Logo
Become a Client

Category: Software

In the early 2020s, most IT dashboards looked deliciously green – until you cut them open. That “watermelon problem” summed up the gap between what SLAs said and how people actually felt at work: 99.8% uptime on paper, but slow logons, clunky multi-factor authentication, and chatbots that couldn’t understand what anyone really wanted. Experience was an afterthought, AI was a sideshow, and creativity was nowhere to be found in the contract.​

When SLAs ruled the world

Back then, three things defined the status quo. AI was narrow and local, sitting on the edge of workflows answering FAQs or routing tickets rather than orchestrating work. Experience measurement lagged reality, with annual or quarterly surveys surfacing issues long after the damage was done. And creativity simply didn’t exist in the metrics; contracts cared about uptime, not whether people had the cognitive space to experiment or innovate.​

The result was a strange split-screen. On one side, leaders proudly cited their SLA success. On the other, employees wrestled with friction that didn’t fit any KPI: context-switching between tools, re-entering the same data, and watching “helpful” chatbots miss the point. XLAs were occasionally piloted  (an NPS here, a satisfaction score there) but rarely changed actual design or investment decisions.​

Now: XLAs as control towers for human-AI work

Fast forward to 2026, and AI is no longer the sidekick; it is the backbone of digital work. GenAI assistants, low-code agents, and orchestration platforms now sit inside service desks, digital workplace platforms, and line-of-business apps. XLAs have emerged as the language that decides whether all this AI is genuinely helping humans do better work or just adding more noise.​

Three big shifts define the “now.” Agentic AI makes XLAs real-time and contextual, correlating technical signals like latency and crashes with human signals such as sentiment, task completion, and time to productivity. It can trigger automated remediation, from self-healing endpoints to conversational agents that guide users through fixes, and spotlight experience hotspots for specific personas or workflows. IDC’s 2025 Future of Work survey shows 79% of organizations now actively measure the relationship between employee and customer experience, with two-thirds having proof of causal linkages, while 94% of AI-enabled work adopters report productivity gains and over half see significant improvements.​

Making creativity a measurable outcome

The most interesting XLAs no longer treat creativity as a fuzzy aspiration. They track uninterrupted focus time per persona, link AI automation to freed-up hours, and measure innovation throughput:  ideas submitted, prototypes built, experiments completed. Instead of only asking if AI is fast or accurate, organizations track “human-plus” metrics: how much better decisions, proposals, and options become when humans and AI work together.​

Governance grows up

This evolution is forcing governance structures to grow up fast. AI-focused Centers of Excellence increasingly use XLA dashboards as strategic instruments, challenging deployments that look great on technical metrics but poor on human outcomes. They prioritize changes that build trust and agency, such as better explainability, robust feedback loops, and human override capabilities, and retire tools that consistently score badly on ease of use or learning curve.​

Metrics are diversifying accordingly: about 69% of organizations use productivity scores such as task-based speed and throughput to assess AI, while 42% also track employee satisfaction and 44% monitor skills proficiency. XLAs have become a proxy for hard questions: Are we making it easier for people to solve novel problems? Are AI tools empowering experts or boxing them in? Where is digital friction quietly killing initiative?​

Tomorrow: XLAs as the OS for co-creation

Looking ahead, XLAs are set to become the operating system for human/AI co-creation. Emerging “experience-risk” indices predict burnout or disengagement, while creativity capacity scores combine focus time, use of exploratory tools, and psychological safety indicators. Agentic AI will increasingly use XLAs as experience-intent parameters  – goals like maximizing focus time for data scientists or ensuring frontline staff resolve most issues in under three minutes  – and autonomously orchestrate tools, notifications, and workflows to hit them.​

Contracts will catch up too, moving from green dashboards to models that reward innovation, protect against “experience debt,” and explicitly safeguard time and cognitive bandwidth for meaningful work. For service providers, the mandate is clear: anchor XLAs on outcomes only humans can deliver, make creativity visible on the dashboard, build strong feedback loops, and use XLAs as guardrails against over-automation. XLAs are no longer just a friendlier way to measure IT; they are becoming the central platform for keeping human potential at the center of an AI-driven future of work.

