• 
• 
• 

Tweet

• Future State Technology
• PMO & Portfolio Management

04 Jun, 2020

AI: A Threat Or Enabler Of The PMO Of The Future?

While many tend to put AI in the same bracket as automation solutions, such as Robotic Process Automation (RPA), it’s important to make a distinction between AI and automation solutions as follows:

• Automation solutions, be they based on a user writing code or using an automation user interface, involve automation of repeatable, rules-based processes that only require logical decision making. The solutions can include mimicking what a user would do manually in one or more applications or moving data between systems. 
• AI on the other hand seeks to duplicate more complex cognitive human functions, such as learning and problem solving.

This distinction is crucial. Automation is already commonplace across many functions, including PMO, while there is no widespread application of AI. AI, commonly known as machine learning is task specific and is trained on historical datasets, which are not commonly available in this context.

But in the context of PMO this does not mean AI should be disregarded. Chaucer identifies at least two areas where AI has an impact on PMO:

• Intelligent automation: AI-enhanced automation. Machine learning to provide decisions and case management, creating flexibility for end-to-end case management success.
• AI applications such as optimisation tasks, predictive analytics and deciphering patterns from previous experiences.

AI could soon play a crucial role in performing tasks to increase the benefits of PMO. These include:

• Improved insights with through increased data collection.
• Scenario planning for budget and scheduling
• Risk management.
• Forecasting performance issues.

This is especially true for AI solutions that cut out the typical requirements of large data volumes, such as genetic algorithms and reinforcement learning on simulated data. 

Although the idea of adopting AI in the context of PMO is at present something many organisations simply aspire to, there are strong parallels with the role PMO played when it began to use agile practices. 

Given its far reach within organisations, PMO can act as a kind of AI evangelist. It can help sustain the use of AI after implementation, while clearly guiding and informing stakeholders of its potential benefits.

To make a success of this, future PMO must acquire new data skills to be more effective when liaising with data scientists. By having a strong understanding of AI generated data analysis and the ability to communicate this to diverse audiences, PMO should be able to cut through distrust of automatically generated insights and end users’ resistance to change. 

AI will without doubt take the process of extracting insights from data to a higher level. But this does not make the human aspect of PMO redundant. 

Applications of AI are as highly performing as they are specialised. This can however, make them inflexible. To reuse or repurpose an AI model it is likely to require retraining, as it were, to solve new optimisation issues when they arise.

The ultimate decision maker will always be human. There will always be times when exceptions and anomalies cannot be correctly analysed by AI. This is where PMO experience comes into its own to resolve them. 

AI enhances the process of translating data into valuable recommendations that can be acted upon, and in so doing creates the need for effective PMO in the future. 

For this reason, PMO should take steps to support and sustain the uptake of AI, as businesses continue the road to full digital transformation.