| | JULY 20199CIOReview"colleague," Marvin. Named after the robot in A Hitchhiker's Guide to the Galaxy, SWN created Marvin entirely in-house, cost-effectively, within about six months, using artificial intelligence (AI) to process thousands of invoices a year. Facing a backlog of over 2,000 invoices, Marvin took over from his D&C colleagues. With some programming tweaks, this new digital `associate' got better, faster, and more precise. An SWN planner collaborated with IT to develop the SWN "FTC" app to capture field tickets more quickly. Now, users can phone to select a field from a menu via their cell phones and take a snapshot of the invoice to capture it. Marvin analyzes this input and matches field tickets to invoices more efficiently, emailing a complete report to approvers in much less time and with greater accuracy than would be "humanly" possible.The payoff has been impressive. Marvin's autonomous ability to process about half the 60,000 D&C invoices is saving SWN an estimated $1 million per year. Invoices have never been treated so quickly, which reduces reliance on accruals for capital estimation. More importantly, the success of SWN's Marvin program is stimulating more ambitious projects intended to capture even greater efficiencies across its business operations.Use a Decision-Matrix. Developing Marvin prompted broader thinking about how other digital solutions could be deployed. SWN filters out projects using a decision matrix to estimate perceived strategic value and degree of difficulty to implement. Determining a framework is an excellent first step in considering how and where to start. Consider whether the digital project might be better served using external consultants or service providers, or in-house staff, or some combination of both, to execute the proposed solution. A decision matrix is a clear protocol for assessing the scope and potential investment decision. CAPTION: Is your digital project a Tier-1 ("high value" / "high availability")? If it can be more efficiently outsourced to experts with access to marketplace-wide technology, e.g., a completion optimization project, this might be best. A Tier-2 "lower value, low availability" project could serve as a specialized training project for new engineers or interns interested in digital. Involve People. Companies have learned the value of diversified teamwork. Placing data scientists, together with operations or other functional teams, improves the design, execution, and understanding of digital. Creating a collaborative, supportive culture that sees failure as a learning opportunity is critical to building viable digital projects faster. Mixing up teams to produce early wins can showcase how digital -- machine learning, AI, big data analytics, IoT ­ actually makes peoples' jobs easier. Communicating regularly about progress on digital helps stoke demand and sustains interest in longer-term transformation. Measure Success. Articulating how digital supports the company's business strategy, its goals and performance make a difference. Southwestern and companies like Micro Focus noted that digital interest rises quickly if management sees a return. Before undertaking a digital project, model what it will do and how it can impact key performance indicators. Link execution to strategy and KPIs to digital efforts. Credibility is crucial for more ambitious digital projects. A small project at SWN became an eye-opening experience for digital's potential. Marvin's success is stimulating new ideas for digital transformation within the company. Across the industry, the digital party is constantly re-energizing. Seema Menon
< Page 8 | Page 10 >