For more information see IDCs upcoming research documents: “Measuring What Matters: XLAs and the 2026 Digital Workplace” and “Control Towers for Human Potential: The Growing Importance of XLAs in the Age of Agentic AI”.

If you have a question about this or any other IDC research, drop it in here.

Meike Escherich - Associate Research Director, European Future of Work - IDC

Analyst Profile
Meike Escherich is an associate research director with IDC's European Future of Work practice, based in the UK. In this role, she provides coverage of key technology trends across the Future of Work, specializing in how to enable and foster teamwork in a flexible work environment. Her research looks at how technologies influence workers' skills and behaviors, organizational culture, worker experience and how the workspace itself is enabling the future enterprise.

By Bo Lykkegaard, Associate VP for Software Research Europe with advice and review by Ewa Zborowska, Research Director, AI, Europe

Providers of SaaS solutions across the world have been through the market capitalization bloodbath during the past six months. Despite presenting solid indicators of growth and margins for 2025, almost all publicly traded companies have seen share price reductions 10% to 60% with the average reduction being in the 30-35% range.

Forget about looming trade wars, recession fears, missed revenue goals, and other conventional share price depressants. This is about AI disruption of the current SaaS user experience, licensing model, and product architecture. Investors are starting to fear that the SaaS ‘rental model for software’ will become invisible ‘featureware’ inside an AI agent layer.

What Are the Market Cap Reductions Telling Us?

We have examined the market cap reductions of public traded SaaS vendors over the past six months. Based upon this, we can make the following observations:

  • All SaaS vendors are affected across solution areas, geographies, size of vendor, recent growth KPIs, and size focus (SMB vs. enterprise). This means that investors are reexamining their assumptions related to SaaS growth prospects in general.
  • Vendors of workflow automation solutions and vendors targeting small and medium-sized businesses appear particularly exposed. Commercial workflow software is seen as exposed to replacement by new AI agent technologies. Also, vendors targeting small businesses are seen as more exposed to churn and price pressures.
  • SaaS vendors headquartered in EMEA do not appear harder hit than those headquartered in North America and the market cap correction has hit the largest as well as the smaller SaaS vendors.

Changes that All SaaS Vendors Are Facing

Firstly, the conventional SaaS user experience must change. In a conventional SaaS application, the user executes tasks manually within defined workflows. In an AI-powered application, the system adds to these structured workflows with probabilistic outputs, where it generates, predicts, recommends, or executes. Also, AI-powered applications can accept and react to all kinds of conversational user inputs. Furthermore, just like today’s LLM-based apps, business applications understand context and remember past interactions, which make recommendations and predictions more relevant and precise. Finally, AI-powered business applications are more proactive in nature and help users with monitoring tasks and relevant notifications.

Secondly, the conventional SaaS licensing model must evolve. The talk of the town these days is ‘outcome-based pricing’, i.e. the notion of pricing an application on outcomes (e.g. number of invoices issued) as opposed to number of users. If agentic workflows increasingly automate core business processes in the future, the user of a, say, financial application will be an agentic workflow as opposed to a human user. As AI agents increasingly become users of business applications, the user-based revenue model of SaaS application collapses. Investors are looking for SaaS vendors to at least align licensing better to business outcomes.

Thirdly, the conversional SaaS product architecture must be rethought. Adding AI to a conventional SaaS solution in the form of a chatbot or other form of AI-generated add-on does not make a meaningful difference. Real modernization requires rethinking the SaaS workflow from the ground up. AI changes all levels of the SaaS product stack and needs foundation model(s), embedding layer, vector database, retrieval-augmented generation (RAG), orchestration layer, guardrails, monitoring, and prompt/version management.

