Case Study: ASUG and Chevron Phillips Chemical on Driving AI Adoption Enterprise-Wide
Artificial Intelligence (AI)’s potential to accelerate growth and transform the face of the energy sector can no longer be ignored, especially for industries that are built upon complicated processes, and rely heavily on data and analytics for innovation. Within the energy sector, where efficiency, safety, and innovation are paramount, the integration of AI holds immense potential.
However, getting buy-in from C-Suite and the frontline can prove to be challenging, along with identifying the right digital decisions that align with overall business objectives. We spoke with Martin R. Gonzalez, Innovation and Technology Principal at bp, Laura Tibodeau, Global Digital Innovation & Transformation Leader at ASUG (Americas' SAP Users' Group) and Ra Inta, Head of Data Science at Chevron Phillips Chemical Company to share their insights on how to navigate the realm of cutting-edge AI in the energy sector.
Read as they offer individual perspectives into the strategic deployment of AI, illustrating how organizations can leverage technology to unlock new opportunities, drive sustainable growth and shape the future of energy operations.
How are you employing cutting-edge AI within your operations?
Laura Tibodeau, ASUG: Cutting-edge AI technologies offer a wealth of opportunities to revolutionize operations within the energy industry.
For us, harnessing AI for operations relates to:
1. Predictive Maintenance (Versus Time-Based Maintenance)
AI can analyze vast amounts of data from sensors and equipment to predict potential equipment failures before they occur. By harnessing predictive maintenance systems (powered by AI), companies can minimize downtime, optimize maintenance schedules, enhance reliability and safety, extend the lifespan of critical assets, and optimize maintenance costs.
2. Cyber Security
AI technologies enable cyber security functions to proactively identify and respond to anomalous activities and potential security threats through Machine Learning Algorithms, Behavior Analytics, User and Entity Behavior Analytics (UEBA), Anomaly Detection and Real-time Monitoring and Alerts.
3. Asset Management
AI streamlines asset management processes by providing insights into the performance, health, and optimal utilization of infrastructure assets. By employing AI-powered asset management tools, companies can make data-driven decisions on asset maintenance, replacement, and investment, leading to improved operational efficiency and cost-effectiveness.
Martin R. Gonzalez, bp: Starting with Generative AI (Gen AI), we are learning to successfully deploy and derive value from large language models. To achieve the high level of trust required for an LLM-based application, we are employing retrieval-augmented generation and working to balance the advantages of a foundational model with the imperative to stay true to our internally curated knowledge stores. RAG is particularly intriguing due to its potential to use vector databases more broadly than the automation of contextual relationship mapping. Additionally, we're exploring other technologies capable of extracting relationships from various knowledge sources or datasets. By using knowledge graphs, reinforcement learning, causal AI, and hybrid physics/ML modeling, we help engineers and operators to understand context and make predictions. We use these technologies, individually and in combination, to support decision-making in the field, enabling our frontline workers to spring into action more quickly.
Ra Inta, CPChem: At Chevron Phillips Chemical (CPChem), we are lucky to operate within an environment that encourages innovation in the digital space. We have an internal team that builds customized machine learning products and pipelines. These applications cover various aspects of our business, ranging from optimizing chemical processes to enhancing supply chain management and forecasting accuracy. We have reached a mature stage in traditional machine learning workflows, including our approach to Machine Learning Operations (MLOps). Currently, we are scaling up our internal capacity for building Gen AI-based applications. While many of these applications are still in the proof-of-concept stage, several have progressed to demonstrate tangible business value. As with most enterprises, we possess a substantial volume of data, both structured and unstructured, brimming with untapped potential. Under our Performance by Design initiative, whereby every member of the company is encouraged to submit ideas, a significant portion of these ideas relate to Machine Learning and/or Gen AI.
How can you effectively communicate a digital innovation strategy, with its tangible and intangible benefits, to C-Suite and the frontline?
