fibre lights

Meet Eggplant GAI, the AI-Augmented Powered Test Platform That Harnesses All the Possibilities of Test Automation

You’ve probably heard someone say recently, ‘You won’t be replaced by an AI, but you will be replaced by a human using an AI.’

If that’s true, we have to ask ourselves how we, as testers, can make use of generative AI, harness its superpowers, and thus ensure we keep pace with the progress of the test industry.

The answer, I think, lies in how we use generative AI in test automation. Let's take a step back to understand why that’s the case.

The Trouble with Test Automation Today

Figures from Forrester suggest that the industry average for test automation is just 22%. You might think this sounds incredibly low. You might have test projects that are 70% or more automated. It’s certainly true that some projects are highly automated.

But across the entire landscape of test requirements in a typical enterprise, many projects are still manual – which is reflected in the 22%. And there are very good reasons why automation isn’t being embraced more widely too.

The Efficiency Issue of Test Automation Coverage

For the past four decades, automation evangelists have focused on the time-saving potential of test automation. We've shown how, by capturing the tests and creating a model out of them, you reduce the time needed on future test runs.

But a chunk of time is still needed to understand the test requirements and generate the necessary test. A typical core IT project involving business technology will involve the digital grind of communicating with multiple teams and understanding multiple use cases. Automation in its current form can’t help with that.

And then consider the maintenance requirements. if your test automation platform doesn’t self-heal test assets, the more tests you automate, the more tests need to be manually maintained and documented.

The faster the software development lifecycle becomes, and the more we shift left in our development practices, the bigger these issues become.

For me, the biggest potential use case for generative AI lies in how it can be harnessed to carry out test design and test maintenance alongside test execution. It’s how we, as testers, will be able to keep pace with DevOps processes and benefit from the extraordinary potential of generative AI.

However, it’s not as simple as using a public-facing generative AI.

The Risks of Open Generative AI in Software Testing

You’re no doubt familiar with the privacy and data leakage problems of feeding IP into open generative AI platforms. Looking further, my sense is that legislation, such as the EU AI Act, will limit the value of any open AI in any case.

Like you, these are issues that our customers are very aware of. They want to use the power of generative AI, but they don’t want to connect their own systems and data to any of the open tools that are available.

The solution is to have a dedicated generative AI tool deployed within an organization’s walls, keeping it 100% secure. An offline generative AI that can be deployed on a standard laptop or put into the cloud as part of a dedicated and secure AI infrastructure.

This is where Keysight’s Eggplant Generative AI comes in.

Enter Eggplant Generative AI (GAI)

Eggplant GAI is the local generative AI the test industry needs.

When it’s first out of the box, you can think of Eggplant GAI as the equivalent of a highly experienced human tester who’s new to your team.

It comes to you as a specialist large language model trained on trusted material in multiple areas:

With this knowledge, we can think of Eggplant GAI as being at the bottom of the knowledge graph. The next step is the training you give it on everything in your organization to take it to the top of the knowledge graph.

You connect Eggplant GAI to your systems and give it all your structured and unstructured data so it can understand your projects, your teams, your confluences, and so on and so on. (And remember, because it’s offline, it’s secure.)

Like every organization, you’ll have unique quirks that don’t conform to standard interpretations. Eggplant GAI will question these and ask for definitions so it starts to understand your particular organization, something it will continue to do throughout its life.

Equipped with this generalist and specialist knowledge, Eggplant GAI is a tool that gives you a much better quality of response than you’d get from an open generative AI model. It’s the AI equivalent of the person who knows everything there is to know about all of your systems and why certain things work in certain ways.

It gives you a genuinely valuable assistant for the humans in the loop.

Here’s How Eggplant GAI Helps Humans in the Loop

Eggplant GAI has a retrieval augmented generation capability, so you can use it in the same way as you’d use an open generative AI to carry out prompt engineering or tuning exercises. But its capabilities go far beyond simple questions and answers.

You can ask it to do test design and to generate manual and executable tests. It will tell you from a risk-based testing perspective which tests to run first and why. It will do test deduplication and test case optimization. If there are changes to the requirements, you can ask it to regenerate all the tests or tell you which tests need to be updated and which lines of code.

In short, its ability to go through paths, discover paths and understand systems moves you further up the Automation Maturity Model index (AMMi). It also allows you to automate more of your software testing requirements across your organization.

Achieve the Full Potential of Test Automation

Eggplant GAI means test automation can finally achieve its full potential – as a solution that delivers huge productivity gains.

To understand how Eggplant GAI works in more detail, watch my Automation Guild 24 session AutoOps: Harnessing the Power of AI-Augmented Testing with Generative AI.

limit
3