AI and the 80/20 Rule: Humans and Robots Will Collaborate to Improve Our Work Lives

AI and the 80/20 Rule:
Humans and Robots Will Collaborate to Improve Our Work Lives

The mainstream and business press

The mainstream and business press is suddenly flooded with horror stories about the threat posed by automation to white collar jobs. Amidst the warning cries from visionaries like Elon Musk, however, there are voices of optimism that foresee robots as our allies, rather than overlords.

One of those voices for optimism is Eden Shochat, who, in a December 2017 Wired Magazine article, describes a vision for how human and artificial intelligence work best in collaboration. The reason that Shochat sees a bright future for human workers in an automated world is that while automated systems have made huge strides, systems that are truly self-learning are in their infancy. As a result, robotic process automation systems still depend upon human intelligence.

As Shochat puts it, “By feeding terabytes into neural networks, computers are now able to understand voices, recognize faces and sift through data with unprecedented accuracy. And yet, advances in so-called unsupervised learning - which finds the structure or relationships in data inputs without training in the way that a child learns from experience - are almost non-existent.”

When it comes to typical back office processes like accounts payable or sales order processing, “the future is here,” as science fiction writer William Gibson famously put it, “But It’s not evenly distributed.” Companies that have embraced automation to manage manual, inefficient vendor invoice approvals and payments and sales order processing are no longer cutting edge, but they remain in the minority. These firms, who rely on machine intelligence to “read” a document, extract relevant information and handle the approval routing and transaction entry with minimal human input, typically achieve a return on investment in 6 to 12 months. The case for embracing automation for these kinds of routine, repeatable processes is well established, and implementing them has becoming easier and more cost-effective.

Even so, these fully automated systems need to be “taught” how to walk before they can run. Existing AP staff members, along with finance leadership, need to be involved in defining processes for these “robots” to execute. Once these systems are in place, they typically do an amazing job of automating the routine—while flagging exceptions to be reviewed and handled by their human coworkers. As a pattern emerges for exceptions and how to handle them, these “exceptions” also can be managed as part of the routine.

At the same time, today’s intelligent, automated systems are able to extract meaningful information from unstructured data sources like invoices, quotes or order forms, in a way that provides greater visibility to human staff members and managers about the business impact of the process. In the case of accounts payable, this allows finance managers to have access to more timely accrual reports and a greater ability to monitor and control cash flow. Today, these business decisions remain in the hands of seasoned, qualified professionals who are more empowered to make impactful decisions because of automated processes.

The question naturally arises, “But what if artificially intelligent systems could make better business decisions than a human?” That possibility certainly exists, but today, process automation lacks the precision and reliability go extend into human decision making.

According Shochat, “AI…has to be perfect - and most software engineering projects are not perfect. Unlike a bridge that can't be 90 per cent done, software can be "good enough". In software…the initial 80 per cent of the reward has 20 per cent cost. The remaining 20 per cent of the reward would take 200 per cent of the cost and time.”

It is in that remaining 20% then where we find the cost/benefit of human workers over their robotic counterparts—where human intelligence will continue to deliver better, more cost-effective results. And it is the automation of the other 80% of a task that will allow humans to capitalize on this advantage.

Open in PDF

Optimize Your
Data Capture Processes
with Artsyl docAlpha