The Future of DevEx: 7 Things to Expect From AI-led Software Development

An Accenture study found that 73% of organizations are prioritizing artificial intelligence (AI) over all their digital transformation investments in 2023. Developer experience too is a top priority with 58% of engineering leaders admitting that developer experience is extremely critical for the C-suite. 

The icing on the cake, the two priorities i.e., AI & developer experience are symbiotic as well as complementary in nature. 

Improving developer experience can enhance the efficacy with which your engineering teams adopt new AI technologies. When harnessed right these AI technologies can help you enhance the developer experience (DevEx) you deliver. 

This article dives into the intricacies of the what difference AI-led software development will make on developer experience and productivity in 2024 and the times to come.

Why should engineering leaders be keen on AI & DevEx in 2024?

Today, it is important for engineering managers to understand how AI influences developer experience as stakeholders at almost every organization are getting restless and are eagerly eyeing results from AI initiatives. Tech managers are being bogged down now and then for updates on where their org stands in the AI adoption journey, and how well-equipped your org is to embrace the digital disruption AI will unleash. 

Anyway, no matter how curious the stakeholders get, we can’t blame them. Not their fault. 

Artificial intelligence is the latest tech frenzy and everyone seems to be engrossed in it. The fervor surrounding AI is showing no signs of waning. It’s breaching new heights every day. 

Thanks to AI, today, the entire tech industry is at an exciting inflection point. 

Mostly, the views around AI are optimistic. Dan Ives, managing director of equity research at Wedbush Securities, says that AI is the most transformational tech in the last 30 years. Sundar Pichai, Google’s CEO, holds a similar view. He said that AI is “the most profound technology humans are working on”. 

But the fear lingers too. 

Is AI the new tech bubble?  Is AI just a fad? Or is AI truly a transformational tech? 

Well, time is the best storyteller, and it will pass its verdict. For now, AI seems to be capturing everyone’s imagination— from writers & journalists to photographers, filmmakers, scientists, software developers, and engineering teams at large.

And rightly so. AI gets the closest to the essence of human intelligence i.e., our neural system and cognitive system, and how we reason based on symbolic representations. 

Does that mean AI will become sentient at some point? 

For now, companies are looking to ride this AI-led digital transformation wave and squeeze in more revenue and profits by leveraging AI technologies. The industry as a whole is working towards hacking AI solutions that make developers and organizations build more innovative solutions faster and more responsibly. 

And as mentioned earlier, one of the core fronts where AI is leaving its mark is developer experience & productivity. 

However, to better grasp AI’s influence on developer experience, you must understand how developers are using AI i.e., the key value propositions & benefits of AI for developers, and how AI impacts developer productivity. 

How Developers Are Using AI Tools For Software Development?

As per Github, 92% of developers in the USA are already using AI coding tools. 

Developers are using Amazon’s Codewhisperer, Github’s Copilot, Google’s Codey, Tabnine, Codeium, and a plethora of AI tools for diverse use cases in their SDLC workflows. But a majority of developers use AI tools for the following-

• Writing code and debugging code

• Code documentation 

• Knowledge Transfer

• Testing code & Data

• Code reviews

• Deployment & Observability

• Architecture Designs

In short, AI tools are developers' new best friend and their 24*7 pair programmers. Why else do you think they named it Copilot?

Are Developers Using AI Tools?

• Globally, 44% of developers are already using AI tools in their SDLC workflows. Mostly, they use ChatGPT (83%) and Github Copilot (56%). 

• 83% of developers who have embraced AI developer tools are using them for writing code, and 49% for debugging code. 

• The adoption among upcoming developers i.e., those who are learning to code is higher, 55%. 

Are AI tools Improving Developer Productivity?

These AI coding tools boost developer productivity effectively if the developer is well-versed in his/her tech stack. Otherwise, AI development tools can be equally counterproductive.

A real-life example of AI tools being unproductive.

Let’s say, you are working on developing an intelligence tool that tracks brand mentions across the web. For that, you are supposed to scrape content from the web. You decided to use DynamoDB (a managed AWS database) & Scrapy (an open-source data extraction framework). To speed up your process, you enter a prompt that says-

“I want to use DynamoDB with Scrapy. Give me the code to integrate both.” 

ChatGPT or your code generation tool of preference might give you a solution that uses Boto3 (an AWS SDK for Python). As that’s the most popular one. 

But what if you are not well versed with Boto3? 

You will struggle to use the ChatGPT-generated code. You won’t understand the complete functionality or what the code is doing. You don’t even know if it’s accurate or if it will introduce vulnerabilities. 

In such cases, the AI tools for developers prove to be counterproductive. But let’s say, you are versed in pynamodb (a pythonic interface to Dynamodb). Then you can specify the same in the Prompt. “I want to use DynamoDB with Scrapy Framework. Give me the code using pynamodb.” 

