ARTIFICIAL INTELLIGENCE, BOTS AND ERP

Once considered purely science fiction, Artificial Intelligence (AI) is now a hot topic in the ERP space. We mentioned it in our “What is ERP,” blog post and our Sage Summit recap post.  We’re seeing more and more AI, Machine Learning, Natural Language Processing and bots in consumer technology every day.  But we’re only now starting to realize the impact these technologies will have on our businesses, our jobs and of course, our enterprise software.

What Is Artificial Intelligence?

Artificial intelligence is the science of incorporating human intelligence into machines.  AI is a sub-field of computer science which aims to enable the development of machines/computers that are able complete tasks normally done by humans, specifically those tasks that normally require human intelligence.

However, by that definition, any program can be considered artificially intelligent if it is programmed to do something that we normally think of as intelligent in humans (think of a calculator performing mathematical functions).  How exactly the program completes the task is not the defining factor, rather the ability to do it at all.

Different professionals have different goals with regard to harnessing AI.  There are three main subsets, based on how closely the machines get to replicating human cognitive abilities.  Strong AI aims to genuinely simulate human reasoning (think Terminator or iRobot), while Weak AI aims to get the job done by creating systems that can behave like humans without needing to think or reason (think of IBM’s chess bot from the late 90s).  Somewhere in the middle of these two lies the third camp.  In-between AI systems are inspired by human reasoning, using it as a guide without aiming to replicate it.  This is the camp where many modern AI developments fall into.

Another defining distinction to be made is that between Narrow AI and General AI.  Narrow AI refers to a system designed to complete a specific task or tasks.  General AI refers to systems with the ability to generally reason.  Very few systems (if any) can currently be classified as General AI.

Machine Learning

What makes Netflix so good at recommending binge-worthy tv shows you’ll love? An algorithm that learns your viewing history and the viewing history of those like you.  This is an example of machine learning.  Machine learning can be defined as an algorithm that simulates “learning” by using observations like example data or historical data.  In essence, developing ML in a system creates a system that can learn from experience in order to make predictions, provide insights and produce intelligent solutions based on past data.

Natural Language Processing

Natural Language Processing (NLP) is the component of AI that has to do with language, both written and spoken.  This is particularly challenging because computers traditionally require humans communicate with them through a series of highly structured, clearly-enunciated, unambiguous commands. We all remember the early days of Siri when it was nearly impossible to make her understand what you were asking her.

By utilizing machine learning and natural language processing, you can create an intelligent system with the ability to learn and understand.  In fact, it could even potentially learn to understand by using machine learning to learn or improve natural language processing.

ChatBots

As services based on artificial intelligence improve, those intelligent systems need a way to communicate with humans.  Chatbots are one option.  A chatbot is a computer program that is designed to simulate a conversation with human users either through voice or text.

Similar to AI, bots are very popular in the consumer space.  They allow eCommerce, delivery and transportation businesses to let their end customers order services and goods more conveniently.  Chatbots are also making their way into the enterprise space.  Already, company employees can use third party bots via enterprise messaging apps like Slack.  However, the real value lies in enterprise companies building their own corporate chatbot to add value to their product, create a better interactive experience for B2B customers and improve internal business processes.

Improving ERP

APIs should certainly be a consideration when reviewing a new ERP system.  A robust API portfolio can mean that your business can embrace the multi-cloud world by integrating with other cloud platforms, but it can also allow your business to utilize AI and bots.

Recently we’ve seen large companies like Google, Amazon and Facebook open-sourcing their AI enabled bot platform APIs.  In many ways, open sourcing access to AI represents the culmination of the API economy, something we’ve touched on in previous posts like “What is ERP” and “How to Embrace the Multi-Cloud World.”  This means that developers can accelerate their own development by accessing the work done by others via API hooks.  Basically, links between software or platforms that allow them to integrate and speak to each other.  The API economy may be the only practical path forward in fully leveraging new enterprise capabilities in an enterprise.  AI and bots are just starting to influence the ERP space, but it is a powerful way to efficiently and accurately extend the capabilities of ERP.

Improved Decision-Making

One of the main functions of ERP is to make data and business information more easily accessible in order to allow for better decision making business-wide.  AI and bots extend this functionality by analyzing larger historical data sets.  For example, using historical data, organizations can analyze past equipment management and create a maintenance schedule using predictive maintenance.  Then, the AI/machine learning enabled bot can use that baseline, adjust for variables and apply a predictive maintenance plan that predicts problems before they can slow processes.

Automating Data Input

Data input is still one of the greatest expenses organizations face.  Thousands of hours are spent on manual processes like data input every year and result in duplicated or incorrect data.  AI however, can reduce the time it takes to input data and minimizes the likelihood of errors by integrating multiple applications and cascading data from a single source.  The best part is, it learns as it goes, making it possible to connect data in new ways across applications, recognizing patterns and pushing suggestions to eliminate inefficiency business-wide.

Streamline Business Processes

AI can also use machine learning to make recommendations for the most efficient internal workflow processes.  By analyzing historical data and accounting for variables, AI can provide a prescribed set of steps to complete a task in the most efficient way possible.  Resulting in tasks that are fully completed, without missed steps or incomplete, inaccurate information and employees are prompted to use time efficiently.

Most current chatbots are categorized as in-between AI systems.  They utilize an omnichannel UI to access business functions using natural language processing, focus on retrieving information, getting alerts, executing tasks, and automation through pre-defined rules.  But through machine learning, accelerated by the API economy, the future is bright for systems that are able to self-improve through basic contextual understanding and cognitive capabilities.

By connecting and gathering information from enterprise applications like CRM and ERP through APIs, AI-powered bots can serve many business-critical scenarios.  From a more accessible 360 degree view of the customer to a personal assistant that answers calls and records messages, the range of implementation scenarios of AI and chatbots in the enterprise space is nearly unlimited.

If you’re interested in setting your business up for success with an AI compatible ERP system contact one of our representatives today!