Assess the Potential of Emerging Technologies for Healthcare Payers
19 December 2025 - ID G00834332 - 27 min read
By Mandi Bishop, Austynn Eubank, and 1 more
Health insurance CIOs should regularly assess far-horizon technologies based on their potential impact on their organizations. Use this evaluation framework of 25 emerging technologies — including agentic AI, quantum computing and digital twins — to accelerate and inform strategic planning.
Overview
Key Findings
- Many CIOs lack a disciplined approach to horizon scanning, and technology evaluation is often reactive, not strategic.
- Adoptable technologies like GenAI virtual assistants and AI avatars deliver clear value with low complexity. Enterprise vendors are embedding these capabilities now, enabling rapid deployment.
- Adaptable technologies — agentic AI, domain-specialized language models, and knowledge graphs — promise high impact but require significant integration, IP development, and governance.
- Situational technologies such as quantum computing and decentralized identity remain high-hype, high-complexity bets with long time horizons.
Recommendations
- Embed emerging technology evaluation into annual strategy planning to anticipate disruption and opportunity.
- Position IT as an innovation partner by regularly briefing executives on technology promise, risks, and readiness to build enterprise confidence.
- Prioritize near-term wins and focus resources on adoptable and adaptable technologies with clear ROI potential.
- Invest strategically by only allocating resources to situational technologies if your organization has a research mandate or aggressive innovation goals.
Why This Matters
This document was revised on 26 December 2025. The document you are viewing is the corrected version. For more information, see the Corrections page on gartner.com.
Healthcare payer CIOs must monitor emerging technologies that could disrupt their business, create a competitive advantage or generate new value. In fact, 52% of the 2025 Gartner Business Outcomes of Technology Survey respondents report that they are investing in new technologies to radically transform ways of working — up from 15% in 2024. Yet, Gartner interactions indicate most payer CIOs don’t prioritize horizon scanning for emerging technologies. Gartner’s Emerging Technology Evaluation Framework (ETEFx) provides an early evaluation of 25 emerging technologies that will stimulate conversation within the C-suite, helping CIOs guide expectations. Technologies are assessed for their potential use for payers and then scored on their suitability and complexity for a typical payer organization.
The emerging technology landscape is rapidly shifting with new technologies frequently surfacing. This year’s ETEFx list includes technologies that will be highly impactful across industries and includes nine new entries along with 15 entries retained from the 2024 list.
CIOs should use Figure 1 as a starting reference point to discuss and engage stakeholders regarding emerging technologies through the lens of their organization’s priorities (see the Risk Profile Considerations section).
Table 1 shows the criteria underpinning the “suitability” and “complexity” dimensions.
Emerging Technology Evaluation Criteria
Suitability criteria | Complexity criteria |
Depth of impact: In the use cases where the technology is adopted, how much new value will be realized? | Integration: When the decision is made to adopt the technology, what level of effort will be required by the adopting organization to realize the expected value? |
Breadth of application: How pervasively can the technology be applied across the value chain, functions and stakeholders that make up the industry? | Contextual IP: How much industry-specific intellectual property (IP) needs to be developed (inclusive of IP created by vendors, system integrators and the adopting organization) for the technology to deliver value in the industry? |
Industry readiness: To what extent is the industry in its current state (cultural, regulatory, political, environmental and societal) positioned to embrace (and not resist) adoption of the technology? | Interdependency: For an organization that would adopt this technology, to what extent are there requirements to achieve the expected value that reside outside the control of the organization itself? |
Source: Gartner (December 2025)
Analysis
Interpreting the Framework
Figure 1 is split into four areas that help healthcare payer CIOs interpret each of these technologies. The gray background color highlights the two areas where CIOs should initially focus: adaptable and adoptable. This assessment reflects the needs and opportunities of CIOs in an average healthcare payer organization — but every organization has its own unique strengths and challenges, financial and operating models, and market context. Thus, CIOs should use this research as an accelerated start to far-horizon scanning to be further refined according to the unique context of their organization and role.
Adaptable
Adaptable technologies have high potential industry value but will need significant effort to unlock. Efforts could relate to the adaptation of intellectual property and integration with business processes or protocols. CIOs should highlight foundational technology investments that might support the use of such technologies, along with any skills or competencies they will need to source or develop.
