PwC: 5 tactics that increase your chance of winning with AI

Artificial intelligence is now mainstream and the companies making it work are building a competitive edge that may be insurmountable, according to a new report from PwC. Successful early adopters are building a virtuous cycle that starts with better customer experiences, which encourages customers to share more data, which in turn powers smarter AI algorithms.

Anand Rao, global artificial intelligence lead at PwC, said the new report explains how companies are using AI successfully and what changes they are making to support that success. The new report, “How to navigate the top 5 AI trends facing your business,” includes data from 1,032 executives at US companies, including more than 200 CEOs.

Rao said that running a business during a pandemic forced executives to be more willing to use AI for decision making when they would have relied instead on previous experience or historical data.

“For the most part, people have no previous experience in dealing with something like COVID-19, and whatever data you might have is useless when you have the whole market come crashing down,” he said.

Rao said that 2020 and COVID-19 illustrated the importance of using AI for scenario planning. Companies also can use AI platforms for scenario modeling and simulation to deal with  uncertainty on multiple fronts: Changing infection rates, regional differences, workforce dynamics, and consumer behavior.

Companies could consider several time frames to map out a plan for supply chain issues, for example, that could be temporary or last six months or even nine months, or to plan for two, three, or even more waves of infections.

“Previously, executives didn’t understand why they needed multiple scenarios,” he said. “Now we can say we don’t know which scenario will play out and in the absence of that, you have to plan for both.”

Rao said that now executives are using scenario modeling to understand whether travel and shopping habits will go back to normal after the vaccine is widely distributed or if habits developed during the pandemic will persist.

“Will travel come back exactly as it was or will people say, ‘I’m not traveling six hours each way for a one-hour meeting,'” he said. “Everyone is asking, what is the new normal going to be and where is it going to be compared to pre-pandemic?”

The top findings are:

1. AI investments will increase
2. AI will fuel faster and better decisions
3. The time for responsible AI is now
4. New talent strategies will emerge
5. The AI reorganization accelerates

Rao said that he’s seen two primary use cases for AI across all industries:

1. Improving the customer experience and understanding demand on the revenue side
2. Making supply chains run more smoothly and efficiently on the cost side

PwC’s study found that these are the top five ways companies are going to use AI in 2021:

1. Managing risk, fraud, and cybersecurity threats
2. Improving AI ethics, explainability and bias detection
3. Helping employees make better decisions
4. Analyzing scenarios using simulation modeling
5. Automating routine tasks

Data scientists adopt agile approach

Rao said that the pandemic has changed how data scientists work as well. Data scientists have shifted their approach to work with less data to turn around an answer within days or weeks instead of months.

Rao described a project with a client who wanted advice about when they could reopen a mill after a COVID-19 shutdown. The question was how many employees would feel safe enough to return to work, after there had been cases in the plant. The client had only eight days worth of workforce data and had a target date of May 1 to reopen.

“We told them we could collect the data we had available now and refine the answer, so it would not be a final answer but our best guess now,” he said.

Rao predicts that this more agile way of working will continue after pandemic conditions have eased.

Why a reorg is required for AI success

PwC found that 76% of companies are barely breaking even on AI investments and only 6% had AI initiatives scaled across the enterprise. AI efforts involve multiple departments, which means that companies need habits and systems to share data, subject matter expertise, governance and AI models across teams and functions. As the report authors note, that might mean reorganizing. Here are PwC’s recommendations about how to reorganize your company to make AI projects successful:

• Bring AI, analytics, and automation together in one effort to increase your ability to monetize data, build a data-driven culture and reduce risk
• Pick an AI platform that fits your unique data sources, business processes and use cases
• Ensure continuous collaboration between engineers, data scientists, and line of business managers and staff

PwC noted that one of the easiest applications for AI–automating routine tasks–has dropped on the priority list. Only 25% of respondents listed this project as a top priority compared to 35% in last year’s survey. PwC sees this as a sign that many companies have progressed beyond this easy win to more strategic use cases for AI, which makes reorganizing “inevitable.”