Strategies to unlock the potential of GenAI for businesses

In our latest podcast discussion, Raajaguru Mylsamy, Co-founder and CEO at Innoboon Technologies, joins Srikrishnan Ganesan, Co-founder and CEO at Rocketlane, to share strategies for deploying generative AI (GenAI) solutions in business.

With extensive experience at TCS, Rohde & Schwarz, and Accenture’s strategy consulting, Raajaguru brings deep expertise in implementing GenAI to create practical, scalable AI applications for enterprises. At Innoboon, he leads generative AI innovation, offering GenAI as a service and building AI applications for enterprises. From navigating the AI hype cycle to aligning business stakeholders, he offers a roadmap for executing quick Proofs of Concept (POCs) that deliver results. Leveraging techniques like rapid prototyping and accelerator bots, Raajaguru sheds light on achieving quick wins in the world of GenAI, making AI more accessible and actionable for businesses today.

Here are some of our favorite bits from the thoughtful insights Raajaguru shared.

Why generative AI is the future of business automation

The decision to pivot toward GenAI emerged from years of foundational experience in mobility solutions, video analytics, and data engineering. While these early initiatives provided valuable efficiencies, GenAI introduced capabilities that go far beyond the limits of traditional automation. Unlike previous approaches, which focused on repetitive tasks and structured workflows, GenAI has the power to automate cognitive functions—complex tasks that demand adaptability and intelligence.

This transformative shift opens the door for smarter, more dynamic applications across enterprise functions. Early adopters who harness models like ChatGPT, OpenAI’s suite, and Gemini are achieving unprecedented results, realizing GenAI’s capacity to drive genuine innovation. This new wave of intelligent automation not only boosts efficiency but also addresses higher-value, cognitive challenges, setting a path for organizations to lead in an increasingly AI-driven world.

The inside-out approach to unlocking GenAI's potential

The shift toward GenAI is driven by a clear vision of its potential, rather than solely by external demand. GenAI technology is still emerging, with maturing standards, regulations, and security protocols yet to be fully established. Conversations with business leaders reveal widespread interest in adopting GenAI, though many remain uncertain about the tangible value, return on investment, security, and functional applications it can truly deliver. This atmosphere of curiosity mixed with caution underscores the approach to GenAI as an inside-out transformation—an investment in potential rather than reactionary demand.

The market may not yet fully realize GenAI’s power, but the demand is anticipated to surge in the coming quarters. This foresight positions early adopters to stay ahead of the curve, strategically prepared to leverage GenAI’s full value as it evolves and becomes indispensable across functions.

Quick POCs to help enterprises adopt GenAI faster

As GenAI technology began to mature, it became clear that a gap existed in the enterprise sector: while companies were eager to explore GenAI, organizational constraints often hindered rapid experimentation. Bureaucratic hurdles, resource limitations, and lengthy approval processes made it challenging for large and mid-sized enterprises to adopt a test-and-learn approach. This is where the ability to deliver fast and focused Proofs of Concept (POCs) added significant value.

In early 2023, conversations with business leaders revealed a cautious curiosity, with many uncertain about the concrete benefits and ROI of GenAI solutions. But in the last year, the market's understanding has evolved significantly. Business leaders today approach with sharper questions and specific requests, such as defining precise functions for automation or incorporating designated documents into the knowledge corpus within large language models or LLMs. The transition from vague interest to focused requirements highlights an encouraging trend where enterprises are ready to move beyond exploration and deploy GenAI in strategic, result-oriented ways. This is a trend expected to grow stronger in the coming quarters.

GenAI readiness and addressing the complexities of deployment

While awareness and interest in GenAI have surged, true readiness to deploy solutions varies significantly across organizations, especially between mid-market and enterprise-level businesses. Many assume that implementing GenAI is as simple as fine-tuning an LLM or uploading documents into a model like OpenAI or Gemini. However, deploying an effective GenAI solution is a multifaceted process requiring more than a surface-level setup.