Emerging technologies often carry bold promises of business model transformation and cost containment. For example, knowledge graphs can unlock advanced personalization and fraud detection by surfacing hidden relationships across member and provider data. But adoption demands new data architectures and cultural buy-in — a tough sell for CIOs facing modernization fatigue. Domain-specific language models are another high-value technology that will require considerable industry- and enterprise-specific IP to realize their promise as foundational components for capabilities such as agentic AI and GenAI assistants.
Adoptable
Adoptable technologies are those that have high potential industry value and do not require significant effort to verticalize. Their adoption will be predominantly in their out-of-the-box form. Ease of adoption means CIOs should look at where and how to apply the technology, as well as the appropriate timing.
Adoptable technologies offer quicker wins than other emerging technologies as they have a thriving vendor landscape with production-ready use cases. For example, GenAI virtual assistants are already embedded in operational platforms, streamlining claims, eligibility checks, and member engagement. Configurable AI avatars support functions like member education. These solutions deliver measurable efficiency gains without major architectural overhauls.
Marginal
Marginal technologies have low potential industry value but will also require low effort to verticalize to the industry. CIOs should steer their enterprises away from spending significant time on these technologies while continuing to monitor them for easy absorption of some marginal value or for changes in use cases that increase the technologies’ value potential.
Technologies such as autonomous UAVs, smart spaces and private 5G have negligible (if any) value to traditional payers that do not have care delivery or retail pharmacy businesses. Similarly, neuromorphic computing and spatial computing have limited applicability in remote patient monitoring, where existing technologies are already sufficient for achieving value.
Situational
Situational technologies have low potential industry value for the average organization and would require significant effort to verticalize. For many industry participants, this effort may result in their enterprise ignoring these technologies. However, technologies in this category may still benefit some industries with niche processes or lines of business. CIOs should avoid overspending on these technologies and set expectations on their value and the likely significant effort to implement.
Situational technologies often receive tremendous hype but rarely deliver near-term value. Tokenization promises stronger privacy controls, yet implementation costs and disruption outweigh incremental benefits. Similarly, cybersecurity precrime platforms sound compelling in an era of AI-driven threats, but most payers lack the data maturity and partnerships to realize meaningful ROI. CIOs should monitor these technologies but avoid overinvestment.
Time to Pilot
Table 2 indicates our view on when a technology will be mature enough to start prototyping.
Time to Pilot Emerging Technologies for Healthcare Payer CIOs
Time to pilot | Adaptable | Adoptable | Marginal | Situational |
0 to 2 years | ||||
2 to 5 years | ||||
5 to 10 years | ||||
Source: Gartner (December 2025)
Risk Profile Considerations
Emerging technologies will move toward maturity and adoption at a varying pace. Payer CIOs must decide on their appetite to engage with these technologies based on multiple internal factors, including their individual enterprise risk profiles. Variables include:
- Attitudes toward technology risk
- Geographical relevance
- Line-of-business (LOB) suitability
- Level of technology investment
- Approach to technology exploration
- Success in technology innovation
- Regulatory Implications
- Cultural readiness to adopt technologies
- Value potential beyond existing approach
Next Steps
Healthcare payer CIOs should:
- Educate stakeholders by building executive understanding of emerging technology impact and align enthusiasm with risk awareness.
- Cultivate innovation partnerships with senior business leaders to ensure enthusiasm doesn’t outweigh rigor in introducing new technologies.
- Plan adoption pathways by using this evaluation to assess architecture readiness and resource needs for high-value technologies.
- Engage vendors strategically by requesting roadmaps to confirm how and when they will integrate priority technologies.
Definition
These definitions have been extracted from Gartner research to provide a common understanding of the technology. This research reconciles these technologies with their potential impact on an individual industry.
AI avatars are humanlike virtual personas created using computer-generated imagery (CGI) and various AI techniques and applications, like NLP, synthetic voice, computer vision and emotion AI, among others. AI avatars can be a representation of a real person or a digital being/physical entity to represent the brand or support interactions.