AI is making several other significant changes in SaaS. Development and maintenance as well as running costs have become more volatile and unpredictable. Data management requires new approaches, as application data now serves as a key source for training AI-powered SaaS solutions. Product roadmaps and release cadences are increasingly driven by AI model upgrades rather than traditional update schedules. Software vendors face new risk management challenges related to hallucinations and regulatory compliance. And both vendors and end-user organizations need to adapt their teams with new sets of skills. And most importantly, the overall competitive landscape has shifted, with AI-based startups and hyperscaler offerings emerging as new challengers.

The changes above certainly apply to SaaS vendors in Europe. However, in addition, vendors in Europe – as they adapt solutions and business models to become AI-driven – must pay particular attention to four areas in order to successfully transform.

Firstly, there is the GDPR, NIS2 and EU AI Act compliance, often accompanied by various national or industry-specific regulations. If they cannot document and showcase complete compliance to customers, they cannot sell their AI-powered solutions to compliance-sensitive European organizations.

Secondly, increasingly we see data residency requirements from customers in Europe, particularly in public services, financial services, and healthcare. Buyers in such industries can require EU-hosted data and sovereign cloud guarantees and approaches and can seek to avoid subjection to the US CLOUD Act and to exposing data for foundation model training.

Thirdly, Europe is multi-lingual and buyers require multi-language model performance. A conversational SaaS application is great but only if the conversation happens in the local European language where the application is deployed. We have seen many cases where non-English conversational capabilities are years behind English.

Fourth, European AI-powered SaaS vendors should expect higher demand for transparency and explainability. European customers have a strong preference for understanding how AI systems make decisions, a need often reinforced by regulations like GDPR and the EU AI Act. This means vendors must provide clear logic behind decision criteria, bias mitigation documentation, human oversight mechanisms, and comprehensive audit trails. Black box AI approaches such as “Pick this candidate because the recruiting application assigned a high AI score” simply will not fly in Europe, where trust is key and it heavily depends on being able to trace and justify how conclusions are reached.

Join the Conversation

At IDC, we help you navigate these changes with deep market research, robust data analytics, and tailored custom solutions. Whether you need strategic insights, benchmarking, or support in adapting your business model, our experts are ready to guide you.

Contact us to discuss your unique challenges and discover how IDC can empower your next steps in the evolving, AI-disrupted European software landscape.

Sources:

Bo Lykkegaard - Associate VP for Software Research Europe - IDC

Analyst Profile
Bo Lykkegaard is associate vice president for the enterprise-software-related expertise centers in Europe. His team focuses on the $172 billion European software market, specifically on business applications, customer experience, business analytics, and artificial intelligence. Specific research areas include market analysis, competitive analysis, end-user case studies and surveys, thought leadership, and custom market models.

过去一年，生成式AI迅速从“前沿技术”演变为企业讨论中的常规议题。从董事会到业务一线，关注点已经不再是“要不要用AI”，而是企业在不同发展阶段，应该如何选择落地路径、如何判断投入节奏，以及如何尽量降低不必要的试错成本。

从市场实践来看，企业的AI探索并不存在统一范式：有的企业从具体应用场景切入，有的优先推动流程自动化，也有企业选择先夯实数据和平台基础。这些选择背后，往往与行业属性、组织能力、数字化成熟度和管理目标密切相关，并不存在绝对正确的先后顺序。

在这一过程中，企业级应用、企业级服务以及数据库与数据管理，往往以不同形式、不同权重出现在企业的AI实践中。IDC开展相关研究，并非试图将这些因素“硬性绑定”为成功前提，而是希望更真实地反映市场的复杂性，帮助企业理解不同路径下可能面临的机会与约束。

从“加AI功能”到“用AI做事”：企业级应用的重构正在发生

AI Agent正在改变企业应用的基本形态

在很多企业中，生成式AI最初的落地方式是“功能叠加”：写文案、生成报表、自动摘要。这类能力提升了效率，但并没有改变应用的本质。

IDC的研究发现，真正具有颠覆意义的变化来自AI Agent（智能体）的引入。企业级应用正在经历从“被动工具”到“主动参与业务执行”的转变：

  • 应用不再只是被人操作，而是能够理解目标、拆解任务并自动执行
  • 用户界面逐渐从复杂菜单，转向自然语言和流程驱动
  • 应用之间开始通过Agent进行协同，而非人工串联