A successful GenAI deployment involves multiple competencies beyond LLM fine-tuning. It starts with a rigorous data engineering phase, which typically accounts for 30-40% of the development efforts. This stage ensures that the LLM can accurately interpret and use data, much like training a new employee. Data must be organized, cleaned, and structured to make it consumable by the LLM. This could include setting access controls based on organizational roles—such as ensuring finance data is restricted from supply chain management or HR teams, etc. These privacy and specificity controls are essential for ensuring compliance and maintaining data integrity.

Following data engineering, another 30-40% of the workload focuses on model training, pipeline configuration, and data integration with the LLM. A robust architecture allows data to flow seamlessly into the GenAI ecosystem, ensuring the model can provide reliable insights across applications.

Finally, there’s the software development and AI Ops phase. Here, the focus is on creating dashboards, APIs, and cloud infrastructure for operationalizing the solution. Integrating with databases, setting up scalable pipelines, and maintaining AI-specific operations (AI Ops) are crucial steps to streamline and support the system in production. 

Ultimately, deploying GenAI solutions is a multi-layered effort that requires expertise across data engineering, model development, software integration, and operational management—underscoring the need for readiness across various functions before meaningful results can be realized.

Accelerating GenAI adoption with strategic prototyping for quick wins

In GenAI development, speed and prototyping are key to capturing client interest and proving concept value quickly. Rather than skipping essential stages, effective prototyping involves strategically choosing components that require minimal engineering. This approach helps showcase the potential of GenAI without getting bogged down in complexities that would delay initial results.

For early prototypes, selecting manageable data sources—like text-heavy documents without complex images or mixed data types—minimizes data engineering requirements. This focus allows for a quick, impactful demonstration of the solution’s capabilities, often using powerful models like OpenAI or Gemini to connect data through a straightforward pipeline. These prototypes create a “wow” factor, showing clients how GenAI solutions could look and function in practice.

Once clients are on board and see the potential, a deeper decision-making phase begins. This is where scaling decisions are made—whether to maintain reliance on external models or transition to in-house solutions with open-source LLMs. At scale, the project expands to integrate various data sources, connectors, cloud infrastructure, and specialized resources, such as GPU management. This phased approach to deployment and adoption allows organizations to evaluate GenAI’s impact and feasibility before committing to a full-scale deployment, balancing rapid results with thorough planning for long-term success.

Ensuring POC success and cross-team collaboration

The success of a GenAI POC hinges on a close collaboration between functional and technical teams, rather than solely on the idea itself. Effective POCs require input from both the technology provider and the client’s functional experts. 

For instance, in developing a POC for a supply chain function, the functional team’s involvement is required to clarify the specific pain points, prioritize challenges, and shape the solution’s intended impact. This alignment ensures the POC addresses a real, high-value issue, setting it up for valuable results.

From the technology side, experienced GenAI providers, like Innoboon, bring the envisioned solution to life with the right tools and expertise necessary. Having a robust library of tools, like pre-built bots or customizable modules, enables technology teams to efficiently stitch together solutions tailored to the client's unique needs. This synergy between functional insight and technological prowess is essential for unlocking the full potential of a POC.

It is not just about how great an idea is, but the collaborative execution that determines success. A skilled, cohesive team can elevate even a modest idea, while a misaligned team might fail to realize the value of a promising concept. Organizations seeking to leverage GenAI should prioritize partnerships that foster collaboration and align functional and technical expertise—turning innovative ideas into impactful outcomes.

Validating GenAI investment and securing budget

POCs are an integral in GenAI adoption, born out of the need to demonstrate value before committing to larger investments. They provide a structured approach for enterprises to test innovative solutions, proving GenAI’s worth in a controlled, risk-managed manner while setting the stage for scalable, impactful deployments.