Martin R. Gonzalez, bp: Inevitably, we must rely on our leaders at the highest level to establish the strategy for our enterprise. I have been fortunate to have an executive team and board that are forward-thinking and believe in leading our industry by showcasing the value that can be attained through the adoption of digital technology.
The goal, however, is to pursue the broader transformation stemming from the collective impact of individual project. While value must be clearly articulated and measured for each project, the extent to which each effort moves us toward our strategic goals is also considered in project resourcing and prioritization.
Laura Tibodeau, ASUG: Effectively communicating your digital innovation strategy, along with its tangible and intangible benefits is critical for driving organizational alignment, engagement, and successful implementation.
When communicating to senior executives, it’s important to frame your digital innovation strategy in terms of how it aligns with the organization's overall business goals, objectives, and long-term vision. Emphasize how the strategy will drive revenue growth, cost savings, market differentiation, and competitive advantage.
Additionally, involve frontline employees (subject matter experts) early in the communication process by directly soliciting their feedback, ideas, and concerns regarding digital innovation opportunities. Foster a culture of open communication, transparency, and collaboration to empower employees to contribute to the strategy's success.
Ensure that your communications are clear, concise, and consistent across all levels of the organization. Use multiple channels, including town halls, newsletters, intranet portals, and one-on-one meetings, to reach both the senior executives and frontline effectively.
Ra Inta, CPChem: Effectively communicating abstract strategies can be challenging, especially when the technology is new. For Generative AI, we strategically chose several proofs-of-concept that showcased potential value for various use cases across the business. These demonstrations highlighted techniques that could be applied to other scenarios. Additionally, they provided concrete illustrations of the potential risks associated with the new tools, presented in a context our business leaders could understand in a tangible, and actionable, way.
How do you align digital decisions with corporate values and market demands?
Ra Inta, CPChem: We utilize an internal system to assess the value-to-effort ratio of potential use cases. The challenges are determined by market dynamics, operational demands, and corporate values.
For example, at CPChem, we prioritize projects that align with our commitment to safety and sustainability. Hence, initiatives incorporating these values, ceteris paribus, are given precedence. Our digital decisions are driven by the potential value and our current technical capabilities. It's worth noting that this can sometimes be a moving target; showing what's possible can often stimulate further demand.
Martin R. Gonzalez, bp: When developing digital products, it's crucial to explicitly connect the goals of a project to the objectives of the organizational unit and the overall enterprise. This requires implementing a system for managing nested objectives and key results across every digital project.
The investment of time in creating this structure not only facilitates resource prioritization but also fosters clarity of purpose for individual employees who seek to know their time is meaningful beyond just their day-to-day tasks.
Laura Tibodeau, ASUG: Aligning digital decisions with corporate values and market demands is essential for ensuring strategic coherence, organizational integrity, and competitive relevance. Start by clearly defining and articulating your organization’s core values, mission, and vision. Take time to understand the principles that guide decision-making, shape culture, and define your organization's identity.
Next, take the time to conduct thorough market research, which includes competitive analysis and customer insights. This step is crucial for understanding current market trends, consumer preferences, and industry dynamics. Finally, engage your key stakeholders (such as employees, customers, partners, and industry experts) in the decision-making process to gather diverse perspectives and valuable insights.
What methods do you use to establish a transparent and resilient framework for AI governance?
Martin R. Gonzalez, bp: For security and ethics purposes, we have established an internal policy governing the use of Gen AI. We provide mandatory compliance training and maintain an active digital security organization that identifies and addresses vulnerabilities.
In terms of decision-making regarding AI initiatives, we follow a product-led model where the problem or opportunity is clearly articulated before a solution is selected. AI-based solutions have proven to be advantageous in terms of cost-benefit balance and time to value, largely due to the efforts of our innovation acceleration team. They showcase the practical application of AI to improve user workflows and generate business value. However, we don't view AI as an end in itself, and so, we don’t have a governance process dictating its application across our businesses.
