Guest Post: Marlabs

AI is already getting boring, and organizations are better off for it.

The novelty of AI and the initial hype have worn off. Early on, organizations chased flashy, high-profile AI projects, but there was a huge problem: widespread failure. 

MIT’s August 2025 study found that only 5% of organizations were getting significant value from generative AI, while 95% failed to achieve ROI or move beyond pilots and experiments.

Now the market is sobering up. Leaders are moving to a more pragmatic phase and taking time to figure out what actually works. 

It turns out that the successful 5% focused their AI efforts on more boring, practical applications, such as automating routine work and improving how work gets done.

Today, leaders are prioritizing “boring” AI, the silent background automation of time-consuming, high-volume tasks. Instead of a separate AI app, intelligence is embedded directly into existing workflows and systems.

Organizations taking the boring route see that AI is not just delivering results but simultaneously transforming enterprises by improving operations from the inside out. 

Two areas where boring AI creates measurable impact are agentic automation and knowledge management. 

Though overlooked during the initial AI frenzy, these approaches are delivering the most reliable efficiency gains and ROI because they are more attainable and cost-effective, are easier to deploy, and yield faster returns than moonshot ideas. 

If you’re disappointed with your AI efforts so far, it might be time to regroup and rethink how to move forward. Boring AI provides a safer, more practical path with proven success. Let’s explore why. 

What Is Boring AI?

Boring AI refers to practical, non-glamorous uses of AI focused on execution. These are not moonshot projects or experimental pilots created to impress. They quietly improve how work gets done. 

Boring AI emphasizes infrastructure over interface. AI is embedded into existing systems and workflows, allowing people to keep working the way they already do without having to learn new systems and processes. 

These typically lower-investment initiatives are smaller in scope and deliver value faster than hype-driven projects. 

They also pose less risk. Boring AI relies on well-understood data and repeatable tasks, avoiding many of the ethical and organizational concerns that surround more ambitious AI promises. 

Plus, boring AI does not replace people. It makes them more effective at what they already do. 

With wins like these, you can see why leaders are shifting AI efforts toward boring projects. 

So, let’s dive into where AI consistently delivers value: agentic automation and knowledge management.

Agentic Automation: Where Boring AI Proves Its Value  

Agentic automation is a form of AI-driven automation where software autonomously determines what needs to happen to achieve a goal and takes actions accordingly. 

While traditional automation is rule-based and follows fixed processes and predefined steps (executing exactly as programmed), agentic automation is goal-driven and adaptive. 

It is built to handle routine decisions and repetitive actions at scale, and it augments human judgment rather than replacing it.

Traditional automation has been around for a long time, but its rigid nature has limited how much it could do. Apply AI, and it’s a whole different story. 

What Agentic Automation Does

Agentic automation uses AI systems to autonomously execute routine, multi-step tasks within defined guardrails, reducing operational costs, improving speed and accuracy, and enabling efficient scaling — all without requiring sweeping organizational change. 

Agents within an agentic automation system assess data, understand context, learn, reason through options, ask questions, and continually optimize and adapt to changes in conditions or objectives. 

They execute tasks to achieve defined outcomes with minimal human intervention. 

Each agent has a narrow, specific scope and handles very specific jobs. One might classify requests, while another routes them, another generates a response, and another triggers follow-up actions. 

The system orchestrates the agents to achieve shared goals, much like a conductor guides musicians in an orchestra. 

The individual agents derive their intelligence from the system, and this coordinated approach is far more effective than relying on a single, generalized agent attempting to handle everything. 

What Technologies Drive Agentic Automation?

The technologies that fuel agentic automation include large language models (LLMs), large action models (LAMs), generative AI, robotic process automation (RPA), machine learning, predictive analytics, and others.

Working together, they enable agentic automation systems to plan, act, and adapt across complex workflows.

Agentic Automation in Action

Common use cases for agentic automation include:

• Document processing for loans, real estate transactions, and legal briefings
• Customer support, like automated ticket routing, response generation, resolution suggestions, and follow-ups
• Marketing applications, such as automated lead scoring, campaign prioritization, and personalized content delivery 
• Finance workflows, including invoice processing, expense validation, and fraud detection
• Operations and supply chain applications, like predictive order routing, inventory adjustments, and exception handling

These tasks already follow defined rules and occur in high volume, which are ideal conditions for agentic automation. 

The Impact of Agentic Automation 

AI-driven agentic automation is already delivering meaningful operational benefits. 

Gartner projects that by 2029, agentic AI will autonomously resolve up to 80% of the most common customer service issues, potentially reducing operational costs by about 30%.

Organizations adopting agentic automation report fewer errors, shorter cycle times, and less manual rework, outcomes that contribute to improved efficiency and lower costs.

Another Gartner survey found that 54% of infrastructure and operations leaders said reducing costs is their top goal for adopting AI, reflecting a broader shift toward operational, cost-focused use cases.

Because agentic automation builds on workflows and data that already exist, AI accelerates what businesses are already doing without requiring them to reinvent how work gets done. 

