Maximizing AI Success: Strategies for Delivering Genuine Business Value

A staggering 95% of AI projects fail, as reported by MIT research. This daunting statistic raises eyebrows in boardrooms and among IT leaders, prompting a quest for answers about why such a substantial percentage of initiatives do not realize their intended value. Resources are being poured into AI pursuits, but the disparity between expectation and outcome continues to widen, leaving many to wonder where they are going wrong.

One major barrier to successful AI implementation lies in the fragmentation of data across organizations. Data often exists in isolated silos, making comprehensive access difficult. As noted by many experts, AI models are only as good as the data fed into them; without unified access to quality data, even the most advanced algorithms can produce results that are irrelevant or inaccurate. This underlines the critical concern surrounding data quality and governance.

“Data needs to be accurate, clean, and managed under clear governance,” conveys Rusty Searle, interim CIO at Elastic. “You can push all the data into an AI tool, but if the quality isn’t there, the output won’t tell the right story. Poor or siloed data will always show up in the results.”

Neglecting this foundational step can leave organizations with AI systems that generate impressive demos yet deliver little in the way of consistent business value in live environments. Jay Shah, senior director of enterprise applications at Elastic, echoes this sentiment by stating, “Garbage in, garbage out. It’s paramount to have fast access to quality data. It directly impacts the contextuality, reliability, and accuracy of the response output.”

Focus on Solving Business Problems, Not Just Technology

To navigate this challenging landscape, leaders must pivot from a technology-first mindset to a problem-first approach. Instead of asking, “What can AI do?” they should start with the question, “What specific business challenges need resolution?” This focused strategy ensures that AI initiatives align with measurable outcomes and correspond to genuine user requirements.

“Every AI initiative should connect back to a defined problem with measurable outcomes,” emphasizes Searle. Rather than chasing the latest technology for its own sake, clarity on the use case must guide investments and efforts. AI cannot thrive as a mere IT initiative. Effective implementation necessitates collaboration with various business functions—from HR to legal—to define requirements and maintain data quality. This collective effort not only drives strategic goals but also enhances user adoption: ultimately, it’s irrelevant how cool AI may be if it’s underutilized.

Elastic’s journey illustrates these principles through its development of ElasticGPT, a generative AI employee assistant aimed at significantly boosting productivity. This tool enables employees to locate crucial information—ranging from onboarding checklists to company policies—efficiently, drawing from multiple data sources. The goal was not just to build something flashy but to establish a scalable foundation for future AI endeavors.

Designate an AI Champion

Another cornerstone for successful AI adoption is identifying a dedicated leader—an “AI champion”—to drive the initiative forward. “Successful AI implementations involve identifying a single-threaded leader who can guide your organization’s vision,” advises Shah. This individual should lead a focused team working toward clear objectives and measurable outcomes. By navigating the intricacies of IT infrastructures, compliance issues, and data evaluation, this core group can effectively orchestrate the initiative across business lines.

This leader should also facilitate cross-departmental collaboration with business stakeholders, ensuring that the AI tools developed truly meet user needs. Conversely, trouble often arises when AI applications are restricted to the confines of IT without any buy-in from other departments.

Meeting Users Where They Are

Successful AI applications thrive when they seamlessly integrate into existing workflows. Users are naturally more inclined to adopt AI solutions that enhance their current practices rather than require them to learn new tools or processes. When AI functionalities appear within the platforms where users already operate—like communication tools or existing applications—adoption tends to happen organically.

“It’s critical to embed AI into existing workflows, ensuring people are trained and comfortable using it. Otherwise, it becomes shelfware,” Searle cautions. An excellent illustration of this approach is ElasticGPT’s integration into Slack, a platform employees already use daily. By embedding AI in familiar environments, it enhances user experience rather than altering established habits.

Change management is equally vital as AI technologies evolve rapidly. Organizations should prioritize ongoing education to cultivate a culture of growth and experimentation, helping team members adapt to new tools and methodologies.

Recover When Things Go Off Track

The complexity and novelty of AI technologies make it unrealistic to expect that all implementations will proceed flawlessly. Elastic learned this early in its AI journey, facing challenges as it sought to attach AI assistants to existing systems. It soon became apparent that better orchestration was necessary to ensure secure, contextual AI interactions. Adopting Langchain allowed Elastic to structure its framework more effectively, enhancing access controls and aligning AI with existing systems.

Real-World Success with a GenAI Employee Assistant

The effectiveness of ElasticGPT serves as a practical demonstration of these principles. Developed on Elastic’s Search AI Platform using retrieval augmented generation (RAG), the tool successfully connects disparate data sources, making information readily accessible. The results are compelling: employees save approximately 63 hours annually, with a two-month payback period and an impressive 98% user satisfaction rating.

Several key decisions contributed to this substantial success:

• First, HR and IT teams engaged with users to directly address their common challenges, such as difficulties in finding important information or complex upgrade procedures.
• Second, Elastic prioritized data quality and accessibility, cleaning up its internal resources as part of the development process.
• Lastly, the organization allowed multiple access points for employees—such as via the company intranet and Slack—integrating AI into existing workflows and minimizing friction in adoption.

Building Successful AI Applications

Successful generative AI applications require strong leadership, high-quality data, and seamless integration into daily operations. The journey calls for moving beyond isolated software solutions and towards comprehensive platforms that scale with business needs. Pragmatic steps can transform AI from a mere concept into a vital component of organizational success.

For organizations looking to drive real business value with AI, following Elastic’s eight steps to building a scalable generative AI app can yield measurable results.