IDC将这一变化总结为企业级应用的“Agentic演进路径”，并指出未来几年内，Agent将从辅助角色逐步走向主导角色。

哪些业务场景最先受益？

从企业实际落地情况来看，生成式AI和智能体的应用并未集中在单一部门，而是优先出现在高频交互、高度标准化或知识密集型的业务与技术场景中，包括：

  • 客户服务与智能联络中心：AI被广泛用于自动应答、坐席辅助、工单分流与服务质量监控，在不完全替代人工的前提下，提高响应效率和服务一致性。
  • 办公自动化与知识管理：会议纪要、文档整理、企业知识问答等场景逐步由AI承担基础工作，降低员工获取信息和跨部门协作的成本。
  • 内容生成与市场营销：从内容和素材生成，延伸至客户洞察、活动优化和线索管理，营销决策开始更多依赖数据与模型驱动。
  • 职能流程自动化：在财务、供应链、HR、采购、法务等职能领域，AI被用于规则明确、重复性高的流程自动化、合规检查和风险识别。
  • 研发与IT运维：代码生成、测试、故障定位和运维自动化成为AI落地的重要方向，直接影响研发效率、系统稳定性和运维成本。

IDC之所以持续追踪这些细分场景，是因为企业在做AI投资决策时，往往需要回答一个现实问题：哪些应用场景已经具备规模化条件，哪些仍处在早期探索阶段。这类研究的价值，在于帮助企业避免“平均用力”，而是将有限资源投入到最有可能产生业务回报的方向。

没有服务能力，AI很难真正跑起来

一个在客户中反复出现的共识是：AI Agent的成败，不仅仅是模型本身，很大程度还取决于项目实施过程中对于数据治理，安全合规，流程重塑，平台整合等环节的设计和把控，以及后期的运营和维护。

企业在推进过程中普遍会遇到：

  • 业务流程是否适合被Agent接管
  • 多个Agent如何协同、治理和监控
  • 如何持续评估ROI，而不是一次性交付

这也是为什么企业级服务在AI时代的重要性被显著放大。IDC在软件与服务研究中，将AI咨询、Agent设计与开发、系统集成、运维与持续优化视为一个完整闭环，而非单一项目 。

对企业而言，这类研究的价值并不仅在于“推荐某一家供应商”，而在于帮助管理层理解能力建设的先后顺序：哪些能力需要长期内生，哪些可以借助生态伙伴补齐，从而避免“试点成功、规模失败”的常见陷阱。

AI走得多远，取决于数据和数据库走得多稳

数据库正在从“后台系统”走向“AI基础设施”

如果说企业级应用决定了AI“做什么”，那么数据库和数据管理决定的则是AI“能不能做、做得好不好”。

在生成式AI快速演进的同时，中国数据库市场也正在经历深刻变化：一方面，AI对数据实时性、多模态和向量能力提出更高要求；另一方面，国产化进程推动本土数据库厂商在功能和市场份额上持续提升 。

AI for Data：让数据库更智能

IDC在数据库研究中发现，AI正在反向赋能数据库自身：

  • 自动调优与容量预测
  • 基于AI的异常检测和安全防护
  • 更智能的运维和资源调度

这些能力直接降低了数据库复杂度，使企业能够用更少的人力支撑更复杂的业务和AI负载。

Data for AI：让AI真正可用

更关键的是，数据库正在成为AI应用的“能力上限”：

  • 向量引擎和多模数据管理决定了Agent是否具备“长期记忆”和上下文理解能力
  • 数据治理和权限体系决定了AI是否可信、可控
  • 实时数据能力决定了AI是否能够参与业务决策，而不仅是事后分析

