Business
Belton: A Case Study in Industrial Resilience and Civic Longevity
From a railroad junction to a modern manufacturing hub, Belton demonstrates how historical industrial foundations suppor...
Pablo Gomez Tena bridges deep math-driven insight with product strategy to turn AI from buzzword into measurable value across consumer banking. By balancing innovation with risk, governance, and cross-functional alignment, he shows how resilient AI roadmaps can deliver trusted, compliant, and scalable impact.
The financial services industry is undergoing a profound transformation, largely propelled by the rapid advancements and adoption of AI and ML. No longer confined to experimental labs, AI has emerged as a cornerstone technology, fundamentally reshaping how financial institutions operate, innovate, and compete. The scale of this shift is underscored by significant investment and market growth.
The scale of this shift is underscored by significant investment and market growth. According to a KPMG survey, a striking 84% of leaders at U.S. financial services firms plan to increase their investments in generative AI over the next year. This escalating commitment reflects a clear understanding that AI is pivotal for driving revenue growth, enhancing operational efficiency, and delivering superior customer experiences.
This escalating commitment reflects a clear understanding that AI is pivotal for driving revenue growth, enhancing operational efficiency, and delivering superior customer experiences. As organizations move beyond pilot programs to full-scale AI implementation, the need for astute leadership capable of navigating this complex technological frontier has never been more critical.
At the vanguard of this intersection between AI innovation and strategic deployment within the financial sector is Pablo Gomez Tena. His professional identity is centered on AI leadership and strategic product development, showcasing a distinct ability to spearhead AI product roadmaps, oversee rigorous strategic development and feasibility assessments, and articulate compelling overarching AI visions. With twelve years of experience dedicated to solving complex problems through data and critical thinking, predominantly within the consumer banking segment of financial services, Pablo brings a wealth of practical industry knowledge.
Pablo’s expertise spans a wide array of critical domains, including data and analytics, AI, ML, technology strategy, and product management. His proficiency extends to specialized areas such as credit risk management, consumer lending, data management, data governance, business intelligence, analytics engineering, data engineering, consulting, business strategy, and AI strategy, including the nuanced application of design of experiments. This comprehensive skill set directly addresses the evolving demands of the financial industry, which is increasingly focused on delivering personalized financial solutions and strengthening risk management frameworks through advanced technological capabilities.
Leaders like Pablo, who can meld technical mastery with product management acumen and a clear strategic vision, are indispensable for organizations aspiring not merely to adapt but to lead in this era of AI-driven transformation.
The journey from abstract mathematical principles to the strategic leadership of AI-driven product initiatives is one that requires a unique blend of analytical rigor, practical application, and effective communication. Pablo’s career trajectory exemplifies this path, beginning with an inherent fascination for mathematics. He was drawn to “numbers and the structure that comes with solving problems through logic and proof.”
This early inclination shaped a way of thinking that he describes as “scientific, curious, and always looking to understand the ‘why’ behind things.” Such a mindset is a natural precursor to a career in analytics and data science, particularly within the data-intensive financial services sector, where he began applying mathematical concepts to tangible business challenges.
As Pablo’s career progressed, he found himself increasingly involved in the strategic dimension of analytics—less about the intricacies of model construction and more about conceptualizing how these models could address significant business problems, such as refining approaches to credit risk modeling within a major banking institution. This transition highlights a critical capability in AI leadership: the ability to connect technical possibilities with strategic imperatives.
“My math background gave me the technical grounding to understand what’s under the hood,” Pablo explains, “and my communication skills helped me bridge the gap between data scientists, engineers, and business leaders.”
This “bridging” role is exceptionally valuable in AI projects, which often involve diverse teams speaking different technical and business languages. Effective communication strategies, including storytelling and the use of visual aids, are essential for making complex AI concepts accessible to all stakeholders.
Pablo’s journey illustrates a key aspect of effective AI leadership: the most impactful leaders often serve as “translators,” possessing both the deep technical insight to understand the nuances of AI and the business acumen to articulate its strategic value. His mathematical foundation is not merely a historical footnote but an active component of his leadership approach, enabling him to critically evaluate AI capabilities and limitations—a vital skill for strategic decision-making.
