Chatbots are invading the business world. Among enterprises, we’re seeing an uptick in their use for customer support automation by the likes of American Express, Royal Dutch Airlines, AT&T, and others across a wide swath of industries. And while customer support is the most obvious use case, bots today are underused as efficiency drivers, and they can have a much larger impact on all aspects of a business for those who learn how to leverage them.
In fact, with businesses increasingly adopting a multitude of cloud solutions, chatbots are uniquely positioned to help companies bring together different software into one streamlined communication channel. Booming growth in open APIs has made this easier than ever across a wide array of applications. The payoff will be significantly greater efficiency: Companies using support bots report efficiency gains in excess of 30 percent. As bots spread to sales automation, automated recruitment, intelligent Q&A, and other applications, I predict we’ll see even greater impact on bottom lines.
The chatbot advantage
Why are chatbots so well-suited to this work? For the same reasons that enterprise messaging is increasingly the new OS for businesses. First, they are easy to implement in any conversation. “Installation” often requires nothing more than inviting a bot into a chat stream, and users interact with them largely in the same way they interact with other users, and on any device. Second, bots address needs across a spectrum of complexity, from sales opportunity updates broadcasted to the relevant team, to a pricing lookup bot, to insurance claim workflows that involve multiple systems and departments. Chatbots facilitate these varying use cases with a familiar interaction model and preordained context (e.g. participants and case numbers), driving higher task completion and accuracy.
Finally, chatbots are increasingly becoming accessible to the point that the everyday knowledge/IT worker can use publicly available bots or implement custom integration bots across a wider array of messaging platforms. For example, message.io offers a “translation” service to enable bot builders to distribute their Slack bot onto a customer’s messaging platform of choice. Another company, Troops, builds “sales bots” that aid sales teams in interfacing with their Salesforce account from their messaging platform.
For companies that have in-house development capabilities, bot frameworks such as Hubot, Lita, and Botkit have spurred the creation of thousands of chatbots (many of them powering software integrations) in the open source community, and many of these codes come with a friendly “deploy to Heroku” button.
Getting started with bots
Now that you know how beneficial bots can be, here are some quick guidelines for creating a sound bot strategy for your business.
1. Learn from your employees
Take a close look at what your employees are doing on the company messaging platform. The blurring of lines in tools used for work and outside of work, especially among millennials, can be an asset to help you assess the best tools to bring in. Security issues aside, shadow IT can be a powerful teacher. Tech-savvy employees can easily bring public bots into a corporate messaging platform, and this is an opportunity for IT to discover new efficiency-driving cases via this type of bot discovery by employees.
2. Choose a messaging platform that has enterprise expertise
Look for providers that have an enterprise-level customer base, and ask them about activity logging and monitoring, data loss prevention setup, compliance support, bot and integration administration, and other enterprise concerns.
On the functionality side, SMS is still a good channel for enterprise messaging communication, and most bots can work on SMS. Find a messaging platform that is SMS capable. One of the early chatbots built by a RingCentral customer was done entirely on our SMS service, enabling sales reps to quickly query and receive info on latest pricing on specific SKUs.
3. Set proper expectations
There will always be limits to what can be presented in a message stream; users will not be conversing their way through spreadsheet edits anytime soon. So figure out what your employees vitally need and what they should go to the apps to perform. Expectations of AI bots with perfect NLP replacing common job functions also remain largely unfilled (just think of your last frustrating conversation with Alexa).
With esoteric business workflows, today’s bots work better as facilitators of information and process. For example, a bot that proactively moves a support case through different teams with automated reminders and contextual support case data will increase the productivity of all human participants by shaving minutes off each case as opposed to eliminating the need for a specific person in the chain.
4. Develop an information and security governance strategy
Bots introduce new issues within information and security governance. To cover the former, make sure you can answer some key questions: Who will own governance? How will access permissions set by IT in integrated software be adhered to by a chatbot? How will you ensure employees can share partial information while keeping other data hidden? Like email systems, enterprises should also scrutinize their messaging platforms to ensure DLP and e-discovery technology can be applied.
On the security side, determine whether the ability to bring in chatbots will be democratized or controlled, and whether your corporate policy permits the use of public as well as private bots. Like other cloud-based technology, bots can be publicly or privately hosted, with each having pros and cons. I predict that enterprise security companies like CipherCloud and Skyhigh may look into this space soon, and new services will help companies navigate the bot landscape more securely.
We’re just beginning to see the impact that well-implemented bots can have on companies, but as the importance of integration shows no signs of dimming, bots will only grow more relevant in driving efficiency forward.
This article was written by RingCentral and David Lee from VentureBeat and was legally licensed through the NewsCred publisher network.