top of page

How a Plug-n-Play Rookie Beat a Gunslinging Veteran Quarterback

As the NFL season comes to a close with its crescendo moment – Super Bowl LVII – I just wanted to take a few minutes to shed light on how machine analytics and human inconsistency played out on the gridiron and in the biggest moments to determine the outcome.

Let’s hit rewind to the San Francisco 49ersDallas Cowboys game on January 22nd…

For much of the game, 49ers quarterback Brock Purdy stuck to the script. Mr. Irrelevant wasn’t about to make any mistakes against a tough Dallas Cowboys defense. He was methodical, resulting in a tie score deep into the third quarter.

Then came the play of the game. Purdy rolled to his left and saw both primary and secondary receivers covered. Rather than throw it away, he fired the ball back across the middle of the field — a cardinal sin for an NFL quarterback — to his scrambling tight-end George Kittle, who bobbled the ball before securing it for a 30-yard pickup. The gutsy play sparked the game winning drive.

Winning comes down to mastering split-second decisioning

Purdy’s performance is a nearly perfect example of the melding of machine analytics and human intuition. The NFL has the Next Gen Stats program that collects some 3TB of data from sensors around stadiums and on player equipment during games every week. AI and machine learning use this data to help teams improve their scripted “playsheets” and even in-game calls in certain cases. Yet, there are times when injecting split-second human intuition leads to a better outcome.

Immersed in a technology culture, the 49ers run a sophisticated offense in a plug-n-play system, whereby a third-string rookie quarterback can win high stakes divisional playoff games (and possibly the Super Bowl if injury wasn’t a factor). In all my years as a football fan, I can’t recall this ever happening. (For the record, only nine backup quarterbacks, none of them rookies, led their teams to championships since the NFL-AFL merger in 1970.)

In contrast, the Cowboys’s fortunes rest largely on the shoulders of a wildly unpredictable, gunslinging quarterback. To say Dak Prescott is reckless would be an understatement. He co-led the league with 15 interceptions despite playing only 12 games, and threw two interceptions in the game against the 49ers. Most people agree he is the reason the Cowboys lost.

“Dak Prescott gave away the ball twice in the narrow loss to the 49ers, in a matchup the Cowboys had a chance to win if they didn’t again generate self-inflicted wounds,” tweeted the Dallas Cowboys.

Self-inflicted wounds, foibles and inconsistency are what can happen when an organization leans too much on human intuition and not enough on analytics. Decision-making in the moment is a balance between analysis of what’s worked in the past and a human understanding of nuances and special circumstances.

Unlike machine learning, Prescott didn’t learn from his mistakes. His first interception against the 49ers was eerily similar to his interception against one of the league’s worst teams, Houston Texans, during the regular season. In both cases, Prescott took too long to throw a short pass toward the sideline against a defensive player sitting on the receiver’s route. Later, some analysts questioned whether he even watched the tape.

“We’re watching Dak Prescott throw very simple interceptions,” said ESPN analyst and former NFL player Ryan Clark. “That’s the most alarming, trying to throw the short corner into a Cover 2 against the Houston Texans.”

If you think this is just about football, think again.

Your contact center agents are your quarterbacks

Combining analytics and intuition is something we do every day. We’re always striving to make the best decision in the moment with the most information available. Art and science are our collective strengths. From quarterbacks to contact center agents, we rely on our right and left brains working in harmony.

With advancements in AI, the left brain is becoming more machine than human. Like the Next Gen stats program, AI and machine learning in marketing, sales, and customer service mine mountains of data to help people make smarter decisions. According to Aberdeen’s CX Executive Agenda 2021 survey, companies using AI experienced:

  • a 3x greater year-over-year increase in customer retention rates (10.5% vs. 3.2% among contact center users not using AI).

  • 5x greater year-over-year increase in CSAT rates (10.1% vs. 2.9%).

  • 8x greater year-over-year improvement in customer effort score (8.8% vs. 1.1%).

For the Brock Purdys among us, a contact center agent dealing with an agitated customer must focus on the needs of the customer without the minutiae of toggling between screens/apps, writing post-it-notes, manually inputting data in tons of fields, and other tasks that prevent the agent from focusing on the customer and the conversation. Today AI technologies that assist agents can provide:

  • real-time AI-generated workflows that adapt to the customer conversation

  • scripts and pointers for agents developed from machine learning

  • dynamic data that automatically surfaces so agents don't have to look it up

  • sentiment detection and meaningful and impactful responses

These foundational capabilities take significant cognitive load off of agents and frees them up to play quarterback when human intuition suggests a new course of action in the moments that matter most for the customer. This is what enables the plug-n-play ability where a newly hired agent can slot in and begin resolving customer issues, up-sell and cross-sell, and troubleshoot confidently.

It’s human and machine (aka collaborative intelligence) making the right call.

Extra point – I didn’t mention ChatGPT even once (at least not in this article)!

To learn more on how to leverage human intuition and machine intelligence to elevate your customer and agent experience click here or contact us at


bottom of page