In enterprise businesses, POCs serve a dual purpose:

• They validate the technical and business value of GenAI solutions
• They help secure further budget for scaling these solutions

Enterprises typically allocate specific budgets for innovation, but these funds are limited and must demonstrate clear ROI and strategic impact before additional resources are committed.

Most POCs arise from two core motivations. The first is addressing pressing pain points—the immediate issues that demand quick, effective solutions. The second is the pursuit of innovation, where business leaders test GenAI's potential to transform operations and future-proof the organization over the next five years. In either scenario, POCs offer a tangible demonstration of GenAI’s benefits, enabling leaders to make a compelling case for further investment. As GenAI is still relatively new and misunderstood, especially in terms of production-ready applications, seeing concrete results through a POC is often essential for securing buy-in.

But enterprises do face significant roadblocks in executing POCs. Talent acquisition alone can delay GenAI initiatives by weeks, and assembling specialized teams within large organizations can be a 10-12 week process. Additionally, if a POC doesn’t meet expectations, redeploying resources or justifying ongoing roles adds complexity. This is where technology partners can be valuable, offering a rapid setup for POCs—typically forming teams within a week and delivering results within 10 weeks.

What happens when a client decides to shelf a project mid-POC?

POCs don’t proceed to full-scale production due to factors such as budget constraints or shifting organizational priorities.

Another common cause for stalled projects is the pursuit of POCs with only short-term objectives. If a POC is initiated without a clear, long-term business case or simply to meet immediate needs, it’s far more likely to face delays or deprioritization.

For enterprises and their technology partners alike, ensuring that the POC targets a core, high-value pain point is essential for demonstrating meaningful impact and justifying further investment. This requires a deep initial understanding of the client’s business landscape, including the potential ROI or cost savings associated with solving a specific problem. Without this, even the most well-executed POC may fail to resonate with decision-makers.

Two critical aspects during this phase are collaboration and clarity. Both the enterprise and the technology provider must work together to outline the POC's intrinsic value from the start. When POCs are aligned with the client’s strategic objectives and clearly address top pain points, they’re positioned for sustainable impact rather than temporary gains. This strategic approach enables organizations to leverage GenAI solutions that are not only feasible but also truly transformative, setting the stage for full-scale adoption.

Driving GenAI efficiency with accelerators and bots

In the world of GenAI, accelerators and pre-built bots serve an important role in bringing ideas to life for clients, especially when exploring new and complex use cases. These tools allow business leaders to see how a solution might function in their specific context before investing in a full-scale implementation. 

When Innoboon began discussions with clients, it became evident that illustrating the potential of GenAI solutions was a challenge. This led to the development of over 250 “bots”—lightweight, functional prototypes built to showcase practical applications and inspire client confidence.

Initially, these bots were intended solely for demonstration. For example, an HR bot might simulate resume filtering and automated interviewing, while a finance bot could showcase capital management or financial modeling. And by developing lean models for various sectors, the team could provide clients with tangible examples of how GenAI could address their unique needs. Over time, however, these front-end demonstrations naturally evolved into a robust library of back end functionalities. What began as specific use cases led to the creation of reusable components that now streamline the GenAI development process.

These backend mechanisms—from voice-to-text transcription pipelines to structured data context setting—have become versatile assets. Each bot contributes to a growing ecosystem of APIs and modular components that can be easily integrated into diverse solutions. For instance, a knowledge graph conversion initially built for one project can now be applied across different applications with ease. This library of bots allows for quicker, more cost-effective development, reducing the time and expense for clients.

Rather than selling these assets as off-the-shelf products, the approach involves incorporating them seamlessly into custom solutions. This advantage enables Innoboon to deliver projects that might take 100 days and $100,000 to complete to be delivered in 60 days at a significantly lower cost. These accelerators are a testament to the value of sustained development and deep technical investment, allowing businesses to access GenAI solutions that are faster, more efficient, and, ultimately, more impactful.

