PwC’s 2020 AI Predictions report included a fascinating survey result: across all sectors of the economy, 20 percent of executives at companies with AI initiatives are planning to deploy AI enterprise-wide in 2019.
Sound improbable? It shouldn’t. First, remember that this survey was only of companies that already have AI initiatives. It excluded the vast majority of companies which, regrettably, haven’t even started with AI yet. Next, consider just how much value AI is adding right now.
Here are some of the examples I’ve seen from capital project owners, financers, builders, or all three:
Many companies working with capital projects have what AI needs most: a treasure trove of historical data. It’s historical data that make AI so powerful, by teaching it how to interpret new data.
Every capital project is unique, but one of the things that sets AI apart from older automation tools is that it can draw insights from complex, irregular data. So for capital projects, AI can find the significant patterns in historical examples, even if they’re not identical, then apply them to a new project to generate time- and money-saving forecasts and analysis.
The top challenges that AI leaders cited in recent PwC research were ensuring that AI systems are trustworthy, training current employees to work with AI, and managing the convergence of AI with other technologies, such as the IoT and enterprise software.
These challenges all have answers: follow the principles of responsible AI to build trust; adjust workforce strategies for AI, with an awareness that only a few users will have to be AI experts; and get your data cleaned up so you have a solid foundation for that technology convergence — and so that AI can use your data in the first place.
That may sound hard, but it shouldn’t be. You start with AI to solve a small, specific business problem — such as one of the four listed above — get it generating ROI quickly, then scale it up. That’s what the leaders are doing.
You may also want to look at explainable AI, which offers transparency into its decision-making. You can then monitor and validate AI’s algorithms for accuracy.
The transformation tomorrow
Imagine insights and forecasts drawn from geography, the financial markets, the properties of raw materials, the skills in your labor force, the track record of subcontractors and vendors, the regulatory environment, even the weather. Imagine sensors everywhere in every stage of the project lifecycle, feeding data into AI for constantly evolving, real-time insights.
The result would be a remarkably powerful modeling tool. With the more accurate forecasts and better planning that it could enable, cost overruns and time delays might be a thing of the past.
That transformation of the industry isn’t for today, though I do think we’ll get there. What is for today is specific use cases, where AI is already able to cut costs, improve forecasting, and support better decision making.
It’s worth the time to start looking at some of these use cases to add value right away, while laying the groundwork for fully AI-enhanced capital projects tomorrow.