Potential for healthcare payers: AI avatars can differentiate payer services in member engagement and wellness programs. While internal use cases like employee onboarding face fewer barriers, consumer-facing applications must overcome trust and potential regulatory hurdles. However, there are niche vendors providing avatar-enhanced solutions to payers today for member-focused use cases, such as Mediktor and Prsonas. Integration complexity is reduced because most avatar capabilities will be embedded in existing platforms.
Advanced behavioral detection analytics are the emerging set of technologies that analyze the behavior and activities of users, systems, applications and devices within a network or system. Using machine learning (ML) and AI algorithms, these technologies can dynamically adapt and learn from correlated data, enhancing their accuracy in recognizing evolving threats that traditional rule-based methods would miss. Such behavioral detection analytics solutions not only aid in detection, speed and confidence, but also enable future prediction of threats based on proposed security control or network changes. With the evolution of adversarial attack methods, the advancement of AI-based behavioral detection algorithms must and will continue to progress. Part of this projected progress is the trend of detection algorithms being paired with graph databases to more effectively analyze and learn from the behavioral relationships of events that make up an attack. Graph databases are able to stitch these events together in an attack sequence to increase detection fidelity.
Potential for healthcare payers: These analytics combine ML and graph databases to detect fraud schemes and insider threats that rule-based systems miss. They can identify collusion rings, anomalous access patterns, and upcoding behaviors. Adoption will strengthen compliance and security but requires robust data environments with advanced architectures such as knowledge graphs.
Agentic AI is the various architectures, techniques, and frameworks for creating single-agent or collaborative multiagent systems capable of unsupervised task execution. AI agents are autonomous or semiautonomous software entities that use AI techniques to perceive, make decisions, take actions and achieve goals in their digital or physical environments. Agentic AI will mature into fully autonomous systems where agent decisions and actions are necessary, permissible and trusted. Agentic AI is characterized by agents imbued with the following — access to historicized views of context-based decisions and workflows; adjustment of goals, decisions, plans and actions based on changing conditions; variable levels of autonomy for decision-making capabilities; adaptive execution, including multiagent collaboration to learn to complete multistep processes in digital and physical environments; use of multimodal, multimodel capabilities for domain-specific context and instructions; and “reasoning” abilities for making context-based decisions and judgments. Agentic AI differs from earlier agent systems (and robotic process automation [RPA]), traditionally guided by scripted guardrails. Agentic AI can combine predictive AI, symbolic AI and generative AI (GenAI). Often, generative models are trained for domain specificity to develop task-specific expertise. The promise of autonomous action will increasingly leverage reinforcement and zero-shot learning, simulation and playout techniques to enhance decision making and improve execution more independently across dynamic environments.
Potential for healthcare payers: Forty-five percent of payers have deployed agentic AI pilots, with most planning adoption by 2028 according to the Gartner 2025 Business Outcomes of Technology Survey. Use cases include call center automation, claims processing, member/provider experience, and data integration. There is a growing vendor market with offerings from companies such as AWS, Basys.AI, Cognigy, ExponentialAI, Five9, Google, Kore.ai, Notable Health, Persistent Systems and Salesforce. While pilots are plentiful, few organizations have achieved production deployments. In part, this is due to a definitional disconnect — 42% of payer tech leader respondents to the 2025 Gartner Healthcare Payer Agentic AI Survey report executives lack a shared understanding of what “agentic AI” means. CIOs must champion clarity, governance, and skills development to scale AI agents effectively.
Autonomous UAVs, or uncrewed aircraft systems (UAS), are autonomous flying machines, mostly used for asset inspection but expanding into small package delivery. Flight control is entirely autonomous, with pilots only required to provide destinations or way points, and intercede when (or if) remote operation becomes necessary. The primary barrier to adoption is regulatory, as concerns over safety, privacy, noise pollution and security must be overcome.
Potential for healthcare payers: This technology has negligible industry relevance, with the only plausible use case today being the improvement of medication adherence by delivering medications as soon as they are prescribed or the refill is due — which would require control over pharmacy operations and favorable regulations.
