Partner POV | The State of AI at Work in 2025

Article written and provided by, Glean. 

Executive summary

The adoption of enterprise AI tools is growing at an unprecedented pace, with generative AI (GenAI) becoming a keystone for growth and workplace innovation. According to a recent ISG survey across G2000 companies in North America and Europe, over 65% of companies are now leveraging GenAI, with the number set to increase to 80% by 2026. However, the rise in the use of AI presents a critical decision point for organizations: whether to adopt layers of multiple disconnected AI tools or centralize AI capabilities within a unified platform.

This report explores the current landscape of enterprise AI adoption, examining how fragmented implementations create an "AI tax": an accumulation of costs, maintenance issues, and inefficiencies from managing multiple, unintegrated solutions due to  redundancy, complexity, and escalating costs. It further examines how centralizing AI into unified platforms offers a more sustainable, collaborative, and cost-effective path forward for enterprises striving to leverage AI at scale.

AI adoption trends in the enterprise

Employee-facing AI tooling is top of mind when it comes to future workplace innovation. Enterprises across the board are looking to spend considerably more on AI, while exploring in which workflows and applications AI applies best.

Enterprise spending on generative AI is projected to grow by 50% next year, with the number of AI-enabled applications increasing by 140% (ISG)—reflecting a growing commitment to leveraging AI.

However, only 6% of surveyed companies have yet to see AI provide even 75% of expected ROI in business growth, and only 39% expect to see true cost savings before H2 2025. (ISG)  

This "AI bloat" muddles insight into the ROI of AI, particularly its impact on company performance. Without a clear strategy to cut through this bloat, enterprises risk overspending on AI without gaining a tangible return—incurring a hidden tax that may cost more than they expect.

The hidden AI tax: Complexity and fragmentation

At this experimental stage of AI adoption, enterprises are deploying numerous AI tools across different departments to keep up with the pace of growing AI capabilities and applications. However, many organizations are facing challenges due to fragmented AI implementations, resulting in the hidden AI tax.

• Redundancy and cost: The average enterprise has 151 GenAI-enabled applications (ISG) across different use cases. Consider a customer service department using one tool for automated email responses, while the marketing department uses another for content generation, and the HR department uses yet another tool for employee onboarding assistance. 

Each of these tools may have overlapping functionalities—such as natural language generation—and yet, due to the fragmented approach, best practices and efficiencies are not shared across departments. This redundancy forces IT teams to manage disparate solutions, making it harder to align on AI usage while increasing costs related to licensing, support, and training. These factors contribute to organizations spending millions on AI tools, but not seeing a return on investment until years later.
• Licensing and user segmentation challenges: The cost of managing AI is compounded by licensing issues. Many AI-enabled tools offer limited user segmentation, forcing companies to either overspend on features for the entire organization or purchase more seats than necessary. For instance, Slack AI may require companies to pay for the generative AI feature across the entire organization, while Microsoft Copilot might require a large minimum volume. This lack of flexibility leads to inefficient resource allocation, as organizations end up paying for features that only a few departments might use effectively.
• Need for dedicated AI implementation teams: Fragmented AI tools can lead to the need for dedicated AI implementation teams, diverting resources away from innovation in a challenging macroeconomic environment. For example, managing disparate solutions often requires dedicated project managers or AI leads who focus solely on tool integration and troubleshooting rather than innovation. The more fragmented the tools, the larger and more specialized the support teams must be, consuming both time and budget.
• Integration and data silos: AI integration poses significant logistical and process-related challenges for enterprises. Imagine a sales team using a different AI tool to predict customer needs than the one used by the customer support team. When data is generated and stored in siloed applications, it becomes challenging to integrate insights across the customer journey. Over 95% of IT leaders report that integration issues hinder AI adoption (ISG), while data silos remain a significant barrier to digital transformation.These challenges prevent organizations from fully realizing the productivity benefits of AI.

Centralized AI platforms: A path to efficiency

To overcome these challenges, enterprises are increasingly turning to centralized AI platforms. These platforms offer a unified interface for managing AI initiatives, making it easier to scale AI across departments and functions.

• Unified management interface for AI initiatives: Centralized platforms provide a cohesive approach to scaling AI, enabling organizations to manage governance and security in a single interface. For instance, a unified security dashboard can provide visibility into all data sources that the central platform has access to, helping eliminate unapproved connections and ensuring that all departmental use cases are linked with appropriate permissions.
• Faster AI literacy: By adopting a centralized platform, companies can streamline AI training and speed up AI literacy across the entire business. Rather than training each department on a different specialized tool (or worse, training users on multiple siloed tools), the same set of building blocks  can be used to design every use case—helping everyone speak the same language when using AI at work. Additionally, a centralized learning process means AI literacy can be standardized across the entire organization, ensuring all employees, regardless of department, receive consistent and high-quality AI education. The industry is changing rapidly enough that the preferred set of tools for a given task can vary from month to month. A central platform that provides interoperability ensures both that these tools can speak to each other, and that employees aren't forced to relearn  everything each time workflows and processes change during this era of rapid experimentation. This facilitates continuous learning as AI technologies evolve, ultimately helping build a comfortable and proficient workforce with AI tools, regardless of their specific role.
• Improved collaboration and innovation: Integrated knowledge repositories and shared data libraries help align AI technology with business outcomes. For example, let's say a marketing team uses an AI content creation tool to build a fine-tuned customer persona model. A product team could find this through a shared resource center, and use it when writing technical documentation. 
This cross-pollination is difficult when tools are fragmented, as knowledge remains confined to silos. AI's need for real- time, relevant data means the data lakes of the enterprise SaaS era must evolve into more complex knowledge graphs to align AI with business objectives.
• Reduce costs in an expensive AI ecosystem: AI tools remain costly, even as API prices continue to drop. Centralized platforms help reduce these high costs across multiple departments, ensuring that AI investments deliver maximum returns without disproportionately burdening specific projects. For instance, a single centralized AI infrastructure could support various functionalities—customer service chatbots, content creation, and analytics—ensuring that each department benefits without individually shouldering the full cost of the AI capability.
• Revolution, not evolution: Rather than "upgrading" one application at a time while paying the AI tax, a centralized platform brings AI-powered capabilities to every content type and data source in a corporate knowledge corpus. No longer dependent on a third party to innovate, a company can instead move at its own pace along the AI maturity curve without having to reinvent the wheel.