This combination of technical depth and strategic communication acumen naturally led Pablo to spearhead AI initiatives at the product level. He perceives AI not merely as a technical instrument but as a solution that must be viable and beneficial for all parties involved. This holistic perspective is further enriched by his experience in risk management, a field that demands a keen awareness of potential downsides and unintended consequences.
“My background in risk management also adds a layer that many product people don’t always bring—I’m always thinking about the implications of what we build, not just what’s possible,” Pablo notes. “It’s all about finding the balance between innovation, usability, and responsibility.” This statement encapsulates a mature approach to product development, particularly pertinent in the financial services industry where risk and regulation are paramount.
The proactive integration of risk considerations into the AI innovation lifecycle is a significant differentiator. While many technologists focus on the sheer capability of AI, Pablo’s financial services background instills a discipline of evaluating what AI should do and what safeguards are essential, reflecting a commitment to responsible AI development that is critical in high-stakes deployments. This preemptive focus on balancing innovation with usability and responsibility is a hallmark of a sustainable and ethically sound AI strategy.
Developing a successful AI product roadmap requires a meticulous approach that harmonizes technical feasibility with overarching business objectives. For Pablo, the cornerstone of this process is a profound understanding of user needs. “To ensure technical feasibility, it is always paramount to consider the user requirements,” he asserts. “Segmenting these into feasibility levels is an important exercise to determine what features are feasible and which ones are not.”
This initial segmentation involves a careful value-effort tradeoff analysis, a practice central to effective AI product roadmap planning, which emphasizes defining core AI goals tied to business outcomes and prioritizing initiatives based on their potential impact and ease of implementation. Modern approaches even leverage AI itself to optimize roadmaps through predictive analysis and backlog automation, further enhancing the strategic planning process. Pablo’s method inherently mitigates risk by front-loading the identification of potential roadblocks through early feasibility segmentation and continuous technical validation, an adaptive strategy well-suited to the often-unpredictable nature of AI projects.
Continuous dialogue with technical teams forms another critical pillar of Pablo’s roadmap strategy. He underscores the necessity to “consistently check with the engineers and scientists to get their availability, the technical difficulty in building and scaling the solution, as well as getting cost estimates.” This iterative validation is indispensable for accurately assessing technical feasibility throughout the project lifecycle.
Such an approach aligns seamlessly with agile development methodologies, which are frequently employed in AI/ML projects due to their iterative nature. Unlike rigid, waterfall-style planning, this dynamic assessment allows for adjustments and refinements as the project evolves and new technical insights emerge. This continuous feasibility check ensures that the roadmap remains grounded in reality, preventing costly overruns or deviations from achievable goals.
Ensuring robust business alignment is the final, crucial component of Pablo’s roadmap planning. Once features are prioritized based on the value-effort assessment, he stresses that “it is important to define criteria for success and design metrics to help measure against it. Having a clear understanding of what the user considers successful will pay dividends in the mid-term while also aligning with the capabilities of the technical team.”
This proactive definition of success, deeply rooted in user expectations and cognizant of technical capabilities, serves as a guiding beacon for the development team. Furthermore, Pablo highlights that “communicating progress and milestones is a key step in alignment for a successful AI product.” This emphasis on transparency and proactive stakeholder management throughout the product’s journey is vital for preventing the common pitfall where a final product, despite technical soundness, fails to meet underlying business expectations. This aligns with strategic frameworks that advocate for linking AI initiatives directly to the overarching mission and vision of the business.
The integration of AI into financial services is not merely an incremental change but a fundamental disruption poised to reshape every facet of the industry. Pablo firmly believes that “AI will disrupt financial services across all practice areas—from customer service to trading to credit risk management to marketing.” While the level of regulatory scrutiny may differ by region and specific application, he emphasizes that a proactive stance on risk management “should be top of mind for any leader.”
This perspective champions engagement with risk as an integral part of the innovation process, rather than an afterthought or an impediment. The financial services sector is indeed in the midst of a significant AI-driven transformation, a shift that inherently introduces a spectrum of new risks that demand careful navigation.
Pablo advocates for a balanced approach to AI deployment, stating, “Deploying innovative AI-powered products is a unique way to improve productivity across dozens of different workflows; the use-cases are countless and the possibilities are endless. However, this does not mean that companies should dive into AI product development or procurement while ignoring the considerable risks.” He specifically identifies several key risk categories: regulatory risk, cybersecurity threats, the potential for AI models (especially Generative AI) to “hallucinate” or produce incorrect information, and business continuity risks.
