Riter AI - Documentation - Riter

Documentation of riter.co — project management tool

Riter AI

Riter software includes a strong artificial intelligence which allows you to organize the workflow with minimal efforts and mistakes. It is now in an active development stage. We are teaching it how to deal with the most of project management duties so that you could free users from complex calculations and long disputes about the development of a project. Riter AI is designed to automate sprints planning, reduce tasks estimation errors and provide a reliable work. Its effectivity is increasing in the process of use as Riter adapts to peculiarities of the team and projects. Meet the release of described AI features in the distant future.

Stories estimation

Riter helps you to turn an approximate subjective task implementation time into a sufficiently accurate estimate that can be used for any long-term planning of tasks. First of all, it insists on a concurrent task estimation by different users to improve the accuracy of the assessment. Moreover, Riter takes into account the individual characteristics of these developers, such as productivity and accuracy coefficients, to define the reliability of their estimates. The coefficient of accuracy depends on how much the actual time spent on work differs from estimated time for previous tasks of the user. The productivity characterized by the ratio of clean and dirty worked hours of a particular user. To determine these values, Riter takes into account all earlier experience and gathered statistics. Thus, it always knows exactly when a task estimate can be considered dependable, and when it still needs to be clarified. Riter also reckons for developers load during the current sprint, their employment in other projects, average number of worked hours per sprints and much more.

Sprint planning

Riter gathers and calculates all necessary information about developers' productivity, workload and planned tasks to predict sprint success or failure and define real deadlines which you can meet. The AI also remembers your team velocity, knows real tasks estimates, which allow to increase its prediction ability. If you are trying to schedule for a sprint more work than you are able to do (according to Riter's AI), it warns you about this so that you could redistribute tasks among sprints and developers in time and avoid breaking terms. In this way, you don't need to keep in mind the errors of your developers' estimates or hope that all tasks will be performed without any complications - they won't. All that is required of you is to create tasks and work on them, while Riter monitors the workflow and ensures its efficiency.

Time intervals generation

Riter software can be run on users' computers and monitor their activity during the workflow. This allows to generate time intervals automatically, that will ensure greater accuracy and reliability of the information provided. With Riter developers don't need to be distracted from work in order to fix their current results. The only thing they have to do is run a background service which will do most of the necessary steps for them.

Riter is able to monitor the screen content using text recognition to determine what you are currently working on. It defines keywords (used web sources and applications, content on opened pages) and compares them with topics found in the stories tracker. After finding a match, Riter can change the story state to "In Progress" and start adding time intervals to it. It remembers the time during which the user is working on a specific task.

Riter is also able to distinguish clean and dirty hours spent on a task. Time spent surfing the Internet, reading documentation, discussing problems on forums, skype conversations can be identified as dirty hours. While the development environment is used or the project software is run Riter increases clean spent hours. Thus, while the developer is working, the tracking service is able to determine a task, a date, spent time and related keywords (which will be added as the interval description). Then Riter adds a corresponding time interval for the specific story. Sometimes if it is not sure what users are working on, it asks for solving this question by the developers themselves.