Low-Code Integration Helps Hardworking Bots Earn Their Keep

August 14, 2018 Thameem Khan

Illustration of various colorful, cute robots.

Imagine a world where all your manual, mundane tasks could be completely automated. Where you have a personal assistant who knows how and where to get things for you. It can constantly learn about you and can be ready and able to help at all times. Utopia!

Well, we’re not quite at a utopia yet. But we’re getting closer. There are a lot of contributing technologies helping move things forward. Bots are one of them.

Bots are sets of algorithms. These algorithms can execute tasks on behalf of users, connecting them or performing a service for them.

Bots come in various forms. Chatbots are designed to converse with people and mimic human behaviors.

An example of a bot for business is KLM Royal Dutch Airlines that created a chatbot to take some of the load off its customer service agents, who handle up to 16,000 cases per week.

KLM’s “BB” (short for “BlueBot”) helps passengers book a flight. It also delivers booking confirmations, check-in reminders, boarding passes, delivers flight status updates, and answers passenger questions.

Customers access BB via Facebook Messenger, Google Home or any device that supports Google Assistant. More than 500,000 passengers have engaged KLM through its bot.

Automating the Mundane but Important

Recent advancements in machine learning and natural language processing (NLP) have made bots of greater interest to enterprises. One key role for enterprise bots is to automate the routine tasks of providing access to information.

When people in an enterprise want to find information, they are usually looking for a subject matter expert — maybe more than one. But, usually, many more people need access to information than there are SMEs who can talk to them. And not everyone in an organization has equal access to the SMEs. So these experts can be a limiting factor and, ultimately, inhibit productivity in an enterprise.

There are many ways and places to store information but offering easy access is still a problem. That’s where chatbots come in. Chatbots can work through many of the most common communication channels, PCs, mobile phones and social apps like Twitter and Facebook. They can be an interface between users and the information they want. And, using NLP, they can interpret everyday speech, understand the context and retrieve information.

Boomi — The Secret Sauce

So, the chatbot framework provides the interface to the communication channel, and via NLP it interprets the request. But how does the bot know which application or endpoint to access? It’s here that Boomi plays a critical role.

Boomi becomes an integration layer that can understand the request from the bot, connect to the right application — say Salesforce — fetch the information, render it in the proper format and send it to the requester.

Here’s a real example of how this might work in an organization. We’ve used the Amazon Lex framework with its NLP capabilities and Boomi’s integration platform to create Betty, an office assistant bot.

The platform includes Boomi AtomSphere for integration, Boomi API design and management, and Boomi Flow. A company has hired a new vice president of sales. When he starts the job, he must deal with Salesforce, Workday, and SAP. He’s familiar with these applications, but every organization uses and customizes them differently. He could spend a lot of time navigating these apps to find the information he needs.

 

Schematic of how Boomi helps integrate Bots with enterprise applications
Boomi makes bots smarter by helping them learn about your enterprise applications and data.

Now, if he has a chatbot, he could say, “I want the latest pipeline report for Mary Scott.” The bot is smart enough to understand that for a pipeline report it must go to Salesforce. Then it can send the report via email or a mobile format to his phone.

The same process works for finding the onboarding status of a new employee. That information lives in Workday. Likewise, Jimmy could retrieve a combined report from Salesforce and SAP that correlates sales orders with products shipped.

And, just like Amazon Lex, Boomi uses configuration and low-code to achieve the integrations. The bot may require ten lines of basic code, which is readily available. You don’t need a developer with coding chops.

Our Boomi-powered slackbot and Bettybot are examples of a low-code approach to quickly linking your bot to the data it needs to do its job.

Bots Behind the Scenes

Boomi-powered bots learn as they go, since the Boomi platform has the ability to gather and interpret the data in real-time. Like, how much do people use that supply chain tracking software that cost your company $1 million to create and deploy?

With the data and learning acquired by the Boomi platform, bots provide a host of benefits:

Hands-free interaction with data and applications. Using an expanded definition of what we do today with Boomi Suggest — canonical data mappings and business objects maintained by AI and machine learning. This allows hands-free management of integrations and data models.

Context-aware endpoint discovery. Based on the metadata, we can interpolate the kinds of questions users commonly ask (e.g. please provide pipeline data by rep and region), as well as the systems that typically provide that data. For example, we know that opportunity data comes from Salesforce. If you add Zendesk or Jira, we understand support cases are handled there.

Purpose-driven data delivery. Also based on the metadata and the question asked, the bot learns the most appropriate approach to responding. For example, emailing a pipeline report in PDF format. And the bot could understand that questions about the number of outstanding customer cases could be simply sent via text message.

Data insight about application usage. Boomi provides data that helps identify the most used app, what objects are used most often, and how data is used. This provides insight into the ROI of your applications and customizations.

AI-driven prediction models for data and business processes. Boomi’s AI engine constantly learns from human behavior and its interaction with data. This can offer predictions for better data models and business processes.

It’s pretty amazing that the Boomi platform, which requires very little coding and has a zero-footprint, can help large enterprises automate hundreds of time-consuming tasks and increase the productivity of their entire workforce.

Want to know more about how the Boomi integration cloud can help you take advantage of emerging technologies like bots? Contact a Boomi integration expert today.

Previous Article
How to Build an Integration Center of Excellence
How to Build an Integration Center of Excellence

In a previous blog post, “Why Every Enterprise Needs an Integration Center of Excellence,” I wrote about th...

Next Article
Integration: Key Ingredient for Transforming Chain Restaurant Industry
Integration: Key Ingredient for Transforming Chain Restaurant Industry

Chain restaurants can relate to the old saying, “Too many cooks spoil the soup.” But I’m not talking about ...