Business Systems are tasked with streamlining business processes with technology. Is the time right to be thinking about introducing AI into business workflows? And if so, what’s the best way to go about doing that?
Let’s begin with some FAQ’s about enterprise automation and artificial intelligence.
What is artificial intelligence, exactly?
Artificial intelligence doesn’t have a concrete definition; in fact, there is much debate and little consensus around metrics to determine when a computer is “intelligent”. Jerry Kaplan of MIT notes that “Early researchers focused on ways to manipulate symbols according to rules,” which was useful for mathematical tasks. These forms of logic were not applicable to computational use-cases such as understanding language or pictures, which require more subtle, nuanced perception. To address this issue, of distilling insights from multi-faceted sensory intake, researchers have turned to machine learning (ML). Machine learning is the process of feeding a quantity of data to a program so that it can “learn” patterns and imprint those pre-existing patterns into its logic.
Kaplan contrasts traditional symbol and logic-based computing with machine learning, and notes that “the mere existence of two major approaches with different strengths calls into question whether either of them could serve as a basis for a universal theory of intelligence.”
The Application of Multiple Intelligences
Business leaders may be most equipped to effectively implement artificial intelligence into business processes not by viewing advanced computing methods under one ontological umbrella of AI, but to consider different applications and technologies as multiple intelligences equipped with different capabilities and propensities.
We can actually see this trend of multiplicity and specialization in the business world today, in the proliferation of SaaS applications used by businesses. Different applications provide unique, specific functionality to solve a business process challenge. Seeking out applications, integrations, and automations that use more advanced computational methods such as machine learning and bots can add to the power and efficiency of these SaaS workflows.
So, let’s look at some specific use-cases for artificial intelligence in business systems.
Examples for Implementing Artificial Intelligence in Business Systems
Image parsing with ParseHub and Google Cloud Vision
The leading worldwide retailer uses an HR automation with ParseHub and Google Cloud Vision to parse images. Previously, the company was having employees do HR research in the geographic location around their stores, by monitoring local ads and jobs postings to see what a typical wage is for a retail employee in the area. So, employees would literally take photographs of competitor’s jobs postings and report the wage information manually. Of course, this was an elaborate, time-consuming process. To solve the problem, they now have ParseHub scrape the web for job postings and place the postings into the HR database in real-time. Now, employees can also photograph job postings and upload them to Google Cloud Vision. Workato parses the data and sends it to the HR database. If the photo is too poor of quality for Cloud Vision, it’s escalated to a human for review. Scaling their research process with artificial intelligence saves time and allows the process to operate in real-time.
Voicemail parsing with IBM’s Watson to create IT support tickets
A leading cafe chain that uses digital devices in their cafe workflow was having problems with their reporting process to create an IT support ticket when a device wasn’t functioning. Their process for creating a support ticket was originally to have the employees go to the back office, call a call center, report that the device was down, and create a ServiceNow ticket. In a hectic cafe environment, the follow-through on this was not at 100%. Employees were often busy or distracted by the front-of-house operations, and the ServiceNow interface can be confusing for someone that’s not experienced with using it. Additionally, the cafe chain was paying a large number of people to sit at a call center and accept phone calls about non-functioning devices. Clearly, they needed a more scalable solution. Now, the cafe uses Splunk to scan the devices to monitor the devices’ code. Splunk picks up code errors early, and Workato creates a ticket in ServiceNow. Additionally, they replaced the call center with a voicemail box and IBM’s Watson. Now when employees call in, they leave a voicemail, which is parsed by Watson and sent back to Workato. Then, Workato creates a ticket in ServiceNow.
Enriching lead data with Google Cloud Vision and Clearbit
Trade shows are a valuable opportunity for companies to cultivate their brand image and build relationships with existing and potential customers. However, actually following up with each lead after a conference can pose a range of challenges- lost business cards, incomplete lead data, etc. However, using Google Cloud Vision and Clearbit, marketers can quickly compile and enrich lead data into a central location. The automation starts by sending a photo of the lead’s conference badge to a Twilio number, where it’s picked up and sent to Google Cloud Vision, which parses the image for their information. Then, a new lead is created in Salesforce, and the lead’s information is enriched with Clearbit. Within minutes, their information is delivered via Slack notification to the appropriate team member. The artificial intelligence eliminates manual data entry, so you don’t have your employees spending valuable conference time typing out lead information into spreadsheets.
