Redefining How Developers Build and Test Code With Autonomous Testing

Reduce Testing Costs And Toll With This New Approach

TL; DR: For this article, we spoke with Adam Sandman, CEO of Inflectra, about Inflectra’s latest update release for its solution Rapise, a codeless test automation tool for dev teams. The update added generative AI capabilities to Rapise, redefining how developers build and test code.

I value transparency in all things, but especially when shopping for new products. If I’m spending money on anything, I would hope it works as advertised.

But unfortunately, way too often, companies tout their products as miracle solutions that can do it all and don’t really deliver those results. Take for example the one time I bought a dark spot removal serum and had to buy another complementary product to reap the full benefits.

Situations like this often happen in the tech and web hosting world where companies offer solutions that claim to be either more reliable or powerful than they are. Many AI-powered products have been marketed as hands-off solutions that can do it all.

That’s not always the case. Oftentimes, these solutions require human monitoring to catch errors and can cause more work for teams, including securing vulnerabilities involved with onboarding the solution.

Inflectra, however, tells it to you as it is. The software platform recently released Rapise v8.2. The latest version upgraded the solution with autonomous testing. The GenAI-powered feature helps users save time while testing and delivers what it promises — nothing more, nothing less.

Inflectra’s Rapise redefines testing with generative AI.

“Now, it’s not magic. Like any AI tool, it requires human intervention to take the automated test that it generates and give it feedback with a thumbs up or a thumbs down. You can train it, so the next time it will do it correctly and improve the model that’s being used,” said Adam Sandman, CEO of Inflectra.

It’s not rare for users to onboard an AI solution and think the solution has it all figured out. Although AI is powerful, it’s not exempt from making errors. That’s why Adam encourages teams to review AI-generated code or content regularly, no matter what solution they use.

But that doesn’t mean Inflectra’s Rapise falters in delivery in any way. It enables teams to accelerate testing and subsequently their development timeline with automation. Below, we’ll take a deeper look into the platform.

The Bottleneck of Software Testing

Inflectra has built its reputation and software on providing users with tools to accelerate development and time to market without compromising on quality. With its Rapise v8.2, it continues this vision.

“If you look at the software development lifecycle, testing is always the bottleneck. There was never enough time for testing. It doesn’t matter whether you have agile DevOps CI/CD, that is just a fact of development,” said Adam.

We’ve all seen the new AI copilots and tools that have hit the market and promised to improve productivity for developers. And they have — some even boosting dev productivity by 70% to 300%. But the problem is that there is also more code to test as a result.

“We created a bigger bottleneck that was already a bottleneck. And so what we found is that the ability to make software testing easier and keep up with the speed of development is becoming more and more important. We simply need to make tests more productive,” said Adam.

Rapise always had codeless testing automation. But the latest version brings the capabilities of generative AI to the table. Its new functionality enables teams to write tests in natural language as opposed to programming language. Rapise then turns that into automated testing code.

a screenshot of Rapise webpage
Rapise allows teams to streamline testing and reduce bottlenecks in their pipelines.

“So it makes the test creation much easier. The second part of this is that test automation is notorious for being brittle. When you write a test and make a change to your application, it breaks. So you spend more time fixing the test,” said Adam.

Rapise prevents this issue by quickly adapting tests to your application. Since the testing code comes from natural language, the natural language can adjust and Rapise will generate a slightly different code and make the test adaptive to your application.

Rapise also provides feedback, generates test data, and allows you to teach its AI how to do testing according to your needs. At the end of the day, Inflectra is helping teams not only be productive in development but also in testing.

“What we’re basically doing is making the cadence of testing match the cadence of development. So testing may still be a bottleneck, but it’s no more a bottleneck than it was before, as opposed to with copilots and development tools making testing an even bigger bottleneck,” said Adam.

How Rapise Fits Into the Agile Development Team

Adam shared a neat analogy with us to show how Rapise can fit into the modern development team’s workflow. He likened Rapise’s AI feature to an intern or a new employee.

“That’s the metaphor we tell people. A new employee may know about testing, but they don’t know your application, so they’re going to make mistakes. And so, you have to be patient with the new employee joining your team,” said Adam.

This is why human intervention is key to using Rapise successfully to improve testing productivity. You as the human basically serve as the supervisor over the newly employed Rapise AI. It automates the testing for you, but it’s your duty to review it for errors and make sure it’s ready to go.

“One of the dangers I foresee would be that it’s so easy to take what the AI does and as a human review and go, oh, that looks good. But a human has to go in there and start to interact with the data that’s come back from the automation to see what the AI might have missed,” said Adam.

Adam said project managers should actively prioritize reviewing the data they receive from AI automation, whether that’s with Rapise or any other AI-based tool. If not, development teams run the risk of putting out lower quality solutions.

The Stakes Are Higher Than Ever

Adam told us developers love to problem solve over anything else, and Rapise allows them to do just that. Rather than doing mundane tasks, developers can focus on high-value tasks and find errors that could be causing latency within their applications.

“The overall Inflectra platform lets you take away a lot of the mundane tasks, like generating tasks, generating sample stub code, and a lot of the things that are effectively copy and paste that developers do,” said Adam.

We also asked Adam how users can maximize the potential of Rapise in the era of innovation. He said teams often take the development approach where they take the nearest application, copy and change it, and start building a new application from there.

But he said to be more productive, developers can use AI to build the foundation of their applications instead.

“They can use AI to build the skeleton or the scaffolds of their building, and then they could focus on the drywall and the glossy finishes,” said Adam.

Rapise provides tools to generate code and testing and find errors in applications. It also has a nifty feature that graphically shows you where mistakes are located in your code.

“So it shortens that conversation between the developer and the tester, which is often frustrating, especially in terms of agility. So it provides more intelligence and insight to reduce the latency between them,” said Adam.

As software becomes more embedded in all aspects of life, quick automation will become more and more essential to the success of overall systems, including in IoT, trading, and healthcare. Cybersecurity has also forced companies to do updates more frequently. Adam told us the stakes have never been higher.

“The cost of failure is getting higher. So the trends are we’re automating more, we’re doing it quicker, we’re doing it early in the lifecycle while doing less comprehensive testing and more testing per release,” said Adam.

AI can help here by allowing teams to prioritize efficiently. It can help users pinpoint risks, choose what to test more frequently, and apply a higher risk approach to more critical areas.

New AI Initiatives on the Way for Inflectra

As it happens, Inflectra has some new AI initiatives on the way. Adam told us some are being developed in house while others have been developed through partners.

“On the third-party side, we’ve got some functionality coming that will allow AI to review requirements and system information, score, and give basic quality control,” said Adam.

This feature will eliminate issues around requirements engineering, which Adam said is often the bane of software testing. He said if the requirements are inconsistent or ambiguous, that can lead to low-quality software. This solution will work to prevent those problems.

Another third-party tool in the works is a code scanner to determine where tests should be run. It will use historical data and code changes to tailor tests to code pushes.

“So depending on what your developers do, it will real-time decide to subset your tests. That means you’ll go from maybe an hour of testing time to maybe two or three minutes. So it should shorten the cycle and optimize the test through the ones that need to be run,” said Adam.

As for in-house development, Inflectra will expand its generative AI capabilities. It will be adding an interactive assistant to give users real-time feedback for their project management needs. It also has a forecasting solution on the way.

“Gen AI will allow us to give users actions and insights they could take to improve quality or avoid risk in the future. So it could say here’s a risk that’s going to potentially happen in a month’s time and here are some steps you could take now to reduce the likelihood. So, much more data forecasting and analytics,” said Adam.