IDC在数据库管理系统市场的研究中强调：未来企业AI竞争的本质，是数据架构和数据能力的竞争。

为什么IDC要持续开展这些研究？

IDC之所以持续在企业级应用、企业级服务以及数据库与数据管理领域投入研究，一方面，这些领域是企业AI价值真正发生的位置：应用决定AI是否进入业务流程，服务决定AI能否规模化运行，数据库管理和数据治理决定AI是否长期可持续。任何一环缺失，AI都很难从“亮点项目”走向“稳定能力”。

从企业决策者视角看，这些研究真正解决了什么问题？

通过持续的市场数据、趋势判断和实践洞察，IDC希望帮助客户：

  • 看清AI技术和应用的成熟节奏
  • 了解行业发展的最新趋势和最佳案例
  • 对于热点领域和技术的评估和参考实践

在生成式AI引领的新一轮技术升级中，真正具备长期优势的企业，往往不是最早“尝鲜”的企业，而是那些能够构建高质量数据资产、完善AI治理体系、深度重塑业务流程并持续融合行业Know-how，实现数据驱动的敏捷创新与可持续落地的企业。这正是IDC持续开展相关研究的出发点，也是客户能够从这些研究中获得的长期价值。

IDC 2026年软件和服务领域研究计划：

如需进一步了解与研究相关内容或咨询 IDC其他相关研究，请点击此处与我们联系。

Lizzie Li - Associate Research Director - IDC

Lizzie Li is Associate Research Director of IDC China's Enterprise System and Software Research that focuses on research and analysis of the China Datacenter, Cloud Computing, and IT infrastructure markets. She also provides intelligence and consulting services in customized projects for local and multinational corporation (MNCs) IT vendors. Lizzie’s research domain covers Datacenters, Cloud Computing, Virtualization, and her duties include providing consulting proposals to IT vendors on sales, marketing, and research fields. Lizzie Li has seven years of experience in the IT industry, including Internet datacenters, cloud computing services, mobile telecommunication systems, and enterprise markets. Prior to joining IDC, Lizzie Li worked for 21vianet, Nokia Siemens Networks, and Huawei, and was responsible for sales analysis, project management, and technical support. Lizzie graduated from Huazhong University of Science and Technology with a Master’s degree in Pattern Recognition and Intelligent Systems.

In December 2024, one year ago, Microsoft CEO Satya Nadella declared on the BG2 podcast that “SaaS is dead.” The comment set off a shockwave across the technology industry and many felt provoked. After all, software-as-a-service (SaaS) has defined enterprise computing for nearly two decades, representing a massive share (over 10% according IDC’s Black Book) of IT spending in 2024 and forming the backbone of digital transformation strategies worldwide.

Yet, when we cast a cold IDC analytical eye beyond the provocative statement, a crucial truth emerges: SaaS, as we know it, is being disrupted, not by decline but by evolution.

The Status Quo: SaaS at Its Peak

Today, most of the world’s leading software vendors are, in some form, SaaS companies. Among the ten most valuable software players, including Microsoft, Salesforce, Oracle, SAP, and Shopify, SaaS delivery models dominate. Enterprises have grown dependent on the SaaS ecosystem, licensing countless applications to manage HR, payroll, CRM, expenses, and vertical industry workflows.

However, the sheer sprawl of SaaS adoption has created complexity for business users. Employees navigate dozens of interfaces daily, shifting context between multiple systems that rarely communicate smoothly. Despite efforts to simplify workflows through integrations and APIs, SaaS remains a patchwork of interfaces and data silos, forcing users to adapt to the software rather than the other way around.

The Complexity Problem and the AI Opportunity

This complexity is the Achilles’ heel of the SaaS model. Each SaaS application demands its own learning curve and user interface, often used sporadically and inefficiently. In this environment, AI offers a compelling remedy.

Instead of navigating multiple dashboards, users could interact with agent-driven, conversational interfaces that perform tasks across systems. Imagine instructing an AI agent to “approve last week’s expense reports” or “generate next quarter’s sales forecast” and having the agent orchestrate workflows across HR, finance, and CRM systems behind the scenes.

