journey. While we aren’t seeing health systems hand over the keys to autonomous agents just yet, the major platforms are building the infrastructure to make this possible.
What We’re Watching in the DXP Market
We are tracking how the leading platforms are re-architecting their cores for this agentic future. Here is what stands out:
- Optimizely and Opal: Optimizely Opal is focusing on agent orchestration. We’re particularly interested in how agents can now autonomously query content structures within the Optimizely Graph to identify which care-finding content is converting best and propose data-driven variations without a human needing to run a manual audit.
- Adobe and the Experience Platform Agent Orchestrator: Adobe Experience Platform is positioning their agent as a reasoning engine. Their Audience Agent can analyze first-party data from your CRM to find specific patient groups—like those who searched for a specialist but didn’t book—and then autonomously build that audience and trigger a follow-up journey.
- Sitecore and Sitecore Stream: Sitecore Stream focuses on brand and clinical safety. Their Agentic Studio uses brand kits to ensure that any automated content or patient interaction remains compliant with your specific clinical tone-of-voice. For a $2B+ health system, this level of governance is the floor, not the ceiling.
- Acquia and Acquia Source: Acquia AI is embedding agents directly into their SaaS CMS to solve bottlenecks. Their Site Builder Agent allows teams to create multi-page campaign sites from a creative brief in hours, while their Web Governance Agent autonomously identifies and fixes accessibility and compliance issues.
The Clinical Core: Connecting the EHR
An agent is only as smart as the data it can see. For these workflows to matter in healthcare, they must connect to the Electronic Health Record (EHR). We are continuing to see and help clients execute the shift toward HL7 FHIR and SMART on FHIR as the standard.
- Secure Environments: Platforms like Adobe use Healthcare Shield to ingest ePHI directly from systems like Epic or Cerner into a HIPAA-ready environment. This allows agents to “reason” across clinical data—identifying care gaps and triggering outreach—without compromising security.
- Real-Time Scheduling: We are watching how platforms pull real-time scheduling data into their AI layers. This allows an agent to see a last-minute cancellation in the EHR and immediately prioritize that slot in the “Find a Doctor” results on the public site.
- Seamless Handoffs: Integration with portals like Epic MyChart ensures that a patient can move from the public site to the secure portal with their context intact, making the transition to booking as frictionless as possible.
How We Deliver
At Modea, we don’t view these developments as flashy experiments. We see them as a way to work smarter and faster for you. We are actively incorporating these types of evolving models into our own internal delivery workflows—using them to bypass manual bottlenecks in technical architecture and project management.
This allows our human experts to stay focused on high-level strategy and the operational stability your system requires. We aren’t here to add to the word soup of AI hype; we are here to help you use machine power as a multiplier for your team, solving the friction points that directly impact your margin.
The models are getting smarter. The goal now is to ensure your DXP and data infrastructure are ready to support these goal-oriented workflows as they become viable.