Cybersecurity precrime platforms are technology platforms that are designed to proactively anticipate and prevent cybercrimes before they occur. By leveraging a combination of large historical datasets, AI and predictive analytics software, these platforms can analyze information from past criminal activity and identify patterns in both real time and historical data to predict likely crimes. This information can then be used by crime prevention teams as part of precrime cases, or investigations, that can be tracked and monitored based on assigned risk levels. Currently, the most common use cases for these platforms are in the areas of law enforcement, fraud detection and prevention, and in preventing various types of online banking and financial services crimes. However, future use cases include, but are not limited to, detecting cyber extortion, cyber espionage, cyberbullying and cyberstalking.
Potential for healthcare payers: Cybersecurity precrime platforms could eventually support several important capabilities for payers, such as preemptive ransomware and data-exfiltration detection as well as insider-threat and privileged-access risk prediction. Once knowledge graphs have become mainstream, this proactive approach could deliver meaningful value in areas like compliance, fraud reduction and reputational risk management.
Decentralized identity (DCI) allows an entity (typically a human user) to control their own digital identity by leveraging technologies such as blockchain or other distributed ledger technologies (DLTs), along with digital wallets. By establishing trust, privacy, and security through identity attributes contained in decentralized verifiable claims, DCI provides a more secure alternative to storing identity information centrally.
Potential for healthcare payers: Decentralized identity offers strong privacy protections but faces systemic adoption barriers. Without regulatory mandates, payers are unlikely to invest due to the integration complexity and lack of the foundational infrastructure as well as the lack of a national patient identifier. If mandates emerge, the disruption risk is high, given healthcare’s scale and sensitivity.
Decision intelligence (DI) platforms are used to create solutions that support, automate and augment the decision making of humans or machines, powered by the combination of data, analytics, knowledge and artificial intelligence (AI) techniques. DI platforms must have collaborative capabilities for decision modeling, execution and monitoring. DI is used to design decision-centric solutions, explicitly model decisions, orchestrate decision execution flows, and evaluate and govern decisions and audit their outcomes. DI features can include machine learning, business intelligence, natural language processing, graph technology, AI agents, simulation, real-time event stream processing and multistructured data preparation. DI can provide advanced data and analytics insights into current and future data relationships, predictions, and simulations of “what if” scenarios.
Potential for healthcare payers: Most payers already use some of the technology components of decision intelligence and have mature practices in place to leverage capabilities such as machine learning, business intelligence and NLP. However, payers have not yet integrated AI, simulation or knowledge graphs into these capabilities. As the payer data, analytics and AI environments evolve to support these approaches and architectures, decision intelligence will deliver substantial value in areas such as risk stratification, care management orchestration, clinical operations and payment integrity.
Digital twins are a virtual representation of an entity such as an asset, person or process. It is developed to support business objectives. Digital twin elements include the model, data, unique one-to-one association and monitorability. The three taxonomy levels of digital twins are discrete, composite and organizational. These digital twin elements are built, used, and shared in enabling technologies such as analytics software, IoT platforms or simulation tools.
Potential for healthcare payers: Digital twins replicate transactional flows such as claims processing and enrollment, enabling process optimization. However, maintaining parallel environments and adapting identified improvements limit ROI. Adoption is most relevant for payers with advanced analytics maturity.
Disinformation security is an emerging category of technologies and trends aimed at providing methodological approaches for discerning trust, assessing truth and tracking the spread of information. Primary enterprise use cases for technologies within this domain include content authenticity, narrative intelligence, fraud prevention, fact checking and brand reputation. Foundational technical elements of disinformation security center on the following: deepfake detection, impersonation prevention and brand protection. Gartner has noted a focus in this category on validating integrity of real-time communication, ensuring authenticity of third-party multimedia, monitoring systems powered by large language models that track narratives on social media and dark web channels, reducing generative AI hallucination and leveraging digital polygraphs.
Potential for healthcare payers: Disinformation exacerbates existing trust challenges for health plans, especially in this new era of consumers using social media and consumer-facing LLM interfaces such as ChatGPT, Gemini and Perplexity as primary information sources. Many consumers believe widely circulated deepfake videos declaring that the health plan denies all claims or sells member data, and some will ignore valid outreaches from their plan because they’re suspicious about any communication. Early detection and response could avert membership churn or provider network disruption as well as provide auditable action for regulators. Adoption is slow due to immature vendor ecosystems and unfamiliar data sources.