This agentic, “flow-of-work” user experience could replace much of today’s direct interaction with SaaS applications. The result? AI as the new interface layer, which is one that abstracts away complexity, automates repetitive processes, and redefines how enterprises consume software.

The Disruption: From Seats to Outcomes

Such a shift has profound implications for how SaaS is bought and sold. The traditional per-user, per-month licensing model becomes increasingly obsolete as digital labor replaces manual interaction. IDC predicts that by 2028, pure seat-based pricing will be obsolete, with 70% of software vendors refactoring their pricing strategies around new value metrics, such as consumption, outcomes, or organizational capability (please see IDC FutureScape: Worldwide Agentic Artificial Intelligence 2026 Predictions, IDC #US53860925, October 2025).

This agentic IT disruption will impact IDC’s existing forecasts for the various levels in the IT stack differently as shown below. Also, the impact will change over time, as for examples SaaS Applications and IT Services will feel a negative impact in the short term, while recovering if we look five years out to 2030.

For infrastructure hardware, IDC sees a different impact with a short term boost, followed by headwinds as inference costs drop exponentially.

Source: Charting the Agentic Future: 10 Vision Statements for 2030 (IDC #US53909225, November 2025)

Inside the enterprises, this evolution changes the economics of enterprise software. Companies optimizing AI agent development to reduce licensing costs will need to revisit their roadmaps as vendors adjust to these emerging pricing paradigms. Meanwhile, process owners may gain more flexibility, designing application-neutral operational efficiencies that transcend the limitations of current SaaS systems.

Business and IT Implications

The rise of AI agents doesn’t just alter pricing, it transforms how technology functions within organizations.

From a business perspective, enterprises may initially lose the tactical benefit of reduced software costs but gain strategic control over innovation and process optimization. Process teams will design workflows around end-to-end outcomes rather than application silos, supported by a new breed of “headless” software modules accessible via APIs and marketplaces.

From an IT standpoint, this means a fundamental re-architecture of the enterprise tech stack. Where today’s stack is built around SaaS interfaces, tomorrow’s will revolve around AI agents that interact with modular backend services. Data lakes and live data connections become critical enablers, while vendor relationships evolve from UI-centric engagement to agentic enablement partnerships.

Guidance for Technology Buyers

For IT and procurement leaders, this transformation demands foresight and experimentation. Buyers should assume that software vendors will increasingly position their offerings to accommodate or counteract the impact of digital labor.

Before adopting agentic systems, IDC advises enterprises to:

  • Build proofs of concept (POCs) and define clear ROI metrics around cycle time, productivity, and revenue improvements.
  • Evaluate end-to-end process efficiency, not just individual task automation.
  • Explore packaged AI agents offered by existing SaaS vendors, integrating them as part of broader operational redesigns.

In other words, the transition to AI-driven enterprise software should be intentional, data-backed, and aligned with measurable business outcomes.

The Road to 2030: SaaS Reimagined

By the end of this decade, the enterprise technology landscape will look radically different. The AI agent will become a new enterprise SKU, purchased via marketplaces and powered by modular backend capabilities rather than monolithic SaaS platforms. User interfaces will still be critical to productivity but so will orchestration of more-or-less autonomous workflows.

SaaS is not dead, but it is metamorphosing. The software industry is entering a new chapter defined by AI, automation, and outcome-based economics. For vendors, it’s a challenge to reinvent their business models. For buyers, it’s an invitation to rethink how software delivers value.

Either way, the next generation of enterprise technology will be less about screens and more about agents.

Got a question? Drop it in here.

You may be interested in listening to IDC EMEA’s predictions for 2026 and beyond.

Bo Lykkegaard - Associate VP for Software Research Europe - IDC

Analyst Profile
Bo Lykkegaard is associate vice president for the enterprise-software-related expertise centers in Europe. His team focuses on the $172 billion European software market, specifically on business applications, customer experience, business analytics, and artificial intelligence. Specific research areas include market analysis, competitive analysis, end-user case studies and surveys, thought leadership, and custom market models.