Rapid Pace of Change and the Need for Agility

In today’s modern enterprise landscape, the adage “change is the only constant”—coined by the ancient philosopher Heraclitus—has never been more relevant. This phrase encapsulates the perpetual state of flux that defines the contemporary business environment. Rapid technological advancements, especially in the Generative AI (GenAI) space, are disrupting traditional enterprise models and forcing organizations to rethink their strategies. Enterprises must empower their teams to innovate and bring ideas to fruition swiftly, while ensuring responsibility and maintaining transparency with consumers. At the same time, they must scale rapidly with strong security and compliance measures—a significant challenge. In essence, organizations must adapt swiftly to technological changes but do so responsibly.

According to a recent Gartner report, enterprises have allocated approximately 4.7% of their IT budgets to GenAI in 2024, with this figure projected to rise to 7.6% by 2027. However, Gartner also predicts that by the end of 2025, at least 30% of GenAI projects will be abandoned. The three key reasons for the project failure include challenges in scaling and compliance of the platform across the enterprise, the inability to maintain high data quality, and maintaining the costs under control.

Can a Central Enterprise Platform Approach Address the Challenges?

At Mars, a family-owned company with $50 billion in annual sales and a diverse, expanding portfolio of quality snacking, food, and pet care products and services, we faced a similar challenge of needing to adapt our technology quickly and responsibly. At Mars, using generative AI responsibly is non-negotiable and our top priority, which led to the creation of a Responsible AI board, community, and policy. Mars not only empowered Associates to use generative AI in a way that is fair, inclusive, and sustainable, but also prioritized the protection of privacy and security for our people, consumers, and partners. 

Within this Responsible AI framework, Mars empowered its business units to explore and experiment with GenAI initiatives. However, as experimentation grew, the IT team encountered challenges with core processes such as provisioning infrastructure and standardizing LLMOps practices. Moreover, teams were using multiple tools and platforms for content generation, and it was critical to ensure that content generated by these platforms complied with policies and guidelines. Mars’ executive team quickly realized that these siloed efforts were duplicative, costly, and difficult to scale enterprise-wide due to non-standardized tools, multiple partners/vendors, and varied LLMOps processes.

To fully harness the potential of GenAI, Mars adopted a unified multimodal platform and central operations approach—one that would scale across the enterprise, enable cross-business collaboration to realize true value, consolidate GenAI use cases, enhance user experiences, strengthen risk controls, ensure compliance, and empower Associates to unlock their creativity and boost productivity. This unified multimodal approach also streamlines operations and accelerates innovation by leveraging cutting-edge tools such as pre-built models, plug-and-play modules, and processes like standardized LLMOps, with a strong focus on AI observability.

Additionally, by implementing guardrails for content generation that align with organizational policies and tone of voice, business units can reduce time to market and minimize duplication of efforts. The unified multimodal platform approach provides a structured framework, enabling faster deployment of use cases while maintaining consistency, compliance, and adherence to Responsible AI principles.

Industry-Wide Shift Towards Centralized Platforms

Tredence’s experience at Mars reflects a broader movement across industries. Global enterprises are increasingly recognizing that while experimental projects are valuable for testing ideas, they often fail to deliver long-term value without a scalable framework. As the GenAI landscape evolves, businesses must respond by building flexible, future-proof platforms that can grow as the technology matures.

Enterprises, such as the largest beverage and personal care companies are now prioritizing the development of playbooks, frameworks, and standardized operational tools to drive efficiency and eliminate duplication of effort arising from siloed GenAI projects. By centralizing operations into a unified multimodal platform, businesses can streamline processes and ensure consistent results across multiple use cases. The Mars example underscores how companies must adopt such frameworks to evolve their AI strategies from isolated experiments into cohesive, scalable solutions.

Identifying the Right Use Cases for Impact

One of the critical decisions that the Mars team had to make was to identify the GenAI use cases that would deliver the most value to the business. Of the use cases, 60% could be solved directly by advanced prompt engineering techniques, 30% could be solved using RAG-based solution approaches, and the remaining 10% required fine-tuning of existing models and, in some cases, a domain-specific LLM model. In addition, success with GenAI (across many enterprises within the industry) often hinges on selecting strategies that not only improve productivity but also enhance user experiences.

Based on our experiences, here are some strategies for identifying high-impact GenAI use cases in large organizations:

1. Align with Business Objectives: Start by identifying areas where GenAI can directly support your organization’s strategic goals. Focus on use cases that enhance employee productivity (that allow them to save time, which in turn can be used to produce higher-order value), improve customer satisfaction, and reduce costs to ensure executive buy-in.

2. Assess Impact and Feasibility: Prioritize use cases based on their potential impact and technical feasibility. Use an impact-feasibility matrix to identify projects that offer significant benefits with manageable implementation efforts.

3. Leverage Existing Data Assets: Choose use cases that can utilize your organization’s existing data. High-quality, abundant data accelerates model training (if needed) and increases the likelihood of successful outcomes.

