AI enables dynamic transformation

Asset managers are finding themselves facing rapid transformation due to technological innovation, market economics, political, societal and regulatory changes, as well as increased global and local competition. For global firms especially, the rates of change combined with large investments in various asset sectors across a diversified geographical scope mean large volumes of information must be constantly acquired, absorbed and processed.

Change also means high levels of uncertainty, where shifting relationships between causes and effects are difficult to extract from the flood of information. Uncertainty also means higher risks, whether financial, technological, political, regulatory or caused by disasters and pandemics.

Artificial intelligence and machine learning can analyse large amounts of data and detect patterns in situations where humans are defeated by the volume or complexity of information. Furthermore, AI/ML can create significant operational efficiencies in complex infrastructure assets – for example, it can help organisations go from being reactive to proactive and take advantage of global business opportunities or cut costs on the supply side.

We surveyed more than 50 decision makers and 40 data scientists at global infrastructure organisations. Certain common features emerged:

Asset managers need to make strategic resource allocation decisions in a dynamic global environment. This includes taking into account the ESG impact of their investments.

The resilience of operations is increased by a more complete understanding of the external factors influencing projects as well as existing portfolio companies.

Currently, no predictive software exists to inform decisions for asset managers, corporations, multilaterals and construction companies based on global data.

Specific to infrastructure, three user classes emerge:

• LPs and governments want to know the extent to which pipeline infrastructure projects are in keeping with sustainable development aims.
• Infrastructure investors are seeking screening and due diligence tools for early warning for large projects and existing portfolio companies.
• Construction companies are seeking advanced data and AI analytics to discover project opportunities and related project risks.

Greenfield projects

When large infrastructure projects fail, it is estimated that one-third of the causes of failure are due to unforeseen risks and events. Accounting for different risk perspectives and incorporating accurate and reliable forecasts at early stages can help risk-management practices by ensuring decisions are based on more complete and prospective intelligence.

Asset and project managers have the most effective control in the early stages of the project and need to be vigilant throughout the asset lifecycle. As project phases proceed, the costs increase and the managers’ ability to influence the project decreases. Advanced data and AI can be used at the earliest stage of project scoping to inform project-specific and portfolio-wide global strategy.

To help the infrastructure sector become more resilient, it is important to bring a project-centric portfolio view. To this effect, building the largest database of infrastructure projects in the world helps to predict a project’s status, benchmark similar projects, and monitor country and location risk pertinent to the project.

It also enables organisations to do the following:

• analyse emerging global opportunities in large infrastructure projects across sub-sectors in a data-driven and dynamic manner. For example, the Luhri Hydro Project, estimated the longest tunnel in the world, with a budget over $1 billion generating 775MW of power was initiated in 2010 and dropped in 2015 by the World Bank. A system was developed to help infrastructure investors and engineering, procurement, and construction (EPC’s) companies find early in the process the project would fail, helped find other similar projects in size and impact but higher success rate;
• predict price movements and influencing factors (for assets such as commodity prices, equities and foreign exchange). For example, by predicting price movement at different frequencies and factors driving them, investors and multinational companies procuring raw material (such as steel or copper) were able to build hedging strategies to reduce costs;
• move the operator from a reactive to a proactive mode, avoiding (or mitigating) major disruptive events. For example, multi-stakeholders involved in infrastructure projects were able to be aligned about external project risks and take mitigation actions earlier in the project lifecycle;
• inform a more complete view of value chain mapping, including global and local risks and identifying influencing factors related to ESG or UN Sustainable Development Goals for suppliers or assets;
• create uncertainty-bounded predictions in context-specific analytical dashboards. Developers and small analyst teams can reduce the time to aggregate data and enable organisation-wide AI transformation by using domain knowledge, internal-external data, custom-intelligence and notifications for specific business use case.

In addition, data and AI capabilities can be used to help executives find high-value global opportunities and reduce uncertainties from external domains, such as the macro or local economy, finance, climate or geopolitics, or transform the effectiveness of large infrastructure projects by identifying challenges up front.

Reliable data and decision-support systems assist in the discovery, screening and location risk monitoring of infrastructure projects. They also help in providing early-warning risk alerts for active projects and site intelligence for greenfield assets.

Brownfield projects

Brownfield projects are also facing increasing risks due to volatile capital markets, regulatory and political changes and social empowerment. Many of these events can be predicted as trends (whether regional or global) by using AI, which gives asset managers the opportunity to hedge their respective risks or sell the assets in advance of a material event.

Over the past decade, events including the decommissioning of nuclear and coal power plants in various counties as well as the rise of Extinction Rebellion could have been predicted by following either social media or speeches made by regulators or members of governments. These events have cost investors hundreds of billions and are likely to impact more investors in other regions. AI solutions are proving invaluable to monitor the interplay between the sector, project, location and country risks to brownfield assets, and help investors find the right partners.

AI for decision support

Asset managers are being required to implement AI in order not to miss global growth opportunities or make inappropriate decisions on their existing portfolio companies by failing to connect diverse external data and events. As large infrastructure organisations learn to predict supply chain disruptions and dynamic fluctuations in global demand, this will help them to actively manage risk on their large infrastructure projects and assets.

These predictions will allow them to take mitigating actions before they are affected by major disruptive events, thereby increasing operational and financial resilience in a changing world. To bring value to individual asset managers, collaborative AI should include all relevant information. We suggest a process where managers, together with a larger group of experts and advisers, identify all the factors that could be relevant to the decision, particularly when the relationships are not easily defined.

Infrastructure executives responsible for complex multinational strategic decisions have extensive domain expertise and understand the strategic imperatives and objectives of their organisations. However, they tend to have personal biases induced by their own experiences. They often overemphasise internal data and underemphasise external data and are not able to continually monitor and absorb the vast amount of information about global demand and supply patterns and the many associated uncertainties.

Individual managers have widely varying experiences and viewpoints, which lead to substantial differences of opinion. Managers often cannot articulate their levels of uncertainty in a way that allows for a consensus to be reached.

AI can process masses of data to extract and combine weak signals and aggregate the results in a useful way. The unbiased nature of aggregating data and advanced AI-based analytical tools can provide a basis to level the playing field of information across silos within an organisation. Furthermore, AI methods can provide predictions at multiple time-horizons, quantify the related uncertainties, and track changes and shocks dynamically as they occur.

Given the increasing complexity and dynamic changes occurring in the infrastructure sector, AI is becoming an integral tool for infrastructure executives to properly assess the risks and opportunities on large projects and on their existing portfolio companies. Several infrastructure investors are in the process of implementing AI and many more will be examining how to do so as part of their key decision-making processes.

Saurabh Mishra is the CEO of enterprise software platform Taiyō; Jeffrey Altman is senior adviser at independent consultancy Finadvice; Uday Karmarkar is a professor at UCLA Anderson School of Management