Balancing grassroots innovation with centralized AI strategy

AI-centric employees create AI-centric companies, and establishing this culture now is crucial. Early adoption provides a competitive edge that compounds—high-quality data feeds productive models, which generate more efficient workflows and outputs that, in turn, improve the standard of data used in the next generation of models. Delaying this transition can put companies well behind the curve. As other organizations build AI fluency and capture resulting market opportunities, catching up can be considerably difficult. Achieving this speed requires synchronized AI literacy across all departments, which takes time, effort, and resources to develop.

Step 1: Provide access to tools

The foundation of an AI-centric organization is providing access to tools that meet governance and security standards. By ensuring all employees have access to sanctioned AI tools, companies can create a baseline of AI literacy that empowers every department. 



This is crucial to avoiding the risks of shadow IT; unsanctioned tools used by employees can lead to data breaches, inefficiencies, and, in the case of highly regulated industries, compliance issues. By standardizing tool usage, organizations can limit exposing their data to insecure tools and ensure consistent AI adoption across all departments.

Step 2: Enablement through training

The next step is enablement, where employees learn prompt  engineering and best practices. They learn where and how to be critical of AI usage, how to identify biases in AI models, and how to ensure AI outputs align with business objectives. This is also where we see the power of synchronized AI literacy across departments.

Consider a procurement process: speeding up just the legal review of contracts isn't enough; every step, from vendor identification to cybersecurity checks to onboarding, must be enhanced to achieve true productivity gains. When companies purchase piecemeal products for different departments, they risk creating new bottlenecks while removing old ones. Centralized platforms ensure that every department has the appropriate speed and agility to keep the organization running smoothly.

Acknowledging the value of grassroots experimentation here is essential. Early AI adopters and experimenters within organizations often drive innovation by identifying impactful use cases. These early efforts are vital for discovering what works and where AI can add the most value. Organizations should  encourage and enable safe grassroots experimentation, recognizing that these pioneers pave the way for future enterprise-level AI strategies.

To effectively harness these grassroots efforts, companies need a system to collect, evaluate, and document successful AI use cases discovered at the edges of the organization. Centralized tools (like Glean's Prompt Library and Apps) that integrate grassroots learnings help transition isolated successes into standardized best practices. By packaging these practices and making them accessible, organizations can empower a broader employee base to benefit from them. This way, individual successes can be scaled enterprise-wide, maximizing impact.

Step 3: Integrate AI into production

Up to this point, employees are still primarily consumers of AI—often using tools or processes developed by that small team of early adopters. However, to create truly AI-centric companies, AI must touch all people, processes, and products to drive a transformative impact.

For people, this means encouraging employees at all levels (from interns to C-suite) to constantly experiment with new ways of working, fostering a culture of innovation. For processes, it means overhauling internal workflows to ensure they are AI-native—representing a revolution, not just an evolution. For products, companies must embed AI capabilities directly into their offerings and allow their business to be flexible in turn, whether this manifests as outcome-based pricing models or through leveraging Glean's contextual knowledge graph via APIs, making AI a fundamental part of their customer value proposition.

Future-proof AI integration through centralized platforms

Harnessing grassroots innovation with a unified AI strategy ignites scalable and sustainable change. By embedding AI into the core of the organization's culture, enterprises can unlock sustainable growth, resilience, and lasting success while cutting inefficiencies and complexity. Centralized platforms provide the infrastructure needed to reduce costs, foster collaboration, and spark innovation. A unified AI approach guarantees consistent impact, enables scalable training, and builds an inclusive, future-ready ecosystem. With centralized AI, organizations are empowered to thrive, adapt, and lead in our new AI-driven world, where grassroots innovation takes flight from an adaptable and future-proof foundation.

The time to act on enterprise AI strategy is now. As organizations grapple with mounting AI costs and fragmented tools, the path forward lies in harmonizing grassroots innovation with a unified AI strategy. By thoughtfully embedding AI into the organization's core through centralized platforms, enterprises can cut through the complexity and inefficiencies of scattered solutions and unleash their full innovative potential.

This unified approach does more than just reduce costs—it creates a foundation where ideas can flourish, departments can collaborate seamlessly, and employees at every level feel empowered to drive transformation. The companies that take decisive action today to consolidate their AI capabilities will build the resilience and agility needed to thrive in an AI-driven future, turning the challenge of AI fragmentation into an opportunity for sustainable growth and market leadership. With centralized AI as their cornerstone, organizations can confidently navigate the evolving technological landscape, knowing they're building not just for today's needs, but for tomorrow's possibilities.

Looking to advance your own organization's AI adoption and acceleration in order to unlock the full capabilities of AI? Glean's Work AI platform provides next-generation generative AI capabilities that help you and your teams easily accelerate and automate day-to-day workflows.