Domain-specialized language models (DSLMs) are designed to understand and respond to knowledge requests within a particular knowledge area. These may include functional domains (such as sales, HR, marketing and research), industry domains (such as finance or retail) and use cases (such as recruitment or lead generation). Unlike general-purpose models like GPT4 or Gemini, which have been exposed to a wide range of topics, DSLMs are trained on focused datasets representative of a specific domain. They often rely on a task-specific architecture in order to deliver more effective results than domain-agnostic models. Because of their focused nature, they can be valuable as constituent elements of agentic solutions (broad systems that combine and take action across multiple domains).
Potential for healthcare payers: DSLMs are already powering intelligent prior authorization and claims triage, reducing manual intervention. Vendors such as Cohere Health, HiLabs, Itiliti, Ricoh and Ushur offer models trained on payer-specific policies and adjudication logic. These models will accelerate automation and serve as foundational components for agentic AI and GenAI assistants. However, most high-value use cases require considerable industry and enterprise-specific IP.
GenAI-enabled virtual assistants (VAs) represent a new generation of VAs that leverage large language models (LLMs) that deliver superior functionality well beyond previous VA methods. GenAI is being used to improve VA performance, add new functionality, extend task automation, and support new value outcomes.
Potential for healthcare payers: GenAI assistants’ adoption is accelerating for use cases such as contact center optimization, compliance analysis, eligibility checks, care management, member engagement, provider self-service, utilization management and claims reviews. These capabilities are available from vendors like Hippocratic AI, Inovaare (UsherAI), Virtusa and Wipro (PayerAI). These assistants not only improve operational efficiency and experience — they could disrupt roles such as brokers over time.
Human-centered AI (HCAI) is a common AI design principle calling for AI to benefit people and society. In some cases, it also redresses the negative consequences of human behavior. HCAI assumes a partnership model of people and AI working together to enhance cognitive performance, including learning, decision making and new experiences. HCAI is sometimes referred to as “augmented intelligence,” “centaur intelligence” or “human in the loop,” but in a wider sense, even a fully automated system must have human benefits as a goal.
Potential for healthcare payers: For many AI-based analytics and automation capabilities, health insurers should only use HCAI — human in the loop is not only prudent but also increasingly publicly and legislatively demanded. The consequences of AI-driven autonomous workflow execution have already been exemplified by lawsuits against national insurers for AI-driven claims denials. These incidents resulted in new mandates about the use of algorithms and AI in prior authorization and claims review processes, which now require HCAI and limit the circumstances for decision automation.
Intelligent applications (IAs) are applications that are augmented with AI and connected data, from transaction and external sources, to generate a system that provides contextualized features, experiences and processes, and can continually learn, improve and adapt.
Potential for healthcare payers: In the short term, intelligent applications are beginning to enter the payer technology environment via horizontal enterprise-class applications such as ERP and CRM. There is significant opportunity for these capabilities in areas such as claims, benefits configuration, HR, care management and network management. Systemic constraints will hinder the pace of enterprisewide adoption and scale of utility for industry, such as cultural resistance to hyperpersonalization and entrenched legacy systems with long-term vendor contracts.
Knowledge graphs are machine-readable data structures, representing knowledge of the physical and digital worlds including entities (people, companies, digital assets) and their relationships, which adhere to a graph data model — a network of nodes (vertices) and links (edges/arcs).
Potential for healthcare payers: Knowledge graphs enable use cases such as advanced fraud detection, healthcare and social care provider network analysis, and care coordination by uncovering complex relationships across datasets. They are becoming a must-have feature of data fabric architecture and are foundational for AI techniques such as GraphRAG. However, they require significant integration effort. Adoption will depend on modernizing data architectures and governance frameworks.
Multimodal user interface (UI) is a high-level design model in which user and machine interactions can occur simultaneously via a combination of various interactions — spoken or written language and voice, brain and muscle control, motion and gesture, and gaze. Data can be processed from various data sources beyond text, including images, video, tables, maps, audio, gesture, motion, myoelectric, brain-computer interface and eye movement.
Potential for healthcare payers: Multimodal interfaces will elevate customer engagement by combining text, voice, images, and other inputs for richer, context-aware interactions. For payers, this means claims decisions can integrate text and imagery for faster, more accurate outcomes, while risk inspections can merge multiple media sources to improve pricing precision. Adoption will complement existing engagement tools like chatbots and virtual assistants, creating a seamless experience across multiple technologies.