These are widely acknowledged challenges in AI adoption across industries, and they take on heightened significance within the financial domain. His core message is clear: “It is paramount to understand what risks a particular organization is exposed to and how introducing an AI-powered product could expose or enhance those risks.” This understanding forms the bedrock of responsible innovation.
A pragmatic, cost-benefit analysis is central to Pablo’s philosophy on AI adoption. “Financial services organizations should always consider the cost-benefit of introducing an innovation to their processes, AI or not,” he advises. “AI products provide great promise, but the reward needs to be greater than the downside.”
This counsel underscores the imperative for meticulous financial justification before committing to AI projects. He cautions against the allure of using sophisticated AI to solve minor problems, especially if the associated costs and risks are disproportionately high, warning that “Many times, financial services organizations dive in first to solve a $1M problem, but fail to realize the exposure and cost of procuring an AI product is double or triple the benefit.” Achieving a holistic view, which balances potential rewards against inherent risks and implementation costs, necessitates proactive and transparent communication across all stakeholder groups.
The failure to accurately assess this value-effort tradeoff can lead to “AI washing”—implementing AI for superficial reasons without generating real ROI, thereby potentially increasing risk exposure for minimal actual gain. Ultimately, Pablo posits that the transformative power of AI can only be harnessed effectively under specific conditions. “AI is a game-changing technology, but it needs to be deployed and used effectively,” he concludes. “This means consistent communication, transparent usage policies, robust training, and effective measuring of the impact.”
These elements are foundational to establishing robust AI governance and fostering responsible AI practices. Without this comprehensive framework for deployment and oversight, “the risks outweigh the benefits.” This approach signifies that innovation and risk management are not opposing forces but complementary aspects of a successful AI strategy, creating an environment where innovation can flourish responsibly.
Pablo’s framework for balancing AI innovation with risk in finance can be distilled into several key considerations for leaders. These include defining if the problem being solved is significant enough to warrant an AI solution and thoroughly evaluating regulatory, cybersecurity, hallucination, and business continuity risks.
A critical cost-benefit analysis must determine if the potential reward clearly outweighs the potential downside and costs. Furthermore, a holistic view requires engaging all relevant stakeholders through proactive communication and ensuring enabling factors such as consistent communication, transparent usage policies, robust training programs, and effective impact measurement mechanisms are in place.
The development of sophisticated AI solutions, particularly those leveraging LLMs, often presents a complex maze of technical, strategic, and organizational challenges. Pablo recounts a particularly demanding project: leading the creation of an LLM-powered platform designed to provide users with broad yet safe and controlled access to LLM capabilities. This included functionalities for interacting with documents, such as summarization, drafting, and analysis.
Such endeavors are inherently intricate, involving careful consideration of data handling protocols, robust security measures like encrypted API calls and secure document management, system performance, and user experience. The development and deployment of LLM platforms are known for these multifaceted challenges, requiring adept leadership to navigate.
The most significant hurdle, Pablo identifies, was not purely technical but deeply rooted in organizational dynamics. “A major hurdle was aligning different stakeholders—engineers, compliance officers, product managers—each optimizing for their own priorities,” he shares. “Compliance directives often conflicted with user experience goals, particularly around how document data could be stored or processed.”
This tension between regulatory requirements and user-centric design is a common scenario in the development of AI products within regulated industries like finance. Pablo’s role became pivotal in mediating these differing viewpoints: “My role was to navigate those tensions, translate technical and regulatory constraints into understandable trade-offs, and guide the team toward workable solutions.” This highlights his crucial function as a facilitator and a translator of complex, often competing, requirements, a skill increasingly vital in AI leadership.
The path to resolution was paved with a user-centric approach and transparent communication. “We relied heavily on user requirements to shape the product direction, while managing expectations and clearly communicating what was feasible across all groups,” Pablo notes. This strategy proved key in achieving a delicate balance.