AI for expense reports and financial auditing
AI has matured in the area of expense report auditing. Until recently, auditing expense reports has been a time-consuming manual process, which meant companies were only able to audit a small percentage of reports. Now, companies can automatically parse expense reports with platforms such as Expensify, which can scan a receipt and automatically fill in data fields about the expense.
AI can also be applied to audit billing mistakes and improper payment terms.
Additionally, Tensorflow, an open source software library that can be used for machine learning, can be used to process legal documents, assess invoice accuracy, be applied toward fraud detection, and resolve issues, enhancing BPO (business process outsourcing) productivity.
Choosing an Intelligent Automation Platform
In order to execute on many of the above examples, you need a platform that can execute Intelligent Automation. Intelligent Automation is what orchestrates the workflow between your applications and the AI tool. There are options that range from Intelligent iPaaS (an iPaaS that can work with AI tools) to Enterprise Automation Platforms. Enterprise Automation Platforms not only come with the same capabilities as an Intelligent iPaaS platform, but they also enable more builders and faster deployment times, thanks to the digital native architecture of the platform.
How is RPA with AI different than intelligent automation and enterprise automation?
RPA with AI is RPA that is enhanced with AI capabilities, for example, by connecting to AI tools from IBM or other providers. RPA itself is limited by the nature of its design.
Bandana Ojha defines the limitations of RPA as follows:
- “RPA cannot read any data that is non-electronic with unstructured inputs.”
- “RPA is not a cognitive computing solution. It cannot learn from experience and therefore has a ‘shelf life’.”
- “Applying RPA to a broken and inefficient process will not fix it. RPA is not a Business Process Management solution and does not bring an end-to-end process view from approaches.”
Essentially, RPA performs repetitive tasks well, if the conditions of the task remain the same. However, it’s rigidly structured to perform that task, and it tends to break when conditions change. For this reason, RPA could be described as brittle.
Does your Automation Platform Utilize AI and Machine Learning?
If you want to utilize the power of Machine Learning in your workflows and automation building process, there’s only one platform that has ML built in. Workato’s enterprise automation platform, that allows users to build self-service automations without code, uses machine learning to assist users as they develop recipes. This machine learning feature is called “Recipe IQ”, and it’s powered by the Workato open source automation library.
Recipe IQ: using machine learning to develop integrations and automations
Recipe IQ is a tool that continually analyzes the community recipe library, where open source recipes (automations) from other users have been stored for future adaptation and use by new users. Recipe IQ helps new users build automations by assisting with the process of field mapping between applications, or recommending next steps in automation based on the oeuvre of recipes in the library.
Machine learning in this case could be described as a form of community intelligence that adopts community standards and norms from past recipes to make recommendations. It makes recommendations such as suggesting next steps for field mapping when designing integrations between SaaS applications.
Has AI really progressed enough to be useful for businesses in 2019 and the 2020s?
The answer is yes. Although, the more relevant question might be, has automation progressed enough to be useful for businesses in 2019 and 2020, and how can automation be equipped with ML, data analysis and work bots to perform more effectively? The answer is that intelligent automation is more dynamic and robust than ever. We’ve progressed beyond the capabilities of RPA. When business processes are automated with intelligent automation tools, they become faster, more efficient, and they can work around the clock. No matter how you define artificial intelligence, the tools that businesses have access to today are increasingly flexible, agile, and adaptable, and offer value for businesses that implement them. AI is a particularly good solution right now for handling image parsing and voicemail parsing.
Designing business processes to mimic human decision-making
There is some discourse in the field of architecture that an effective way to design structures is to analyze the patterns of movement that naturally occur in a space, then build to facilitate that movement. For example, if you are designing a park and notice that one area is already being used consistently as a foot path by local residents, to cross from one side to the other, it would be ideal to build a path and bridge in the area that’s already being used as a foot path. Likewise, as a business systems professional, you’re very aware of the most significant data pathways within your business, and the bottlenecks that need to be solved. If the business process is good, but you need to increase the efficiency of it, you can use the Workato development platform to create a flow of behaviors that mimic the processes that employees are already using to manage the task, similarly to the way that the global retailer and cafe chain used artificial intelligence to address bottlenecks.
Implementing AI-enabled enterprise automation
Today’s integration tools, powered by logical decision-making that mimics human capabilities, offer businesses unprecedented subscription-based computational power.