Artificial intelligence (AI) is pervasive: decisions made by AI influence our choices and our lives every day, often without us realising. These decisions, driven by algorithms, are changing society. What does this mean for us as project managers, and how should we adapt? What is AI?

What do we mean by AI? First, we must distinguish between classic or pure AI (‘thinking computers’) and applied AI (the application of specific algorithms to develop systems that exhibit machine intelligence). The latter is what most people understand by AI, and what I am focusing on here.

Applied AI is based on a wide range of algorithms. In the early years, computer science developed decision trees, rule-based systems and genetic algorithms. As computers became more powerful, computer scientists developed voice recognition and natural-language processing, built on Bayesian probability and classification systems. Face and number-plate recognition systems are now widely used, exploiting advances in image analysis, classification and machine learning through artificial neural networks. The field of robotics is experiencing explosive growth, as algorithms have been developed that allow robots to move through their environment, learning through self-directed trial and error.

The modern economy is now built on these algorithms. They combine to help you choose your next online purchase (influenced through social networks, targeted advertising and recommendation systems). They listen to your commands, take your payment, route the request through the internet, and instruct warehouse robots to dispatch your purchases through an optimised just-in-time distribution system that delivers to your front door. It is a chain of great complexity, built on algorithms and executed millions of times a day.

What does all of this mean for project managers? I believe the application of AI will change our profession in three ways:

• our clients and society will have new expectations;
• we must design and work with new systems and information flows;
• most significantly, AI will change our team dynamics and the skills we require.

New Expectations

The environment in which we deliver projects will change radically. Clients will expect projects to deliver products and solutions with a greater proportion of AI-enabled systems, requiring more AI expertise in teams. They will also expect better information on project progress. Instead of weekly or monthly progress reports summarising project performance, project managers will increasingly need to understand and provide information on the full state of the project in real time. Project teams will be expected to fully embrace the new ways of working. This is already happening in the digital and high-technology sectors, and is broadening into more traditional sectors, such as construction.

Organisations that choose not to work with AI systems will increasingly find themselves at a competitive disadvantage; the cost and schedule advantages of AI will become overwhelming in highly competitive markets. New Systems

AI techniques will impact every stage of the project delivery life cycle.

Natural-language analysis tools will be more widely used. They will automatically read and analyse requirements documents and drive other tools that generate project activities and schedules. At first, these systems will be rudimentary and offer little value over existing experienced project schedulers. But as they improve through machine learning – and we learn to work with them – they will get better.

With pervasive, real-time data flows, project staff will no longer copy timesheet hours, progress information and financial data from one spreadsheet to another. Information will flow faster and further, and fewer staff will be employed in clerical roles. With automated data flows, AI techniques will be able to detect contradictions in the data faster and earlier – for example, a mismatch between claimed progress and invoice. Better-quality data will drive improved decision-making.

Today we use scheduling and optimisation algorithms to allocate staff to projects and predict end dates. We have scheduling tools that can analyse and report on schedule deficiencies. In the future, AI techniques will enable schedules to become richer and more expressive over the set of all possible outcomes, taking into account past experience, events and probabilities. Schedules will become automated, and no longer managed in spreadsheets.

AI will generate and demand ever increasing amounts of data on project progress and performance. On a construction project drones with cameras and image recognition systems could be used to report on the percentage completion without relying on contractors’ interpretations and reports. Similarly, agile software projects exploit ‘bots’ that automatically compile code and perform regression tests. We have the opportunity to do the same in project management: to automate our modelling and reporting systems.

Using expert systems will enable project managers to manage change more effectively. These systems will apply decision trees and organisation-wide rules to identify change earlier and more systematically. Automated workflows could then take the team through the change management process. Baseline management will become more effective and disciplined, leading to better project outcomes and greater trust across the supply chain. At a portfolio level,we can expect to exploit machine learning techniques (now freely available through open source software ) to assess project cost and performance data. These systems will identify failing projects within an organisation’s portfolio earlier, enabling us to intervene and restore delivery confidence.

As automated systems take care of more tasks, we will work at a higher level of abstraction, becoming project designers. What does this mean to project teams? New Skills

Project teams will need to develop new skills and learn to work with AI technologies effectively: not only designing them, but governing them too.

A new generation of project control specialists will take on roles designing, training, integrating and monitoring AI systems. Project managers will need to strengthen their leadership skills, to coach and build trust in these systems, and to manage any conflict they generate. Sales, commercial, finance and operations staff will need to work more closely with project delivery teams, as accelerated data flows and analysis will remove boundaries and expose weaknesses.

What else could this mean for project teams? We could envisage AI making decisions without taking stakeholders with them – and humans may resist this. The project manager exploiting deep AI systems that make such decisions will need to understand what they are doing, to form a relationship with them that means they can trust those decisions. Project teams will insist upon AI that explains the rationale for its conclusions. Ultimately, the project team must take control should the AI not perform as expected. We must build in resilience to handle AI system failures.

Effective AI will be integrated with people, always telling us what is happening, and why conclusions were reached. We will need to understand what the systems are doing and how the data is being processed. To do this, our models of project delivery and information – our descriptions of how the work flows – must get better and more precise. We must stay in control at all times. Conclusion

In the project management profession we have a great opportunity before us, to incorporate AI techniques into our ways of working, and in so doing, to benefit society in an ever-more AI-infused world.

We have a choice. We can be spectators, observing the ever-encroaching application of AI, then play catch-up at a competitive disadvantage. Or we can lead our profession, by embracing and incorporating AI into our project delivery. This calls for a shift in both culture and mindset. Applying AI to project delivery has the power to teach us about ourselves and the discipline of project management, and to develop our skills and profession for the greater good. It’s time to seize this opportunity.

(c) James Lea 2018

If you would like to explore how your organisation can apply AI techniques to gain competitive advantage, please get in touch.

• AI