Virtualization Technology News and Information
VMblog Expert Interview: Zenhub Brings AI to Project Management

interview zenhub 

Zenhub announced a new AI roadmap that will serve as a co-pilot to help software teams reduce manual updates and improve productivity.  To find out more, VMblog spoke with Aaron Upright, Zenhub's co-founder.

VMblog:  Before we get started, can you give VMblog viewers a quick background on Zenhub and its role in software development?

Aaron Upright:  Zenhub is a project management platform that is designed specifically for software development teams. It is primarily used to plan, track, and report on the progress of development projects. We actually started as an internal solution at a tech incubator called Axiom Zen, which is also famous for starting Dapper Labs, the creators of CryptoKitties and NBA TopShot.

In brief, Zenhub enables software teams to build better code more quickly by providing a developer-friendly productivity and project management platform. Zenhub dramatically boosts collaboration and coordination for teams working in GitHub with automated agile features, real-time roadmap visibility, and team productivity insights. More than 8,000 disruptive teams worldwide, including NASA and Red Hat, rely on ZenHub to ship great code faster.

VMblog:  What is Zenhub AI?

Upright:  Zenhub AI is a series of new features and functionality that are designed to make developer teams' lives easier by eliminating the most tedious parts of the job, while also helping them to improve the quality of their output. We're starting with a focus on what we refer to as the "non-controversial" jobs that every dev team has to complete.

Software teams embrace a diverse range of agile methodologies, from Scrum and Kanban to SAFe and beyond. However, the concept of agile often sparks spirited discussions, leading teams to mold their own unique interpretations of the framework. Rather than adhering rigidly to the prescribed doctrine, teams tailor agile principles to suit their specific needs and circumstances. Amidst these variations, certain core "events" remain universally acknowledged and indispensable. Stand-ups, work prioritization, reviews, and retrospectives stand as cornerstones, offering a structured rhythm that fuels collaboration, accountability, and continuous improvement.

Though these events might assume different names - such as backlog refinement or sprint reviews - their fundamental purpose remains unchanged. Zenhub AI helps all software teams, regardless of which framework they follow, by facilitating these must-do events that need to be done in planning and tracking work.

VMblog:  How are Zenhub's customers using AI today?

Upright:  Our own research has shown that 3 out of 4 developer teams are actively using or investigating the use of AI, both for individuals writing code and for teams looking to improve productivity. We've come to the conclusion that the use of AI by developer teams is, for the most part, already here and will be standard for the industry in the near term. The real question is how teams leverage AI, and the true answer is that the industry is still in the investigatory and experimental stages.

The common theme among both individuals and teams is that AI has the potential to serve as a valuable tool in enabling developers to avoid doing the chores that distract them from what they really want to do, which is to write great code. Everything else that comes with the job; meetings, reporting, planning, tracking, etc. - really serves only as distractions.

VMblog:  What makes Zenhub's approach to AI different from other project management solutions?

Upright:  Zenhub AI focuses on the team rather than the individual, serving as a co-pilot to assist in all the ancillary activities to writing code, such as managing projects and tasks, categorizing and prioritizing work, and other productivity-related activities. Zenhub's approach to leveraging AI aims to simplify the day-to-day processes and operational overhead involved in software project management, which can often get in the way of shipping code faster.

Our approach to building AI is also different from most vendors today as we are building in ‘public' with complete transparency. Our objective is to thoughtfully build AI for real-life use cases by actively collaborating with our end users and customers. We've also announced an early access group, where participants work with the product team directly in shaping how Zenhub AI works. It's been an extremely rewarding way to approach this revolutionary new technology and is already paying dividends, including a closer relationship with our customers.

VMblog:  What models is Zenhub using to power these AI experiences within the product?

Upright:  Zenhub is actually using a mix of different types of AI to power this new functionality. Generative AI, sentiment analysis, and good ol' machine learning all factor into our design, depending on the necessary use case to which it is applied. Right now, we're using a mix of OpenAI's ChatGPT versions 3.5 and 4.0. We've found that while 4.0 is superior for some experiences, eg. summarization of information, 3.5 seems to be more performant and faster at other things like writing text. So we're using both for different use cases. We're also exploring IBM's Watsonx models and capabilities as a way to bring our AI experience to our on-premise customers. This has been a bit of a unique technical challenge as most of the models that are available don't support single-tenant environments.

VMblog:  What do you see as the primary impact of AI on the development world and how is Zenhub thinking about those concepts?

Upright:  We've found that the introduction of AI in general, in addition to its specific use in building code, has generated some interesting trends that many have not expected. For example, while AI could very well be the answer that developer teams have been looking for when it comes to improving productivity, it brings up new questions about data privacy, security, intellectual property, and ownership.

One of the surprising trends that has resulted from this focus is a renewed interest in on-premise software implementations. There remain a lot of open questions about data ownership and retention when it comes to both data that gets imputed into LLMs, as well as the corresponding outputs that get generated. While the general outlook towards AI seems optimistic, we're also seeing organizations take a "slow and steady" approach when it comes to their adoption of AI, along with AI tooling policies.


Published Thursday, August 17, 2023 7:29 AM by David Marshall
Filed under: ,
There are no comments for this post.
To post a comment, you must be a registered user. Registration is free and easy! Sign up now!
<August 2023>