Chatbots Meet Enterprise Software

SAP Partnering with Kore to Build SAP Enterprise Bots

For over two years, teams at the SAP Innovation Center Silicon Valley have been exploring natural language-based interfaces and cognitive abilities in the context of enterprise software. The projects range from enabling an executive dashboard on a large-scale video wall using voice recognition, authenticating users to business systems by analyzing the individual characteristics of their voices and faces, to the most recent focus area, Enterprise Bots – the incarnation of chatbots in a business context.

The Innovation Center Silicon Valley is exploring the space with various internal and external teams, one of them is the startup Kore. They provide an enterprise-grade platform to deploy Enterprise Bots on a large scale for a variety of business use cases.

[embedded content]

Source: Kore, Inc.

Messaging as a Platform for Enterprise Applications

We live in times where only on Facebook’s messaging platforms (Messenger and WhatsApp), people exchange over 1 billion messages every single day and this massive trend is not just restricted to millennials. Already today WeChat is used in China by millions of users to order taxis and products, to schedule medical appointments, to pay bills and to transfer money, all using a messaging interface without having to install additional apps.

So it is rather obvious that the move towards a new user interaction paradigm bears considerable potentials also for enterprise software. Bots will not replace mobile apps just like mobile apps have not replaced desktop software. But again they will further diversify the range of alternatives to consume digital solutions and to get the job done.

Sources: Statista, Bloomberg, TechEmergence

The Economist sees bots as the next frontier, TechCrunch announces that we should forget apps as bots are taking over, many commenters see them as a keystone technology for the IT industry going forward, underlining the obvious trend that we are approaching a massive shift towards the AI-enablement of enterprise applications.

Chatbots in the Enterprise

Enabling humans to communicate with systems and vice versa with low friction and natural interactions using message-based interfaces has a number of advantages.

For end users, Enterprise Bots lead to an increase in productivity:

Better Ease of Use: One interaction step to any business function, a dramatically simplified user experience even in complex business scenarios.
Less Training Required: Business knowledge sufficient to use any software, no training necessary, users just need to know common business terminology.
Shorter Time to Insights: Personalized, contextual, dynamic, and robust understanding of user input.

NLP and AI in Enterprise Applications enables a better ease of use with despite an increase in functional richness

On the other hand, for application providers, it can reduce efforts to implement business applications considerably:

Minimal GUI: (Semi-) autonomous generation of bots based on existing APIs, mostly text-based interactions augmented with existing WebViews or shortcuts to existing applications.
Autonomous Improvements: Continuous improvement of capabilities and quality through unsupervised learning based on user input, user behavior, and business data.
Centralized Learning: A holistic NLP and AI engine which provides the interpretation of natural language and intent is trained centrally, serving all bots and users and learning more and more about additional domains and business scenarios. What one user teaches the system also benefits all others.

From Zero to Bot in 5 Days

By leveraging the integrated bot platform provided by Kore, the Innovation Center Silicon Valley was able to implement a pilot bot within less than a week, which already demonstrates the potential of the technology from an implementation perspective.

The use case of the bot was primarily to enable users to check the stock status of a certain product, but also to inform them proactively once a purchase order could not be fulfilled and what the best mitigation actions might be.

SAP Stock Champion Bot (Concept Design)

In addition they provide a semi-automated way to create NLP-enabled bots through the consumption of existing APIs, in this case from an enterprise application on SAP HANA. Kore’s integrated platform also provides the capabilities to deploy chatbots at an enterprise scale – e.g. to different groups, individuals, teams or entire corporations.

This will be a crucial success factor for Enterprise Bots further down the road – it is not sufficient to be able to implement a robust chatbot which interprets user input correctly and corresponds in a natural manner. Enterprises rather need to be enabled to deploy these bots across their organization, to maintain them in a sustainable way and to be able to control the access to it in a secure manner.

Amplifying Employee Productivity Through Enterprise Bots

The variety of use cases for Enterprise Bots with omni-channel access is sheer unlimited – here a list of examples:

A sales representative updating her latest customer opportunities using a CRM bot via Facebook Messenger.
An office worker submitting his leave request using a HR bot via the Slack app.
A project manager being alerted about task delays and potential mitigation actions by a project tracker bot via Telegram./
A technician sending a picture of a broken pump by iMessage to let the service bot identify the model. The bot engine is augmented with image processing capabilities and responds with the detailed product details and suggests to order a replacement.
A team assistant sending a mail to the facilities bot to reduce the temperature in the IoT-enabled building.
An executive requesting a team update from his direct reports using the status bot which gathers the information from all members via WhatsApp.
A presales engineer booking a trip to New York using a travel bot via SMS, which automatically dispatches individual requests to a flight bot, a hotel bot, an Uber bot and a calendar bot that automatically updates the user’s calendar.

The SAP Innovation Center Silicon Valley is exploring various use cases with Kore and other frameworks and is continuously looking for parties who are interested in the space. SAP’s Co-Innovation Lab is working with Kore to enable pilot scenarios for SAP ERP, CRM and Concur on their platform – more on this by David Cruickshank in this post.

Looking Ahead

The Journey from Chatbots to Enterprise Artificial Intelligence

Over the past few months, the team has outlined a potential journey for business applications, which starts with the introduction of chatbots into the world of enterprise software.

Once business applications are enabled with a conversational UI for simple tasks with predefined rules, the next leap will be the introduction of autonomous features where bots and assistants comprehend the holistic context of a user and a certain business scenario. Last but not least, enterprise systems will be augmented with cognitive abilities to bring the amplification of knowledgeable workers to a whole new level.