Securing approval from compliance and risk stakeholders without excessively compromising the core functionality of the platform was a significant achievement. Beyond the strategic negotiations, Pablo also focused on the practical enablement of his technical team: “Eventually, we secured compliance and risk approval without compromising core functionality. At the same time, I made sure the engineering team had everything they needed to build—whether it was computing power, clarity on cost models, or architectural guidance.” This attention to the team’s operational needs underscores a hands-on leadership style.
A final, critical element in bringing the LLM platform from concept to reality was securing the necessary investment. “Securing funding was another critical step, which required making a strong business case backed by projected usage and expected value,” Pablo states. This is a necessity to effectively present an AI venture to investors.
The successful culmination of these efforts allowed the product to launch. Pablo summarizes the key ingredients for this success: “That mix of user insight, strategic communication, and stakeholder alignment helped us move the product from concept to launch successfully.”
This experience demonstrates that in complex AI projects, particularly those involving novel technologies like LLMs within the financial sector, leadership that can effectively manage internal organizational complexities is as vital as leadership that can manage external technical and market challenges. The ability to anchor decisions in user value, even amidst strong compliance and engineering pressures, provides a powerful mechanism for navigating internal disagreements and fostering a shared path forward.
The development and deployment of impactful AI solutions are rarely the work of a single department; they are inherently cross-functional endeavors requiring seamless collaboration between data scientists, engineers, and business stakeholders. However, achieving this synergy can be challenging, as these groups often operate with different objectives and prioritize distinct metrics.
Pablo points out this common scenario: “In many cases, every team is optimizing for different metrics. For example, data scientists could optimize for model accuracy at the cost of more compute expenditures, which would go against business stakeholders’ key optimizations: monetary costs.”
This divergence in optimization goals is a primary source of friction in AI projects, making effective cross-functional collaboration a critical success factor. Pablo’s strategy for fostering collaboration begins with a deep dive into the underlying motivations of each team. “Understanding and mapping each of the team’s motivations will provide a holistic viewpoint of where the development picture stands at the beginning of a project,” he explains.
This initial “mapping” phase is not just about understanding tasks but about uncovering the core drivers and incentives for each functional group. By making these motivations transparent, a leader can then work towards building a shared understanding and ensuring that “each team feels represented and that their requirements are taken into account.” This approach moves beyond simple coordination to achieve a more profound “motivational alignment,” which is essential for navigating the inherent trade-offs in AI development.
Following this crucial step of understanding and acknowledging diverse motivations, Pablo emphasizes inclusivity and a steadfast focus on the end-user. “Once this is achieved, it is important to ensure each team feels represented and that their requirements are taken into account. Then, ensuring the final customer/user is aligned with decisions made is paramount, as they will be the ones who make the final decision.” This principle of user-centricity serves as a powerful, neutral arbiter when technical, business, and data science priorities inevitably clash.
By framing decisions around what best serves the user, internal conflicts can be de-escalated, and collective efforts can be refocused on a common external goal. Effective product managers play a key role in facilitating this user-focused alignment across teams.
The “glue” that binds these collaborative efforts, according to Pablo, is unequivocal: “It goes without saying that transparent, effective, and honest communication is the glue that holds everything together. Many times, teams get confused by other teams’ requests, so providing as much clarity on this and why it is important to the final product is a cost-effective way to ensure cohesive product development.”
This highlights the indispensable role of continuous, clear communication in bridging understanding gaps and ensuring that all teams work cohesively towards a unified product vision. In AI projects, where technical concepts can be particularly complex and prone to misinterpretation, investing time in ensuring clarity and explaining the “why” behind requests is not merely a soft skill but a tangible project efficiency lever. Such transparent communication minimizes confusion, reduces the need for rework, and prevents costly delays, thereby directly impacting project timelines and budgets.
The launch of an AI or ML product is not the end of the journey but the beginning of a crucial phase: measuring its real-world impact and ensuring it delivers sustained value. Pablo champions a multifaceted approach to this measurement, beginning with a strong focus on user-centric metrics. “Adoption and usage are always important,” he states. “To measure this, you want to see how much users are interacting with the product, how often they are using it, and how much are users finding the product useful for their use cases.”
This necessitates establishing consistent touchpoints with users and implementing a robust technical infrastructure capable of capturing these interactions accurately. Key adoption metrics include product adoption rate, feature adoption rate, and time-to-first key action. Furthermore, actively soliciting user feedback for continuous improvement is integral to this process. Metrics such as time savings, usage frequency, and user satisfaction provide critical insights into the product’s perceived value and utility.