Such use cases are easier to justify because they build on existing workflows, which means they carry less implementation risk and deliver measurable returns faster than more experimental AI initiatives. 

Embedded AI also plays a critical role in achieving a better ROI. Instead of deploying a separate AI application, intelligence is embedded directly into existing enterprise systems, such as ERPs and CRMs, improving efficiency within familiar workflows.

As a result, employees remain in their roles without the disruption of sweeping organizational change. They are supported, not displaced, which reduces resistance and allows agentic automation to scale more smoothly.

In the end, agentic automation works almost invisibly to deliver measurable impact without drawing attention to the AI itself — the essence of boring AI.

Knowledge Management: Making Information Work for People

While agentic automation streamlines tasks and execution behind the scenes, knowledge management tackles a different bottleneck: access to information. 

Knowledge management is hands down one of the most effective uses of AI in an organization or field of study. It ensures that employees spend less time searching for information and more time applying it to drive results. 

The Problem That Knowledge Management Solves

For decades, employees have lost hours searching for information, looking for the right person to talk to, recreating work, or unknowingly relying on outdated knowledge.

According to IDC, knowledge workers spend nearly 30% of their time searching for information. This inefficiency compounds across teams. 

When workers act on incomplete or incorrect information, such liabilities lead to errors, poor decisions, lost customers, lost revenue, and damage to the brand.

What Knowledge Management Does

AI-powered knowledge management organizes, retrieves, and contextualizes information inside the systems employees already use. 

It connects documents, policies, tickets, and institutional knowledge so that users have access to reliable information whenever they need it. 

The fact that AI captures institutional knowledge from documents, retrospectives, and conversations across the organization’s communication channels is groundbreaking. 

Before AI, knowledge management systems struggled to stay updated and to capture critical human insights, especially at scale. Being able to document humans’ insights is a major step in preventing such knowledge from walking out the door when employees leave. 

The value of AI-powered knowledge management lies in its speed, consistency of providing reliable information, and accessibility to users across the organization or field. 

Users typically interact with knowledge systems through familiar chat-based interfaces, where they can ask questions and get reliable answers quickly. 

When users get answers faster, decision-making accelerates, onboarding times shorten, and expertise previously trapped in silos is suddenly accessible and usable. 

The Technologies Behind AI-Powered Knowledge Management

Modern knowledge management systems are fueled by a combination of machine learning, natural language processing, generative AI, automated content generation, chatbots, and virtual assistants. 

Knowledge Management in Action

Knowledge management can operate within an enterprise or as a specialized AI solution within a specific profession. 

An excellent example is OpenEvidence, an AI-powered medical knowledge platform. 

Physicians using this tool spend far less time skimming through medical journals to find the crucial information they seek. 

They can access a broad repository of medical research on a topic (including licensed and paywalled sources) from their computers and get summaries of the strongest evidence available. This addresses huge challenges like physician burnout and the growing challenge to keep pace with medical research.

Mechanics also reap the benefits of AI-powered knowledge management by talking to AI assistants to diagnose complex issues, explain unfamiliar systems, and resolve complex issues faster. 

Attorneys rely on similar tools to analyze volumes of legal documents, identify relevant legal precedents, and summarize findings while providing sources.

Easy access to knowledge removes the friction of time spent searching for information, especially in this era of information overload, making knowledge management transformative across organizations and professions. 

By simplifying access to expertise, knowledge management fundamentally changes how work gets done.

The Impact of AI-Powered Knowledge Management

Gartner found that organizations with effective knowledge management practices report higher productivity and faster decision-making, especially in complex, information-heavy roles. 

Like agentic automation, AI-driven knowledge management supports people in their roles rather than replacing them. It improves organizational efficiency, reduces friction in day-to-day operations, and prevents the challenges that come with forcing new tools or workflows on reluctant users. 

Conclusion: Why Boring AI Delivers the Most Value

As the AI hype cycle cools, a clearer path to measurable ROI is coming into focus. 

Across industries, the most effective AI initiatives today aren’t the loudest or most ambitious. They are boring: practical, embedded, and focused on execution. 

• Agentic automation frees employees from repetitive work. 
• AI-powered knowledge management significantly reduces the time spent searching for information.

Both work within existing processes and systems to improve day-to-day operations quietly, reliably, and at scale. They deliver real transformation without disruptive change management, steep adoption curves, or inflated expectations. 

Boring AI produces faster, more justifiable ROI faster because it starts small, builds on what already exists, and ties directly to operational outcomes. 

If you’ve been discouraged by past AI initiatives or hesitant to start at all, the takeaway is reassuring: the problem was not with AI itself. The problem is that organizations were focusing their efforts to look for value. 

In today’s market, boring is not a compromise. It’s where you can make money using AI.

About Marlabs: With 30 years of experience and 2,200 employees, NYC-based Marlabs is an AI transformation partner that helps Fortune 500 organizations operationalize AI and deliver sustained, measurable value across industries.