isento takes a human approach to testing a robot with Qase
isento is a full-service IT provider offering business analysis, requirements engineering (RE), development, testing, project management, and agile coaching services.
At isento, testing is more than just a technical exercise — it’s about creating a system that all team members can understand so that there is more room for collaboration and innovation. To learn how isento manages their testing, we talked to the Head of Test and Quality Assurance, Dr. Christian Brandes.
isento’s Head of Marketing, Anika Danner, also joined the conversation to share details about one of isento’s most innovative projects — a 3D printable humanoid robot that anyone can build, train, and program themselves.
isento needed a tool that they could tailor to their work
When Dr. Brandes joined the team, one of his key responsibilities was to increase testing maturity for isento’s internal projects. To achieve this, the team needed a tool that allowed them to custom-fit the setup to their specific testing process.
For isento, the tool requirements were:
- Integration that allows for requirements traceability (in their case, via Confluence)
- Test suites for managing and structuring test conditions
- Integration with Robot Framework for automation
- Possibility for keyword-driven testing
- Support for data-driven testing
Dr. Brandes explained that while many test management tools claim to offer requirements traceability, the actual functionality is typically limited: “Most tools allow you to add traces directly to single test cases, but not to test conditions or folders containing all the test cases designed to test a given requirement. So, you end up having to enter the same link again and again — dozens of times, typically — to every test case that belongs to an identified test condition, and we didn’t want to work that way,” says Dr. Brandes.
Instead, isento was looking for a tool that offers the possibility to add links to the test condition itself, so the desired link only needs to be entered once. “Of all the tools and plug-ins we evaluated, Qase proved to be one of the few tools that offered this simple but crucial feature.”
This unique ability facilitates better collaboration between the QA, RE, and development team members within isento’s agile teams. They don’t have to go back and forth hunting down requirements and specifications because all teams have full visibility. To get there, isento’s testers work closely with the project’s RE specialists to understand their concepts and artifacts and to jointly define the necessary traceability approach offering the most valuable granularity and seamless tool support.
“With Qase, we have bidirectional traceability. Test cases in Qase have a link to the requirements in Confluence. If you’re working in Confluence there are links to the test cases, so in either tool, the information is one click away. Qase is the only tool that lets us do it our way.”
He goes on to explain how bringing all work into Qase allows them to create and manage a testing plan for their 3D-printable intelligent robot, pib.
A test strategy for a robot
First, the team had to create a testing strategy for pib to determine what to test, which skills each test requires, when to test hardware versus a placeholder or simulation, how to test parts in isolation and interaction with each other, etc.
“The goal is to start with as many quick and cheap tests as possible in the beginning and work up to the more expensive tests that involve the whole robot,” explains Dr. Brandes.
isento relies on Qase to create test suites for managing and structuring test conditions and integration with Confluence via Qase's open API to keep all team members looped into the work. “We strive for a lot of pair testing — a developer working with a tester,” Dr. Brandes says. Developers note what they’ve covered through unit tests while testers explain the test cases from a user perspective.
For isento, how they describe testing activities and write tests is crucial. “We write high-level tests in a way that a human can understand them, not in terms of clicks or technical steps,” says Dr. Brandes.
Qase’s Shared Steps make writing tests for humans easier
The isento team embraces the well-known concept of keyword-driven testing, which involves describing test cases in a way that humans can understand and search for them rather than using very detailed, technical terms to describe test cases.
Dr. Brandes explains that when you write test cases in terms of buttons and mouse clicks, you end up with very elaborate descriptions of UI activity, resulting in lengthy test cases that are closely connected to the UI. “So, whenever the UI changes, you have to go through hundreds of test cases and update them,” says Dr. Brandes.
Qase’s Shared Steps are a critical component of isento’s keyword-driven testing approach as they allow the team to describe and group steps in a human-readable and reusable way. For example, consider a login screen. If each UI action is described in granular terms for every test case needing a login first, any time the login screen UI changes, you’ll have to find and update each test case that involves a login screen.
By describing the test in human terms like “login user,” and using Shared Steps to group all the UI actions linked to a login page, testers can easily search for the Shared Steps group, update it once, and all the test cases using the Shared Steps will be updated automatically. This dramatically minimizes maintenance and automation costs for new test cases.
“With Shared Steps, you can define steps on a higher level rather than in granular terms. They are a container for more detailed steps. And you can reuse these Shared Steps in several test cases. This is quite crucial for us because we typically have hundreds or even thousands of test cases in real-life projects.”
isento’s keyword-driven testing is ideal for scaling automation
Interestingly enough, the human-centric approach that keyword-driven testing provides is ideal for testing a humanoid robot and increasing automated tests over time.
Dr. Brandes uses the example of programming pib to move a finger 45 degrees. A human would understand what “move finger 45 degrees” means and know that they’ll need to test all 10 fingers. To save time and effort, the team writes a few lines of code to explain what “move finger 45 degrees” means to the automation framework. Then, they can search for Shared Steps that mention “move finger,” and easily convert manual tests to automated.
“Most test automation projects fail because maintenance effort increases over time — you have too many cases to update whenever the software changes. Keyword-driven testing is a technique to minimize this effort. At isento, we won’t test any other way than keyword-driven testing, which Shared Steps enable. We love it, we breathe it, and we’re wholeheartedly convinced that this is the best way to do it.”
Dr. Brandes adds that the impact of keyword-driven testing is exponential when it comes to their automation efforts. “If you create a new test case using these human-understandable keywords and those keywords have been automated for other test cases before — like we did using the Robot Framework — you can automate new tests much faster. ”
pib is bringing people together around the world
isento’s human-centric approach to testing can be found in pib's general concepts as well. The robot is designed to be customizable, adaptable, and accessible to various age groups and technical skill levels. “Anyone can 3D print their own pib and adapt it,” explains Anika Danner, Head of Marketing at isento, “It’s an open-source project. We have a community of 1,500 people who are building their own pibs at home or school.”
Thanks to the community-driven work through the open-source project, pib will be able to be programmed or trained to help with everything from fetching coffee to tutoring students. The pib@school initiative is particularly impactful. It provides students with hands-on experience in robotics, AI, and 3D printing. Schools can build and program their version of pib, allowing students to get creative with how they test and program their humanoid robots.
Two students, Helena and Anna, won multiple prizes for their pib adaption — an educational robot designed to tutor other students. Helena and Anna equipped their pib with an AI language module and (after rounds of testing and feedback) designed their pib to offer tutoring in German, English, and Math using advanced voice-assistant technology.
Their incredible work earned them first prize in the Jugend Forscht Mathematics and Computer Science category of the “Schüler experimentieren” junior division. After winning the regional competition, Helena and Anna went on to win a special prize for Human-Machine Interaction.
Start working with a TMS that adapts to how you work
Tools should adapt to how you need to work, not the other way around. Schedule a demo to explore how Qase’s intuitive, all-in-one QA platform can help you deliver work efficiently to deliver higher quality software faster.