Lower for longer has become the prevailing mind-set among the executive corridors of the industry. In an environment where the light of dawn is not visible yet, both survival mode and delivering outperformance mean that a new set of tactics needs to evolve from the wounds that were created over the past three years. During the downturn operators renegotiated their contracts with service companies, streamlined internal efficiencies and leveraged technologies with which they were familiar. The returns from these strategies are starting to plateau. In a lower-for-longer environment, how can operators be profitable and drive outsized shareholder value?

For example, development of core acreage in the major shale basins has matured, and the need going forward will be to develop Tier 2 and Tier 3 acreage as cost-effectively as possible. These reservoirs are more difficult to produce, and lifting costs are more expensive. Thus, asset teams need to sharpen their pencils when it comes to delineating the reservoir, finding sweet spots, shortening drilling time and identifying optimal well spacing and completion parameters.

However, corporate mandates to cut costs run counter to collecting the subsurface measurements and performing the geoscience interpretation that is required. It is exactly the wrong time to make do with less.

Artificial intelligence (AI) is considered an essential technology to drive the new economy. For oil and gas companies that need to do more with less, AI is an important tool that should be in every operator’s arsenal. The right approach with AI is to adopt a barbell strategy of short-term cost savings combined with longterm moonshots. Initial implementation should have substantial, measurable cost savings delivered within one to three months. This is important not only to satisfy budgets, but it compels data science teams to look under a few more rocks than they normally would have to find the right use cases that can help the asset teams. Moreover, if the short-term implementation has measurable cost savings, then the asset teams, data science groups and service providers develop the needed team chemistry—and a better understanding of corporate priorities—to pursue long-term moonshot projects that can deliver shareholder returns in big chunks.

Data are key

One of the hidden obstacles in using AI is that its implementation requires a significant amount of training data to calibrate models to local parameters. In many cases, operators do not have the “Big” in Big Data to leverage AI technologies in any meaningful way. However, this field data can be found from consortia that pool data among operators as well as service companies, which can bring a rich database to the table. Partnering with a service provider that brings its own database means the operator doesn’t have to worry about providing data it might not have nor a budget to obtain them.

The leading AI companies don’t just aggregate data from multiple operators across various basins. They can use well data from one basin and project those AI models into entirely different basins or frontier acreage. They can squeeze more insight out of the data, whether it is about extracting sonic velocity from drilling data or upscaling seismic to near-log scale resolution. These technologies reduce overall data acquisition expenditures while advancing the operations of drilling faster, completing more efficiently and increasing production.

Faster ROP across the rig fleet and lower logging costs are both measurable. One interesting example of a high-value opportunity is that one major has qualified an AI-based service to obtain sonic logs in deep water without running the traditional LWD tools. Using the technology for discrete and measurable use cases can help a company reduce expenditures, which more than pays for future impactful but longer term projects like optimal well spacing and reservoir delineation. Taken together, this barbell strategy is key to leveraging AI successfully so that companies can survive an extended downturn and outperform operationally.