Neuromorphic computing leverages semiconductor devices inspired by neurobiological architectures. Neuromorphic processors feature non-von Neumann architectures and implement spiking neural network execution models that are dramatically different from traditional processors. They are characterized by simple processing elements but very high interconnectivity, although lossless and near-lossless techniques are emerging.
Potential for healthcare payers: For payers, its relevance is limited to enabling smarter wearables and home devices for wellness programs and remote monitoring. While promising for future digital therapeutics, the lack of immediate business value makes near-term investment unlikely.
Neuro-symbolic AI denotes a composite AI system that integrates neural-network-based methods with symbolic-knowledge-based approaches. The synthesis of AI techniques capitalizes on the strengths of network-based and symbolic paradigms and reduces their respective limitations to provide AI systems capable of more advanced reasoning, learning and cognitive modeling. The neural component includes the use of statistical deep-learning techniques foundational to machine learning, while the symbolic element includes rule-based reasoning approaches, commonly utilized in disciplines such as logic and knowledge representation, as well as mathematics and programming languages.
Potential for healthcare payers: Neural networks are embryonic for the average payer, although startups like Cogniswitch AI are promising. Neuro-symbolic AI’s utility depends on mature adoption of knowledge graphs, which is still nascent. Until foundational technologies are in place, neuro-symbolic AI remains a long-horizon opportunity.
Private 5G. A private 5G network is based on 3GPP technology and spectrum to provide connectivity, optimized services and security for enterprises. It can be installed and managed by a CSP, a technology vendor, an SI or the end user for the express use of a single, unique entity. A 5G private mobile network (PMN) is used to interconnect people and also things in an enterprise, and it can consist of a mix of public and private infrastructure (for example, private slice over public network). Deployments can be hybrid with on-premises radio and local breakout and connections to the telco core, and/or a public cloud or fully on-premises.
Potential for healthcare payers: Private 5G networks offer secure, high-speed connectivity but have negligible value for traditional payers. They are more relevant for provider organizations managing clinical systems and physical assets. Unless a payer operates as a payvider, investment in private 5G is unnecessary.
Quantum Computing uses quantum processor units (qubits or quantum bits) to handle complex computational tasks. Unlike classical bits, qubits can exist in superposition, being both zero and one simultaneously until observed. Integrating quantum processors requires a complex hybrid ecosystem of technologies, including low temperatures, vacuum environments, radio frequency (RF) modulation and/or lasers, combined with high-performance computer systems to manage quantum elements. Qubits can be linked with other qubits, a property known as entanglement. Quantum-classical algorithms manipulate linked qubits in their entangled state, enabling future system designs that can potentially address a set of use cases that classical systems cannot handle. Quantum computers will be essential for simulating atomic and molecular interactions, critical for discoveries in materials and drug development.
Potential for healthcare payers: Quantum computing promises breakthroughs in areas such as cybersecurity and actuarial modeling but remains a long-horizon investment. Hardware costs and algorithm immaturity push ROI beyond 10 years. For now, classical computing combined with AI will deliver faster, cheaper value.
Smart spaces are physical or digital environments in which humans and technology-enabled systems interact in increasingly open, connected, coordinated and intelligent ecosystems. The design patterns to create smart spaces are referred to with various names, including “smart city,” “digital workspaces,” “smart venues” and “ambient intelligence.”
Potential for healthcare payers: For payers, use cases are minimal — limited to member-facing wellness programs or internal office environments. With most payers operating remotely or in hybrid models, and lacking retail or clinical footprints, smart spaces offer negligible near-term value.
Small language models (SLMs), also called “light language models” or “small model,” support use cases where traditional large language models (LLMs) are not feasible or not ideal. SLMs represent a trade-off between the generalized power of LLMs and the narrower requirements of resource-constrained environments, such as on-premises deployments, smartphones or edge network nodes. SLMs are much smaller than their LLM counterparts. Examples include Microsoft Research Orca 2 (13 billion parameters), MosaicML MPT (7 billion parameters), Stanford University Alpaca (7 billion parameters) and Stability AI’s Stable LM (7 billion parameters). The number of parameters that categorize a language model as small has changed over time in relation to mainstream LLMs. In addition, the dichotomy of small versus large has become more complex, with various sizes for different use cases. These include 3 billion parameters or smaller for edge devices, 7 billion parameters or smaller for smartphones, 70 billion parameters for laptops or equivalents, and 130 billion parameters or higher for on-premises servers. This spectrum of sizes and use cases is evolving rapidly. Most of the SLMs mentioned are open-source.