Adopting a problem-first approach to GenAI

One of the most valuable lessons learned in GenAI solution development is adopting a problem-first approach. Rather than creating solutions in anticipation of hypothetical needs, the focus remains firmly on addressing specific client challenges. This ensures that solutions are not just technically robust but are also directly relevant and impactful.

The philosophy is simple: don’t build a solution and then search for a problem it can solve. Instead, identify a clear, pressing problem and then develop a solution tailored to it. In doing so, solutions are designed to be as modular and generic as possible, allowing them to be easily integrated across various use cases. When a POC requires a particular functionality, the team looks to leverage existing bots where possible. If a new bot is needed, it’s designed in a modular way, enabling it to stand alone and be repurposed in future projects.

This ensures that each bot or microservice is built with intention and flexibility, providing targeted solutions while expanding a versatile toolkit that can be deployed across multiple client needs. The problem-driven focus avoids the inefficiency of solution-first development, allowing for strategic growth in modular assets that enhance speed and adaptability in GenAI projects.

Balancing quick wins and long-term client expectations

The POC phase in GenAI projects is designed to quickly showcase a potential solution’s value, often achieved by rapidly building a preliminary model that gives clients a glimpse of the end result. But this rapid progress can sometimes set unrealistic expectations for what follows, as clients may assume that moving from POC to production will be just as swift. The reality, however, is that a POC typically represents only 30-40% of the overall effort, with the remaining work involving intensive data engineering, InfoSec measures, and other foundational steps necessary to scale and sustain the solution in production.

This "quick win" approach can lead to misunderstandings if not managed properly. While it’s relatively easy to develop an initial view of the solution, refining it to meet enterprise standards requires significant fine-tuning and deeper integration efforts. It’s common for the last 10% of a project to require a substantial investment of time and resources to address intricacies like data accuracy, security frameworks, and integration with existing systems.

For this reason, business leaders must be careful when selecting a vendor for GenAI projects. It’s essential to work with partners who not only deliver rapid prototypes but also set realistic expectations for the journey from POC to full deployment. Experienced vendors, familiar with the challenges of scaling AI solutions, can proactively plan for potential roadblocks and align POC development with long-term needs. This ensures that while the POC impresses, it also establishes a foundation for sustainable, scalable success—delivering on both initial impact and durable value.

Raajaguru Mylsamy’s favorite quote from a book or movie

Protect your dreams and keep moving forward—no matter the setbacks -  inspired by The Pursuit of Happiness and the film Rocky as it shows resilience, embracing failure, and learning to work smart.

The road ahead for GenAI

The next few quarters promise significant advancements in GenAI as regulations and policies evolve to make the technology more reliable and business-ready. With a wave of industry momentum, GenAI is poised to make a transformative impact across enterprise functions, similar to the dot.com boom's effect in past decades. The ability to automate cognitive tasks at scale is expected to reshape various sectors, unlocking new levels of efficiency and innovation.

As early adopters, Innoboon is excited to be at the forefront of this shift. Our focus remains on solving real business problems and driving measurable value for enterprises. Revenue naturally follows from this commitment to impactful solutions—our true excitement lies in helping organizations leverage GenAI to create lasting, meaningful improvements.

Advice to young entrepreneurs starting out in GenAI

For those entering the GenAI space, deep technological expertise isn't essential at the outset—resources to learn are widely available. The key to success, however, lies in adopting a problem-first approach. Rather than creating a solution and searching for a problem to match, entrepreneurs should focus on understanding the most pressing pain points in their target industry. This requires patience, observation, and often collaboration with domain experts to identify where GenAI can make the greatest impact.

Avoid the common pitfall of being overly committed to a single technology or solution; flexibility in choosing the right tools can reveal better alternatives. For those who have technical skills, partnering with domain experts is invaluable, and vice versa. Combining domain knowledge with technical expertise creates a “sweet spot” where meaningful, high-impact GenAI solutions bring out the magic!