 |
VOOZH | about |
|
VOOZH | about |
We’re so glad you’re here. You can expect all the best TNS content to arrive Monday through Friday to keep you on top of the news and at the top of your game.
Check your inbox for a confirmation email where you can adjust your preferences and even join additional groups.
Follow TNS on your favorite social media networks.
Become a TNS follower on LinkedIn.
Check out the latest featured and trending stories while you wait for your first TNS newsletter.
Jyoti Bansal is the CEO of Harness, a DevOps company that has recently fine-tuned its tagline to “the industry’s first Software Delivery Platform to use AI to simplify your DevOps processes.” Harness isn’t the first company to pivot hard towards AI, of course, but given its focus on continuous delivery, I took the opportunity to interview Bansal about how AI is now used within Harness and by its customers.
I began by asking Bansal what impact generative AI has had on developers and their workflow.
He replied that AI can alleviate the tedious and repetitive tasks involved in the software delivery lifecycle, starting from generating specifications based on existing features, to writing code. Also, he said that AI can automate code reviews, vulnerability testing, bug fixing, and even the creation of CI/CD pipelines for builds and deployments.
“We can easily get 30-40% efficiency in every part of the developer workflow by intelligent use of generative AI,” Bansal said.
Since Harness itself makes heavy use of AI technology, I asked Bansal for details on this.
One key area where AI is utilized, he told me, is in ensuring that code changes do not negatively impact performance, quality, or security. Harness leverages AI models within their continuous delivery pipelines, he explained, which compare code changes against data from monitoring and logging systems like DataDog, Azure Monitor, and Splunk. He said that these AI models can identify any potential issues before code changes are deployed into production in their systems, which enables rapid and reliable deployment pipelines.
Another AI technique employed by Harness is what Bansal called “test intelligence.” This addresses the common challenge of lengthy test execution times, he said. By using AI models, Harness identifies the parts of the code that correlates with specific tests, allowing its developers to optimize the tests that need to be run. Instead of running a large suite of tests for every code change, Harness can determine the specific tests required for a given code change. This significantly reduces test execution time and increases developer productivity.
“We have many other smaller techniques that we have been building to optimize things,” he said.
Bansal also confirmed that it is actively developing generative AI-based approaches to further enhance its platform, but he wouldn’t give any further details about these as yet unannounced features.
One of Bansal’s adages about AI is that it should be embedded in B2B products, rather than sold as a separate add-on. In customer support, he explained, AI can be used to generate appropriate responses to customer queries. Similarly, in sales, AI can assist with outreach campaigns to potential buyers.
As for developer products, Bansal noted that customers are increasingly demanding AI capabilities be directly integrated into the products they purchase, rather than relying on separate AI solutions. He said this approach ensures that AI is tailored to specific use cases and domains.
I asked if Harness has any features that help with the new role of “prompt engineering,” which many developers are now adapting to (for example when getting code completion help from GitHub Copilot).
Bansal acknowledged the importance of prompt engineering and what he termed “multi-agent techniques” in the DevOps process. He said that Harness is actively working on incorporating these concepts into their platform, but as yet it has nothing to announce. But his goal, he said, is to create a model that facilitates the developer workflow by utilizing agents for various tasks. So you’d have agents that create code specifications, write code, generate test cases, and perform code testing.
Bansal is also the CEO of a newer company he co-founded, called Traceable (“Intelligent API Security at Enterprise Scale”). Traceable watches and analyzes every API call, ensuring that sensitive data is not inadvertently sent to GPT, APIs, or other endpoints, he explained. It leverages AI technology to do this monitoring and to detect any potential security breaches or data exfiltration attempts.
Traceable’s AI models are built on an API data lake, Bansal said, which allows security teams to identify and address security challenges in real time. He added that the demand for securing generative AI-based traffic is rapidly growing, especially in industries like banking and financial services where data privacy is a significant concern.
Bansal advises IT departments in enterprise companies to leverage generative AI capabilities offered by their vendors, rather than attempting to build them in-house. He suggests that companies focus on integrating generative AI in areas that are not their core business — such as developer tools, sales tools, marketing tools, accounting, and HR. By pushing their vendors to incorporate generative AI capabilities into these products, companies can benefit from the advancements without investing significant resources, he said.
However, Jyoti also noted that it is worthwhile for companies to explore the use of generative AI in their core business areas, where it can enhance customer experiences. For example, banks can consider using generative AI bots in online banking, and retailers can explore generative AI-based personal shopping assistants.
As for developers, he emphasized the importance of learning how to use AI effectively and incorporating it into every part of the software delivery lifecycle. By harnessing generative AI, he said, developers can be more efficient and complete tasks in less time.