Laura Tibodeau, ASUG: Establishing a transparent and resilient framework for governance is essential to ensure ethical, accountable, and effective use of artificial intelligence technologies within an organization. I recommend that the AI policy handbook encompass thorough AI governance policies and procedures. It should clearly define the guidelines, principles, and standards for the ethical deployment of AI technologies and address crucial aspects such as data privacy, transparency, accountability, bias mitigation, and adherence to regulatory requirements.
Next, it is critical to implement strong data governance practices to safeguard the quality, integrity, and security of data utilized for training the AI models. Data quality assurance processes that track data lineage are essential to uphold data accuracy, relevance, and compliance with data protection regulations.
Finally, foster transparent communication by regularly reporting on AI governance practices, outcomes, and compliance with established guidelines. Provide stakeholders with updates on AI initiatives, risks, controls, and performance metrics to promote accountability, trust, and alignment with organizational values.
Ra Inta, CPChem: We’ve established a cross-functional Gen AI Task Force composed of representatives from Legal, IT Risk, Security, Compliance, Technical/Data, Organizational Change Management, Corporate Communications, and Leadership. This helps forge a cohesive policy governing the usage of Gen AI at CPChem, which outlines key risks such as data exfiltration, bias, and hallucination, and outlines approval and usage processes.
We have also created a Gen AI Community of Excellence which serves as a bidirectional conduit for communication between the business and the Task Force, fostering transparency into how decisions are made. We firmly believe that education is a key pillar in the successful adoption of Gen AI and hence have been tasked with implementing educational initiatives at an enterprise level.
Can you share top techniques to manage fear on the frontline to ultimately drive equitable outcomes?
Martin R. Gonzalez, bp: It's extremely important to engage users in the product development process, including direct participation in proof-of-value exercises integral to innovation and technology development. In exercise, users contribute to shaping how the technology is ultimately deployed and validate that the envisioned solution is likely to be embraced in the field.
In all cases, the objective of digital product deployment should be to reduce the transactional and cognitive load for workers. Engaging users in creating the product interface is also crucial for ensuring a positive user experience (UX), as any missteps here could lead to extended support needs post-initial deployment. The closer you get to the frontline, introducing a new tool becomes increasingly challenging, and users who are involved in early development end up becoming champions during deployment.
Ra Inta, CPChem: Treat Gen AI as a new tool that excels in specific applications involving unstructured data. Nothing more, nothing less. Just like a nail-gun is ideal for constructing a timber-framed house but unsuitable for crafting a Faberge egg. Adopting this mindset helps shield against the constant hyperbole – and the inevitable backlash - when we realize we're probably still far from AGI (Artificial General Intelligence). Instead, we are much closer to leveraging Gen AI effectively with unstructured data than we previously anticipated. Keep your eye on the ball and the rest will take care of itself.
It is also imperative to acknowledge the concerns of those who fear that Generative AI threatens their job security. We emphasize that the most effective utilization of these tools is as decision aids for humans. We do not view the replacement of human labor as either feasible or productive, especially within our workforce, at least not within the next five years.
Laura Tibodeau, ASUG: Provide ongoing education and training programs to frontline employees to increase their understanding of AI technologies, their benefits, and potential impact on their roles. Addressing misconceptions (earlier versus later) and building AI literacy will help alleviate fear and resistance towards the new technologies.
It is also important to implement robust change management processes to support the adoption of AI technologies by frontline employees. Provide resources, support networks, and training to help employees navigate changes, address resistance, and build confidence in using AI tools effectively.
Join the conversation in Houston, 2026!
Learn how to overcome adoption challenges and align digital decisions with business goals with attending bp and CPChem experts, who will share cutting-edge strategies to drive efficiency, safety and sustainable growth.
Register your place today and ensure your AI strategy is built for the future!
