The Power and Limits of AI

The Power and Limits of AI

Practical Applications of Intelligent Automation for the Back Office

When most people think about artificial intelligence, the first images that come to mind from pop culture usually include HAL from 2001 a Space Odyssey, or maybe IBM Watson’s performance on the quiz show, Jeopardy. Whether fact or fiction, both of these examples reference one typical kind of AI—one that is less common in the world of work today, and more likely to cause confusion when it comes to how AI is being applied to solve a wide range of business problems today.

HAL and most of the Hollywood representations of AI that we’ve come to know are representatives of what is referred to as general AI. These systems, which are also reflected by the capabilities of the IBM Watson platform, along with real life offerings from other vendors, can parse human language and respond/adapt to a broad range of conditions and variables in the real world, in ways that mimic human behavior.

One recent example of general AI in action is IBM’s Project Debater. At the IBM Think 2019 Conference in San Francisco, IBM’s Project Debater squared off against a human to debate the subject of “whether we should subsidize preschool.”

A human debater, Harish Natarajan, who happens to be a record-holding debater, took the position against preschool subsidies. Project Debater too the affirmative position. Both sides, human and AI, were given just 15 minutes to conduct research, then formulate an opening statement, a rebuttal and a closing summary.

Natarajan ultimately won the debate, but Project Debater held its own, and was able to pose a clear, coherent argument by parsing data and statistics from current research, parsing human language and constructing an argument that at least held water.

When we think about the complexity of developing software that can read and understand human language, analyze data that includes narrative text, then construct an argument or counter-argument in real time, we’re talking about some pretty powerful stuff.

General AI Hype v.s. Intelligent Process Reality

This is the AI that has captured the public’s imagination, inspiring lots of hype and a certain amount of panic around robots taking over the world.

But, before we all get ready to bow down to our AI overlords, we have to consider that this technology remains costly and complex and requires massive amounts of data and time to teach a general AI system to perform a specific task.

The power and potential are there—but the practical, actionable, real word examples of how to leverage a quiz show or debate-participating AI are few and far between. On the other hand, more narrowly-focused intelligent process automation systems are doing all kinds of amazing things behind the scenes in companies today that promise to have a deep and far-reaching impact.

Simple, Flexible Intelligent Process Automation

When it comes to practical applications of intelligent systems, there are far more to be found that rely on simpler technology, more narrowly focused on solving a particular business problem. Specific, targeted intelligent process automation solutions, for example, don’t have to answer trivia questions or engage you in a clever debate to be able to sort through your mail, enter data in your ERP system and route documents for approval from the appropriate stakeholders. They don’t make coffee, either.

That said, the power if intelligent systems to handle things like sorting through vendor invoices attached to emails or mailed as physical documents, extracting all the relevant data, looking for duplicates and errors, routing everything to the right people, then entering the data into ERP and other business systems is huge.

And it’s happening today, in companies of all sizes.

Powerful, Practical Solutions

Many of them are turning to Artsyl Technologies to apply intelligence to tasks that are far less sexy than holding a debate or answering a trivia question in real time—but also far more impactful to their business. Vendor invoice automation and Customer sales order automation are just two common examples, where AI isn’t there to show up a human being. Rather, intelligent processes make human process owners, by unburdening them from dull tasks so they can better use their brains and business expertise to form better partner relationships, enhance customer service, create budgets, generate forecasts and manage their businesses in more thoughtful, impactful ways.

To learn more about how to make your processes more intelligent and your employees more empowered, contact your Artsyl Technologies representative.


Optimize Your
Data Capture Processes
with Artsyl docAlpha