BOTS through Co-Innovation
I have committed to myself (in writing) that somewhere inside the month of April 2016, I am going to resurrect my dormant blog with news from COIL Silicon Valley. A stretch goal is to try to share a bit, of what is emerging from other COIL locations from around the world-we have some super projects going on this year. For those counting, we have 25 co-innovation projects in flight in just the COIL Silicon Valley so my blogging source material locally is a rich pool from which to draw.
To dive directly into my first topic- Bots, I should begin by recognizing that if you are in the computer industry then I can only presume some exposure to the buzz about bots. You hear about them a lot in the context of personal assistants (not Microsoft Bob one can only hope, although an otherwise amped up “Clippy” might be cool!)
If you are not so familiar with what bots are all about then you have my permission to stop reading and start Googling. If it helps you stay with me in this post, I can offer you a simple description:
A bot is nothing more than some software designed to automate the sorts of tasks you take on for yourself like making calendar appointments or reservations or running out to capture then display some content or information of interest to you. When designed as a messaging app it sets up a user experience akin to IM chatting except that the chat can be with a business system instead of a human being.
You likely hear or read most about chatbots. I am therefore excited to report that COIL has its own bot project this spring. Kore Inc., joins SAP here in the COIL lab to build a few SAP Kore bots designed to aid humans in their work activities that include regular interactions and experience with SAP applications and systems.
As you encounter industry news about bots, (quite a lot of stories already in 2016) you will quickly piece together your own sense of how this technology is advancing, its rate of adoption and future directions but also about the concerns over how easily we can come to trust bots to perform as and when we expect.
As a technology, bots have been in the technical landscape for some time. There are numerous successful commercial uses in play today like WeChat, which supports robo-accounts. WeChat users regularly communicate via bot technology to run job searches or shop. Facebook is betting big on bots, as is Microsoft (despite issues). Even Taco Bell has a bot. I love the name too- TacoBot. (Maybe it can run to the border for you.) It can certainly take care of ordering and paying for you.
Artificial Intelligence (AI) is one technology driving the direction of bot development with respect to evolving natural language processing. Everything delivered from the cloud and numerous platforms is yet another. As these capabilities advance so too will bot ability increase to carry on and to understand conversations that are more complex. Here at COIL Silicon Valley we have AI, ML and IoT projects in flight right now and I’m hopeful that some intersects will occur with one or more of these teams in ways leading up to how a mash up of these capabilities and know how can spawn even newer solutions.
There is still a lot of progress required to render a new generation of intelligent bots that understand spoken language in context. Simply think of Amazon Echo (Alexa), Cortana, Siri and the aggregate investment behind all of it as a leading indicator for what is imagined and likely to come in the near future. There are strong business incentives to realize fully digital and useful personal assistants. (I’ve wanted one since at least 2006 that simply acts a my personal on line security and privacy officer deputized to represent me whenever I get an email saying that my account was hacked and I need to change my password- let the bot do it.)
Bots (today) may not understand human speech well enough to guarantee error free communications or actions but one day perhaps using machine-learning techniques a bot will retain prior conversation transcripts and learn to apply them in context in all future conversations. Nevertheless, we can see today that there is a lot which can be done to manage tasks via discussion. It is cool to see what the SAP Kore project team is doing to demonstrate this at COIL.
I am not going to get into here, but much of what I read suggests some bot zealots think of them as something which will flat out replace apps. Messaging may very well be emerging as a new platform for applications and services. I have heard such arguments before when it comes to big shifts in human behavior due to some new technology so this notion has wait and see written all over it. I will leave it with others to predict if bots will experience widespread consumer uptake, but from a COIL perspective, what works now can be readily applied to a variety of enterprise jobs and tasks.
As I describe this COIL project, I would like to point out one thing that we believe significantly contributes to the momentum of the project and is a marker for its success. It is fact that Kore is already making it possible for more than 250 million knowledge workers globally, spanning more than 10M businesses, whose jobs are becoming increasingly more complex, a simple way to get things done via messaging. The experience, insight, knowledge and passion the team possesses when it comes to the creation and execution of enterprise bots has been evident from the day we all agreed to pursue co-innovation project work. What is obvious is the keen insight for how to capture a consumer quality experience in the context of enterprise IT.
The use cases underscoring the COIL project include demonstrating how SAP Bots communicate with other people, entire teams and systems. Our first use cases explore this using SAP ERP, CRM and Concur. Within the SAP CRM and Cloud for Customer (C4C) use case, Kore is working with SAP to reduce the effort needed to manage the sales cycle by enabling sales, marketing and support to communicate with their SAP CRM system in the same way as these users communicate with each other. The project work includes constructing a network of intelligent SAP CRM chatbots to manage the sales cycle and to glean customer intelligence. Similarly, in an ERP context, there is an effort to demonstrate the streamlining of processes spanning procurement, manufacturing, service, sales, finance and HR by letting users communicate with these systems using chatbots.
Each of these use cases have the objective of creating useful “SAP bots” which can tell you things like when an opportunity changes or an order is delayed or an expense needs approval and suggest various actions that the bot will execute. Some added functionality that I am not able to disclose just yet, is now layered into the lab efforts right now and a few new use cases are also being developed including one of my favorite topics; IoT. I clearly see high potential with the notion of enabling a workforce to be able to communicate with any connected thing or having a network of SAP IoT bots gathering information as a way for a user or team to gain insights into the business.
The COIL project team further focuses upon how these bots function as part of an SAP cloud so there is time now spent creating an App Services Java container application using REST and https. Small payloads containing specific transaction data from the SAP Business Suite is structured using Odata or JSON. For testing, the team has simulated 20 user types exchanging around 50 messages per day.
From my observation, a bot focus on the enterprise makes a lot of sense. Those pioneering with this technology on the consumer frontiers will all be watch worthy but the capabilities of the technology sits a bit further beyond the social dimensions influencing bot uptake like learning to trust a bot not to make unrecoverable errors or the potential for damaging a firm’s brand. On the contrary, in the enterprise where complexity has an adverse impact upon innovation and execution, bots can offer a lot of simplicity and improve upon execution of tasks within business systems.
Innovations will emerge fast as a result of doing this well and I can quickly see those driving a consumer bot experience to steal from the playbooks of those who cast incredibly useful enterprise bots. It is without question, a fertile ground for a whole lot of co-innovation where we begin with SAP and Kore at COIL. We look forward to more use cases, demonstrations and successes in all of the future project work. If you think you have a practical idea for an SAP bot, drop us a note or create a bot and message us.