AI technologies which could enhance Riter chatbots intelligence
Chatbots are not a new phenomenon at all, but in the last year they have attracted especially a lot of attention. According to Business Insider research, by 2020 80% of companies will use chatbots for different needs and this is not surprising. Chatbots are being used in many different ways to automate a significant part of our daily routine, help businesses to reach customers, save money, integrate apps with existing solutions and so on. However, we share the opinion that we have the attention span of a gnat. The fact is that the potential of chatbots is huge, but we continue to use their capabilities in a minimal way. If we direct all the power of AI, which we have already achieved and successfully use in other spheres, to chatbots, the level of business profitability will grow before our eyes.
What exactly could be provided in chatbot technologies to make them intelligent enough? Here are some ways for this purpose:
Semantic parsing. This technology helps AI-based chatbots to understand human speech without forcing people to memorize special constructions, recognizable by AI. With a power of semantic parsing, a manager could write in a common chat, say, "Tomas, you should start working on the landing page when Mike finishes its design" - and a chatbot will be able to create an appropriate task, add a short description, tags, connect it with a corresponding Mark's task if necessary, and assign Tomas to this new task. If AI is smart enough, it could even estimate, when Mark is able to complete his part of job and when Tomas can start working on the website, taking into account his current work load in other projects. Of course, this is impossible just for chatbots, without an AI-based project management tool. In Riter, different AI capabilities are successfully combined and work together, so chatbots may be really helpful.
Automated planning. This feature could let chatbots define necessary steps to achieve a desired goal. It would be really nice if we didn't have to tell chatbots what they should do. Instead, they could figure out by themselves the best sequence of actions in each situation. In other words, we don't need to explain a chatbot what it should do in every possible case. We teach it to adjust its behavior to changeable conditions, describe common rules and let it make its own decisions. Continuing the theme of project management, let's imagine that a developer has worked on a task. AI was watching his actions, recognizing the contents of the screen, calls and messaging with colleagues to understand what a developer was working on. When a developer opens a chat, an appropriate template of report is ready to be sent to the project management software. Time spent on tasks, solved problems, done work are described in a necessary format. The only thing a developer should do is to check everything out and send (or even this is automated). All interested parties are already informed about changes, the task state is changed, subtasks are added, unclear questions are left in a common chat and so on. This is not easy to implement but we believe that one day Riter AI will be able to cope with this challenge.
Natural language generation. This technology allows chatbots to respond to people in their own language. The project can involve people from different countries who are used to different terminology and style of communication. Of course, we can make them adapt to the "language" of the chatbot, memorizing keywords and using template constructions. But often communication can require more detailed information, not be limited to a fixed set of commands. In this case it would be great if our queries are translated into a set of database requests and results then are summarized according to our needs.
This is what we are going to do in Riter - take full advantage of artificial intelligence achievements to ensure complete integration and automation with your workflow. Riter AI along with extensive chatbots support will be responsible for organizing the workflow with minimal efforts and mistakes on the part of developers and managers. A store of bots and a powerful GraphQl API will make it possible to integrate Riter with any existing software such as Slack, Github, Telegram and so on. Riter AI will learn to deal with the most of project management duties so that users could be freed from routine work, complex calculations and long disputes about development process. Based on the successful experience of Jira's add-on market, we can say that chatbots are the future of project management. Here are some possible use cases of Riter chatbots:
- Chatbots for creating and modifying tasks during your conversation with colleagues without having to open the application.
- Chatbots for adding comments, topics, estimated time for tasks, for example, analyzing Slack discussions or Skype calls.
- Chatbots that assign developers to tasks with a chat command.
- Chatbots for automated time logging by monitoring user's activity during the workflow and recognizing screen content.
- Chatbots for notifying users about events, deadlines and last changes in the project on demand.
- Chatbots for learning to work with the program and getting help in the process.
Thus, chatbot opportunities are endless, at the same time, they are limited by developers. It depends on us if we will make full use of their power or expect users to adapt to our primitive tools. In the next 5-10 years the software market will change a lot, as well as project management in companies. The main trends in project management will be more remote teams, a great impact of artificial intelligence with maximum automation of the workplace and large-scale integration of various fields of activity. The role of chatbots and instant messaging in the workflow will continue to increase, while the need for a structured way of task planning will remain in force. We are sure that AI-oriented Riter project management tool is going to fully meet the requirements of this coming world.
Riter development team