4. Enhance Customer Experience: Look for opportunities where GenAI improves employee and customer interactions, such as integrating capabilities within preferred tools like messaging apps, plugins, extensions, and internal portals, to enable personalized recommendations that drive engagement and productivity improvements.

5. Optimize Internal Processes: Identify repetitive or time-consuming LLMOps tasks and processes that can be automated. Automating these processes and leveraging existing frameworks boosts efficiency and frees up employees for higher-value activities.

6. Cross-Functional Collaboration: Involve stakeholders from different business units to uncover diverse needs and pain points. This collaborative approach ensures that selected use cases have broad organizational support.

7. Pilot and Scale: Begin with small-scale Minimum Viable Products (MVPs) to validate assumptions and demonstrate value. Successful MVPs can then be scaled up, reducing risk and building confidence in GenAI initiatives.

8. Ensure Compliance and Ethics: Ensure that platforms are compliant from the ground up with Responsible AI principles and/or can be adapted according to legal and regulatory guidelines. Evaluate potential use cases for compliance with industry regulations and ethical standards. Addressing these concerns upfront mitigates risks.

9. Invest in Scalable Platforms: Opt for use cases that can be supported by scalable GenAI platforms. This approach allows for future expansion and adaptation as technology and business needs evolve.

10. Monitor and Measure: Establish clear metrics to track the performance of GenAI initiatives. Continuous monitoring helps in refining models and demonstrates tangible benefits to stakeholders.

At Mars, we prioritized use cases that directly target productivity improvements for associates so that they can see and realize the value quickly. 

Industries like retail and finance are also adopting these strategies, leveraging AI to deliver more personalized experiences to customers, optimize supply chains, and enhance decision-making across departments. The key lesson is that scalable platforms empower not just AI specialists but business users, allowing them to experiment with pre-approved models and applications with fewer risks.

Reusability as the Key to Efficiency

One of the most significant advantages of our platform approach at Mars—and a trend we see across the industry—is the ability to reuse best-of-breed GenAI models and components. By creating a library of reusable, pre-baked assets, we’ve been able to dramatically shorten the time it takes to bring new use cases to life. This approach is critical for any organization aiming to scale GenAI, as it reduces development time, minimizes cost, and ensures that the solutions deployed are reliable, less prone to hallucination, and aligned with brand standards.

At Mars, we’ve found that in around 50% of the use cases, we can leverage templatized GenAI solution patterns that can be reused across different applications, helping us onboard new use cases far more quickly than before. This shift towards reusability is becoming an industry standard, with companies measuring GenAI success not just by the number of projects completed, but by how quickly and efficiently new solutions are implemented.

Empowering Business Users with GenAI Tools

In summary, Mars and Tredence truly think that the unified multimodal platform approach with compliance with Responsible AI principles is the way forward for larger organizations because it empowers both technology and business users to leverage AI in their daily work in a safe way. 

The democratization of GenAI is essential for driving innovation at scale. Across industries, organizations are recognizing that for GenAI to reach its full potential, it must be accessible to business users, enabling them to integrate AI into their workflows without advanced technical skills. By embedding GenAI into everyday processes, businesses can enhance decision-making, improve consistency, and drive meaningful outcomes.

While companies like Mars have focused on optimizing Associate experiences and delivering business value through GenAI with Responsible AI principles, the future of this technology lies in transforming consumer/customer-facing applications. 

GenAI is not just a tool for operational efficiency—it is a critical enabler of innovation and long-term growth.

About the Authors:

Kaarthik Subramaniam is the Director of Digital Experiences Platforms at Mars Global Services. Kaarthik is a digital transformation leader with expertise in building digital platforms and driving innovation. At Mars, he led the Mars Experiences Platform, launching 300+ digital sites across 65+ brands in 50+ countries and cutting operational costs by 40%.

Kaarthik also works on GenAI initiatives, creating tools that improve productivity and teamwork. He has been recognized in CDO Magazine’s 2024 40 Under 40 and as one of Chicago’s Top 20 Web3 Innovators.

Dr. Ravindra Patil serves as Vice President, overseeing Data Science at Tredence Inc. With over 15 years of experience in Data and AI, he leads efforts to develop practical solutions that address complex business problems.

Ravindra holds a Bachelor’s in Engineering, a Master’s from IIT Madras, and a Ph.D. from the University of Maastricht. He has filed over 30 patents, published several research papers, and was named one of India’s top 100 AI leaders and 40 under 40 Data Scientists by AIM magazine.

Venkatesh Rajagopalan serves as the Director, overseeing Data Science at Tredence Inc. Venkatesh is a seasoned data science leader with over 18 years of experience in building GenAI platforms and solutions. He specializes in translating complex AI concepts into scalable, impactful systems that drive value across industries.

With a proven track record of delivering technology-driven transformations, Venkatesh focuses on creating solutions that enhance productivity, collaboration, and decision-making for enterprises.