Beyond user-facing indicators, Pablo underscores the importance of monitoring technical key performance indicators (KPIs). “It is important to understand some of the technical KPIs as well, such as model accuracy, model drift, and hallucination rates,” he explains. “If the model is getting worse over time, that will ultimately be reflected in adoption and usage rates.”
This critical link highlights how the underlying performance of the AI model directly impacts the user’s experience and, consequently, their engagement with the product. For instance, monitoring for hallucinations or the generation of incorrect information by Generative AI models is crucial for maintaining user trust and output reliability. Additionally, Pablo notes, “Understanding compliance flagged issues is important and a key metric to track a successful launch,” emphasizing the need to monitor adherence to regulatory and internal governance standards post-deployment.
The ultimate measure of an AI product’s success, however, lies in its ability to deliver tangible business value. Pablo advocates for defining metrics tailored to specific use cases: “Once use cases are well understood, measure against whatever the use case merits, be it time saved, workflows optimized, or new customers served.”
Critically, he poses the question, “What sort of improvement are you using from a business perspective? Are there any KPIs that can measure real business impact? Tying back impact to core business goals is often what secures continued investment and drives further adoption.”
This focus on demonstrating a clear link between AI deployment and core business objectives is essential for justifying ongoing investment and scaling AI initiatives within an organization. This comprehensive measurement framework, which connects technical model health to user engagement and then to core business KPIs, creates a vital feedback loop. It ensures that technical excellence is not an end in itself but a means to achieving meaningful business outcomes.
Moreover, the act of clearly demonstrating this impact serves a strategic purpose beyond mere reporting; it builds confidence within the organization, justifies continued investment in AI, and can catalyze broader AI transformation by proving its value. Pablo’s approach to measuring AI product success can be visualized across three interconnected tiers. The first tier, user-centric metrics, includes adoption rates, usage frequency, user satisfaction indicators like CSAT and NPS, task completion times, and time savings.
The second tier focuses on technical model KPIs, such as model accuracy, precision, recall, F1-score, model drift, hallucination rates for Generative AI, error rates, and compliance flags. Finally, the third tier, business impact KPIs, encompasses metrics like time saved, workflows optimized, new customers acquired, cost reduction, revenue growth, ROI, and alignment with core business goals.
The field of AI is characterized by rapid advancements and an evolving regulatory and security landscape, often necessitating strategic pivots during product development. Pablo shares an experience leading an LLM-powered platform where such a pivot became essential. “Midway through developing our LLM-powered platform,” he recounts, “we faced a major strategic pivot. One of our key value propositions was offering access to multiple LLMs, including a popular model that differentiated us from more restrictive, commercial tools.”
However, this differentiator was abruptly invalidated when “new security concerns flagged by compliance required us to vacate that model entirely.” This situation presented a dual challenge: a technical hurdle of removing the problematic model and integrating a viable alternative without disrupting the broader system, and a strategic hurdle of rethinking the product’s go-to-market message given the loss of a key feature. Such pivots driven by unforeseen security or compliance mandates are becoming increasingly common in AI development, demanding agile leadership.
In response to this critical challenge, Pablo proposed and led a decisive course correction. “To address this, I proposed a pivot: integrating another model from the same secure infrastructure we were already using, which met our compliance standards.” This solution demonstrates not only problem-solving acumen but also a deep understanding of the existing technology ecosystem, allowing for a potentially faster and less disruptive transition.
His leadership involved several concurrent actions: “I led the effort by validating the technical feasibility, aligning with the engineering team to make the change with minimal disruption, and securing compliance approval.” This showcases a blend of technical oversight, team coordination, and adept stakeholder management. The ability to quickly identify and integrate a compliant alternative from an existing, trusted infrastructure was a significant tactical advantage, minimizing delays and leveraging established security protocols.
A cornerstone of navigating this pivot successfully was transparent and effective communication. Pablo emphasizes, “I also managed all stakeholder communication, ensuring clarity around what was changing and why, and reframing the user value proposition for launch.” In times of strategic change, clear, consistent communication is vital for maintaining stakeholder alignment, managing expectations, and mitigating uncertainty.