Potential for healthcare payers: SLMs deliver fast ROI with low compute requirements and flexible deployment (cloud, on-premises, edge). They are increasingly embedded in enterprise platforms like Genesys and Workday. Internal AI teams are leveraging SLMs for fine-tuning, making them highly adaptable for payer environments.
Spatial computing is a computing environment that combines physical and digital objects in a shared frame of reference. It involves spatial mapping and identification of people, places and things within the physical world as a foundation for anchoring digital content that intersects with the physical world’s spatially anchored, indexed and organized content.
Potential for healthcare payers: Spatial computing blends physical and digital environments, supporting use cases like remote monitoring and wellness nudges. For payers, integration with care management programs would be moderately complex and dependent on ecosystem partners. Privacy concerns and limited scope reduce its near-term value.
Synthetic data is a class of data that is artificially generated rather than obtained from direct observations of the real world. Data can be generated using different methods, such as statistically rigorous sampling from real data, semantic approaches, generative adversarial networks or by creating simulation scenarios where models and processes interact to create completely new datasets of events.
Potential for healthcare payers: Synthetic data enables model development without exposing protected health information (PHI), supporting fraud detection and engagement analytics. It can expand partnerships beyond entities willing to sign business associate agreements (BAAs). However, its effectiveness depends on clean source data, and bias introduction and proliferation risks remain.
Tokenization is a cryptographically secured representation of value or data. Blockchain technologies provide the ability to represent any asset, physical or virtual, as a token on a blockchain network. They allow you to create (or mint) tokens, assign unique ownership, prevent duplication, transfer or trade such assets, and remove (or burn) tokens from circulation. In addition, tokenization through blockchain offers the additional features of fungibility, programmability and fractionalization.
Potential for healthcare payers: Tokenization secures sensitive data by converting it into cryptographic tokens, reducing exposure risk. While theoretically valuable for privacy, implementation costs and disruption outweigh benefits for most payers. Use cases like tokenizing medical records remain speculative and unlikely to deliver meaningful ROI soon.
2025 Gartner Healthcare Payer Agentic AI Survey. The main objective of this survey was to learn how U.S. healthcare payers are investing in agentic AI technology. This survey was conducted online from 10 September through 29 September 2025. In total, 33 executives at U.S.-based U.S. healthcare payer organizations participated. All 33 participants are members of Gartner’s U.S. Healthcare Payer Research Panel, a Gartner-managed panel. Respondents were all located in the U.S. (n = 33). Disclaimer: The results of this survey do not represent global findings or the market as a whole, but reflect the sentiments of the respondents and companies surveyed.
2025 Gartner Business Outcomes of Technology Survey. This survey was conducted to understand how industries leverage technologies for various use cases. It assessed investment, deployment and implementation strategies for industry technologies. It also examined key areas intended to be impacted by technology investments, including challenges to realizing business outcomes and industry key performance indicators. The survey was conducted online from June through August 2025. The 648 respondents were from midsize, large and global enterprises from North America, EMEA and Asia/Pacific. The respondents were screened for senior IT and some business leadership roles with technology decision-making responsibilities. Disclaimer: The results of this survey do not represent global findings or the market as a whole, but reflect the sentiments of the respondents and companies surveyed.
2025 Gartner CIO and Technology Executive Survey. This survey tracked how senior IT leaders worldwide prioritize strategic business, technical and management objectives. It was conducted online from 1 May through 28 June 2024. The survey includes respondents who lead an IT function, with a total of 3,186 CIOs and technology executives participating. The survey participants are representative of various geographies, revenue bands and industry sectors, including both public and private organizations. Disclaimer: The results of the survey do not represent global findings or the market as a whole, but reflect the sentiments of the respondents and companies surveyed.