The focus of this year’s Enaxis Leadership Forum was “Developing Digital Capability”. The event saw business and technology leaders from more than 40 Fortune 500 organizations across diverse industries come together for a meeting of the minds and to share their own experiences with going digital.
In the panel entitled “Advanced Analytics & Artificial Intelligence: What Every Executive Needs to Know”, participants discussed critical knowledge for executives embracing these technologies.
The panel included:
- Richard Baraniuk, Professor of Electrical & Computer Engineering, Rice University and Founder of OpenStax
- Case Carstensen, Director of Data & Analytics, Baker Hughes, a GE Company
- John Rowe, Partner, Enaxis Consulting
- Hardeep Singh, MD, Chief, Health Policy, Quality & Informatics, Baylor College of Medicine
- Jeremy Graybill, Data Science and Advanced Analytics Leader, Anadarko, who played the role of moderator
Advanced analytics promises to unleash previously untapped insights to ignite revenue growth and significantly lower costs. Artificial intelligence is now being represented as a logical next step along the advanced analytics continuum. This panel addressed the status of these initiatives and what businesses need to understand prior to progressing both of these often-misunderstood technologies.
Here are some of the key takeaways from this panel discussion:
- AI is at Peak Hype, But There’s Much to Be Gained from Machine Learning
Artificial intelligence is a term that is often misused and over-hyped. Many organizations tend to confuse artificial intelligence and machine learning. According to Richard Baraniuk, we are a long way away from true artificial intelligence. Gartner too, seems to think so. According to the 2017 Gartner hype cycle for emerging technologies report, true artificial (general) intelligence more is than 10 years away. That being said, significant progress has been made in the main building block of artificial intelligence, machine learning. Machine learning algorithms detect patterns and from those patterns, predict or optimize performance. The value of selecting and using such algorithms is that they rely on historical data, rather than on explicit programming instructions. As new data is fed into the model, the algorithms adapt to improve efficiency over time. Figure 1 shows a high-level explanation of the different types of machine-learning.
Figure 1: Different Forms of Machine Learning
- Embrace a Holistic View
Far too often, executives spend hundreds of thousands of dollars (and in some cases, millions) on the latest advanced analytics technologies, only to end up disappointed – and in some cases looking for another job – because the expected ROI is not realized. This is typically due to poor strategic alignment, poor implementation and execution, or both. To be successful, a holistic approach to analytics is a must; an approach that includes a vision and strategy, governance, people, process, technology, and data. Figure 2 shows these seven key aspects of the Enaxis Analytics Operating Model (Click here to learn more about the seven key considerations to establish and evolve analytics capabilities).
Figure 2: Enaxis Operating Model for a successful analytics solution
- Keeping Pace with Emerging Technologies
Machine learning is growing faster than it can be implemented, primarily due to a lack of talent and organizational challenges on how to adopt these technologies. Many companies have a dedicated team that focuses on finding and evaluating emerging technologies that can be game-changers in their industries. There are many different sources of emerging technology collaborations:
- Vendors – Conduct POCs with start-ups and other organizations that are developing cutting-edge technologies and looking for industry partners to collaborate with.
- Academia – Join academic consortia or bring in interns (typically PhD candidates) that are conducting research in your function or industry.
- Crowdsourcing – Collaborate with crowdsourced solution providers that harness the power of the crowd to unleash possibilities that come from a variety of different industries. This also helps increase the bandwidth and throughput of internal analytics professionals.
- Agile is the New Delivery Model
Over the last decade and a half, Agile methodologies have gained enormous traction in the world of software development and, more recently, in other functions and industries, as well. But how does one apply Agile principles in an analytics setting? This is typically done by taking a customer-centric approach and defining customer “user stories” that reflect the needs of the customer. Taking an iterative and incremental approach with fast feedback from customers and stakeholders at the end of each iteration is critical to success, especially in an environment where market dynamics and customer requirements change frequently. Through this approach, the primary focus of all efforts is delivering value to customers.
Being customer-centric involves having a robust stakeholder engagement and adoption model. Such a model identifies and includes key stakeholders early in the analytics delivery lifecycle to create analytics solution champions and avoid “black-box” rejection of those solutions. More on this in our upcoming blog, “The Agile Change Curve: Keys to Rolling Out a Successful Analytics Engagement & Adoption Program”.
Click here to learn more about our Data & Analytics service offering. Want to continue the conversation? Contact us at firstname.lastname@example.org, and read our seven-part blog series on our Analytics Operating Model.