Digital Assistants for Work
Updated: May 4
Over the past 5-7 years, our interaction with technology has shifted more and more towards conversations. Siri, Alexa, Google Assistant, or Bixby are some of the well-known assistants that simplify playing music, controlling a smart home, performing searches, finding a recipe, booking a restaurant, getting a taxi, or even listening to a good joke. On the business side, conversational technology has been applied successfully in areas like customer support, FAQ, marketing, sales, insurance, or e-commerce.
The conversational technologies behind the scenes have many names: bots, chatbots, conversational interfaces, contextual assistants, conversational AI, virtual agents, conversational experiences, intelligent digital workers, and digital employees - just to name a few. The landscape can be confusing and the boundaries between them are blurry. However, here's a simple way to get our head around it:
On the left side, we have simple bots able to execute specific commands. On the right side, we have autonomous digital employees able to perform a specific job (unfortunately we are still several years away from that). In the middle, we have digital assistants able to understand and execute higher-level actions. The left side is focused on the conversational aspect of the interaction, i.e. understanding the intent of the user and extracting the relevant information. The right side is focused on intelligence, i.e. understanding the business context and reasoning about when an action needs to be performed.
Conversational digital assistants are hard
Most of the existing technologies, like chatbots, are focused on the left side and they have been applied in transactional use cases, i.e. FAQs, support tickets, qualifying a lead. However, building a digital assistant to help the user on an ongoing basis is much harder. First, the assistant should work with a lot more context data at all times, about the user and the work they're doing. Second, a digital assistant should be able to work similarly to its human counter-part: understand more complex commands with minimal input, adapt to the user's workflow, be able to learn new rules on the fly, react to various events, answer questions about its data and activity, etc.
Of particular importance is its ability to act on behalf of the user. A few years ago there might have been some challenges in this area, however, with the mature APIs that most SaaS platforms offer nowadays and with the rise of RPA (Robotic Process Automation), the cases where a robot cannot perform an action are becoming the exception. The challenge though is the variety of skills a digital assistant should possess, ranging from general ones like working with files, emails, documents, filling forms, to specific ones like extracting a balance sheet from a specific accounting software.
Innovation is needed
Our mission at RoboSelf is to help businesses build digital assistants for their customers and their employees. And get us all closer to developing digital employees. To do that, we believe innovations are needed at multiple levels. First, we believe a paradigm shift is needed with more focus on the business objects an assistant must understand and work with. We call our approach "Object-Oriented Conversational AI" (technical folks might have an intuition about what it means, roughly). Secondly, we believe there is a need for digital assistant frameworks that address the full set of challenges in building a digital assistant: NLU (Natural Language Understanding), context management, actions execution, question answering, data queries, rules, events, integrations, and NLG (Natural Language Generation). Last, but not the least, we believe new platforms are needed to enable businesses to easily create digital assistants and deploy them.
It may take years before we see mature platforms, but we believe that, within the next 5 to 10 years, the first versions of a J.A.R.V.I.S. (Just A Rather Very Intelligent System) for work will simplify our daily business activities!
If you are thinking about building a digital assistant for your SaaS product, come talk to us! We can help you build a prototype in a matter of days. For FREE.