AaaS (Agent as a Service): The Next Transformation in Healthcare

AI hype is everywhere, and the healthcare industry is no exception. Established companies and startups are touting how their use of AI can solve critical problems that have plagued healthcare for decades. However, many of these organizations make assumptions regarding the quality of (and access to) medical data used for their AI models that fail in real-life scenarios.

AI is not a panacea for all that ails healthcare.  In specific use cases, it is very effective, but in a lot of scenarios, AI isn’t needed. Something that can be easily addressed with quality engineering solutions and analytics and brings good ROI to participants is unnecessarily being over-engineered into AI-, Gen AI-, and Agentic AI-based solutions by many organizations, especially consultants who have been the biggest beneficiaries of the AI hype.

Does AaaS (Agent as a Service) have the potential to transform healthcare solutions? For specific use cases, it absolutely can be very effective. This article explores the basics of Agentic AI and how it is different than a typical SaaS-based architecture. 

Agentic AI basics

Agents are essentially autonomous or semi-autonomous codebases that can perform three key sub-tasks: 

• Access different data sources and synthesize data in real time
• Automate decision making process(es) based on analysis of the data
• Automate routine tasks by leveraging process automation tools, integrations to/from different systems and orchestrate workflows

Single as well as multiple agents can collaborate on interconnected tasks.  Single-agent systems can be used to automate standalone processes like claims validation, patient scheduling, and appointment reminders. Multi-agent systems are needed to handle more complex episodic events and workflows across multiple teams and systems like a care transition for knee surgery that involves hospitals, payers, different physician teams, and community health team members. 

In a multi-agent system, for example, one agent is responsible for integration (APIs, batch-based ETLM processes, real time connection to EHRs, etc.), another handles data analysis and memory retention, and a third agent performs task orchestration. Such a multi-agent system can improve coordination between payers, providers, CBOs, patients, and all entities involved in the episodic care of the patient.

Agentic AI-based systems can be useful partners for the healthcare workforce – enabling our physicians, nurses and caregivers the enhanced capabilities of diagnosis, knowledge, and automating of tasks while maintaining the knowledgeable assistant-like feature to ensure the human aspect of healthcare remains intact to provide the best care for patients.

A case for AaaS in lieu of SaaS

The majority of SaaS-based applications that exist today have the following main tiers:

• User Interface (UI) tier for workflows/user interactions 
• Business logic tier that handles the different CRUD (Create, Read, Update and Delete) operations over relational and/or non-relational data stores
• Integration with other systems via APIs and Batch-based Extract-Transform-Load- Modify (ETLM) processes
• Data tier that handles relational and non-relational data, reporting and analytics

Given the pace of AI advances, much of the business logic soon will be handled completely by AI agents. Once that is accomplished, there is really no need to have a traditional SaaS-based model. AI agents will be able to understand what users want/need, anticipate the requests, and eliminate the need for the current model of SaaS applications.

AaaS use cases in healthcare 

Use cases addressed in a typical SaaS-based implementation for value-based care (VBC) include:

• “Network of networks” implementation
• Contract builder, contract modeler, and contract management
• Patient Longitudinal Health Record
• Care engagement (tracking patient data, sending reminders, referral management, analyzing trends for high-risk patients, etc.)
• Outcomes reporting using analytics over different datasets
• To/from integration with different systems (EHRs, different source systems for payers/providers/employers)

Many of the activities listed above require effort from healthcare team members to set goals/objectives, analyze the data, release payments, and take different types of actions. In an Agentic AI-based architecture, some decision-making processes and actions can be automated. Here are four examples:

1. Activity/Task: Identification of at-risk patients (risk stratification) and appointment scheduling

Traditional SaaS: Patients who missed appointments or have worsening vitals are flagged. A caregiver/nurse reviews the list, decides who to contact, and schedules appointments for follow-ups.

Agentic AI: Automated identification of at-risk patients, automated contact via text/email/WhatsApp, appointment scheduling, and respective entries into different systems

2. Activity/Task: Claims processing

Traditional SaaS: Identification of issues (even based on AI algorithms) resulting in claim denials can be achieved, but still requires intervention from healthcare workers in order to trigger different resolution workflows  

Agentic AI: Automatic validation of claims, identification of any missing pieces of information, trigger any workflows that require resolution, and reduce denials.  AI agents, in this case, can utilize LLMs for clinical documents interpretation and extraction/matching for coding accuracy

3. Activity/Task: Chronic conditions management 

Traditional SaaS: Reports are produced showing which diabetic patients need better glucose control. The physician or care team member then needs to review and decide the next steps.

Agentic AI: Automated identification of the patient, preparation and communication of personalized dietary advice, order blood tests (if needed) and alerts to the care team if a patient’s condition does not improve. Since AI agents are context and memory aware, they can recall previous case adjustments for patients and provide the required information to the case managers for action.

4. Activity/Task: Transition of care between care teams

Traditional SaaS: Majority of the task hand-offs are manual in nature today across care teams, and workflow-based tools do not integrate across care settings. 

Agentic AI: AaaS-based platforms can facilitate real-time coordination, making seamless transitions for inpatient, outpatient, and post-acute settings.

Technical implementation of Agentic AI for outperforming a SaaS 

Agentic AI brings automation, personalization, and adaptive learning to healthcare – transforming traditional SaaS tools into proactive care solutions. Instead of just presenting insights, Agentic AI acts on them, improving efficiency and patient outcomes.

The key technologies powering Agentic AI in healthcare are:

• Large Language Models (LLMs): For understanding medical notes and automating communication
• Computer Vision: For analyzing medical imaging (e.g., X-rays, MRIs)
• Reinforcement Learning: For optimizing care pathways by learning from outcomes
• RPA (Robotic Process Automation): For automating repetitive tasks like data entry and appointment booking

The diagram below illustrates the key components of an AaaS platform in a cloud infrastructure:

Figure 1: High Level Block Architecture Diagram of an AaaS implementation

Agentic AI pilots in healthcare

There are numerous Agentic AI pilots currently under way in healthcare. Among them: 

• VoiceCare AI: Agent for providers’ revenue cycle management teams
• Hyro: Enabling health systems to automate workflows and conversations across various platforms and channels using conversational AI
• Notable: Agentic AI to automate healthcare tasks, with AI agents performing these tasks on behalf of health systems, aiming to improve efficiency and reduce administrative burdens
• Luma Health: Agentic AI to create a patient success platform that streamlines patient interactions, reduces manual tasks, and improves staff efficiency by automating tasks like scheduling, communication, and fax processing

Conclusion

While a traditional SaaS implementation gives healthcare teams needed data and insights (analytic outputs), an AaaS-based implementation can analyze, decide, and then act on the data, essentially automating much of the process and thus helping to improve patient outcomes. It can provide proactive care, automate repetitive tasks, personalize patient experience, and proactively prevent serious issues from happening, thus reducing costs while improving outcomes.

Platforms, solutions, tools and utilities that utilize Agentic AI architectures are essentially meant to help and enhance productivity, reduce errors, and provide better care while reducing physician burnout. These cannot replace healthcare workers but can act as powerful assistants that help provide better healthcare.

The success of AaaS-based platforms and solutions will depend upon the measurable results aligned with the needs of the business, while the biggest challenges are going to be in co-existing with current applications while the transition happens for specific use cases.