Equally important is managing user experience expectations when core functionalities or underlying technologies shift. The fact that the product could still launch successfully and receive strong user feedback, despite the removal of a differentiating LLM, suggests that the core value proposition was robust and not solely tethered to one specific technological component. This strategic decoupling of user value from a particular model allowed for remarkable resilience and adaptability.
The project’s successful launch, despite the significant mid-development upheaval, yielded strong internal support and positive user feedback. For Pablo, this experience served as a powerful reinforcement of key leadership principles: “This experience reinforced how important it is to stay flexible, lead with clear communication, and always keep the core value to the user front and center—even when the underlying technology has to shift.” This case is a compelling illustration of the agility required in modern AI product development.
Looking towards the horizon of AI development and integration, the role of AI leaders is set to evolve significantly, demanding a more holistic and strategic perspective. Pablo envisions that future AI leaders “need to have a holistic view of the entire organization to understand where AI can be deployed and used most effectively.” This extends beyond mere technical deployment to encompass a deep understanding of how employees and users interact with AI, where they derive the most value, identify areas of friction, and pinpoint opportunities to enhance productivity.
This perspective aligns with the broader trend of data leaders needing to intricately connect AI strategy with overall business objectives and organizational transformation. Two foundational pillars underpin an organization’s ability to stay at the forefront of this technological wave. “To stay at the forefront of data analytics and machine learning, the most important aspects are data and talent,” Pablo asserts.
Regarding data, he emphasizes the necessity for a “strong understanding of data governance, data management, and data engineering. Ensuring an organization’s data is usable, understandable, and reliable is paramount to being able to harness data analytics.” Indeed, robust data governance and high-quality data are critical success factors for any AI initiative, as AI models are only as effective as the data they are trained on and operate with.
The talent dimension is equally critical. “All of this needs the appropriate talent,” Pablo states. “To attract and retain talent, it is critical to provide employees with the tools they need to succeed. This includes continuously procuring the best tools (database management tools, cloud-based databases, coding assistants, etc.) to extract and use data. This also requires a strong training program, as well as support for employees in technical and business roles.”
The scarcity of specialized AI talent and the continuous need for upskilling existing workforces are significant challenges facing organizations globally. Addressing this requires proactive strategies in recruitment, retention, and internal development. Beyond these foundational elements, Pablo highlights a crucial evolving skill for AI leaders: the ability to demonstrate value and act as a sophisticated intermediary.
“Being able to showcase how data is valuable to your organization is the most important skill an AI leader needs to have,” he notes, especially in an era of cost scrutiny. This leads to a pivotal role: “AI leaders will need to be translators: aligning technical possibilities with business priorities, navigating ethical and regulatory considerations, and building frameworks that scale responsibly.” This “translator” function is becoming increasingly vital as AI systems grow in complexity and pervasiveness, requiring leaders who can bridge the understanding gap between deep technical teams and business executives, as well as interpret complex ethical and regulatory landscapes.
Furthermore, these leaders will “play a key role in change management, as AI continues to reshape roles, workflows, and decision-making,” guiding organizations through the human and operational shifts accompanying AI adoption. This vision of the AI leader as a strategic enabler, data champion, talent cultivator, value articulator, and adept translator underscores a significant evolution from purely technical oversight to comprehensive organizational leadership in the age of AI. The strategic shifts Pablo anticipates—robust data governance, proactive talent development, and value-focused communication—are not just desirable traits but are becoming fundamental prerequisites for any organization aiming for sustained leadership in the AI-driven economy.
The journey from a mathematically inclined problem-solver to a strategic AI leader in the demanding financial services sector, as exemplified by Pablo, underscores a critical evolution in leadership. His approach, characterized by a rigorous analytical foundation, a pragmatic understanding of business value, and a deep commitment to responsible innovation, offers a compelling blueprint.
By consistently balancing technical feasibility with business alignment, proactively managing risk and compliance, and championing cross-functional collaboration through transparent communication, he navigates the complexities of AI deployment. As AI continues its transformative march, the future will increasingly depend on leaders like Pablo who not only possess technical acumen but also excel as strategic thinkers, data stewards, talent developers, and crucial translators capable of bridging the gap between intricate technology and overarching business imperatives, thereby shaping a future where AI is leveraged responsibly and impactfully.