Now, as you’re specific with what exactly you need from your AI developer tool, it may help speed up your coding process and subsequently improve your overall productivity.

A McKinsey report reaffirms this. AI does help speed up code documentation by 50%, code generation by 45%, and code refactoring by 30%. However, when the complexity increases it’s not that effective. In fact, for junior developers, it sometimes takes 7% - 10% longer than without using AI tools. 

Because of this, some developers refrain from using AI tools for high-complexity coding tasks. Is this all? Or are there more pieces in the AI adoption puzzle that are detrimental to the adoption of AI development tools among software developers?

Is There Resistance to AI Adoption Among Developers?

For now, trust is a key challenge in AI adoption among developers. 

As per the same Stackoverfow report, AI is afflicted with Helmsman’s complexity cliff. Basically, after a certain complexity level, AI’s ability to generate accurate solutions drops off.

Low confidence in the accuracy of AI developer tools is a concern among developers-

• 27.2% of developers don’t trust it

• 39.3% do trust the code generated by these AI dev tools to an extent

• 2.85% highly distrust it

But trust isn’t the only impediment, local laws, policies, workplace regulations, privacy, and copyright infringement challenges are keeping developers from rapid AI adoption in their SDLC workflows.

Artificial Intelligence (AI) and Developer Experience (DevEx)

Before we can fathom AI’s impact on Developer Experience, it is important to understand what exactly DevEx means.

In simple words, Developer Experience encompasses every single aspect of software development that influences ‘how a developer feels’--- from how requirements gathering, development, testing, debugging, deployment, and maintenance processes are implemented to the tools they use at each of these stages and the organizational culture, SDLC practices, and employee experience, just everything. 

Well, the best way to improve the developer experience is to minimize the developer friction and developer toil. If you’re a regular reader of Hatica Insight, you would already know that the secret recipe to doing so is to deeply embed intelligence (AI) & automation into your SDLC stages. 

That brings us to the meat of the insight. 

Developer Experience In 2024: How Generative AI Is Making A Difference

AI is influencing almost every aspect of the SDLC. From architecture design decisions to code generation, debugging, testing, deployment, monitoring, and maintenance. Let’s explore them one at a time-

1. Optimized Requirement Gathering & Architecture Decisions

35% of developers believe planning & requirement gathering meetings to be highly beneficial for effective team collaboration and communication— key instruments for developer experience. As GitHub puts it-

Today, you can build AI development tools that would feast on the market signals stored in your data lakes to intelligently gather software requirement data by analyzing market trends, customer feedback, and user behavior in real time. 

AI also assists with architectural decisions. The AI tools can analyze user stories, and project requirements, and accordingly suggest design patterns that are optimized for system performance and scalability. 

Thus, AI dev tools can help save crucial time spent in requirement gathering or architecture design, and facilitate developers to discuss important design decisions in these initial meetings for enhanced efficiency. 

2. Code Generation AI Tools To Expedite Feature Development

87% of developers reported AI coding tools to help them preserve mental effort while dealing with repetitive tasks. 

Wow! Isn’t that awesome for DevEx?

In general, AI code generation tools are brilliant at generating code if developers are good at writing prompts that define the specifics precisely. Github’s Copilot & Amazon’s Codewhisperer are well-established industry-leading AI-led code completion tools. 

These are trained on Billions of lines of code and help autocomplete code based on the existing code your developers may have written. These tools integrate well with your IDEs and can also accept prompts/instructions in natural language to generate code. 

Such tools improve developer productivity by as much as 57% thus saving a lot of developer time to focus on more exciting innovations & challenges. Thus, generative AI code completion tools amplify positive developer experience.

3. AI Code Debugging Tools Save Developers from Smashing Their System

31% of developers identified finding & fixing vulnerabilities as their top daily task. 

It’s not a distant past where a missing semicolon becomes a perfect manuscript for a spine-chilling horror story. Accidental logical errors of creating infinite loops, and small security vulnerabilities that open the door to SQL injections and XSS attacks have always been a nightmare for developers. Thankfully, AI tools are good at comprehensive analysis of the code and can decipher common errors and vulnerabilities easily. 

Not just that, these tools also understand how different functions, classes, and files are interrelated. Hence, they come in handy to developers for knowledge transfer. These tools facilitate developers to understand code written by somebody else. These AI coding tools are good at explaining the nuances, and specifics of a codebase. Helps in quick onboarding of new developers to the ongoing project. 

In a way, utilizing AI development tools helps reduce the learning curve for developers. All this improves the overall developer experience. 

4. AI Code Review Tools To Ensure Flow State Continuum

25% of developers reported spending most of their daily time waiting on code reviews. 

Sad, right? 

Code reviews painfully extend the SDLC timeline. Senior developers often don’t find the bandwidth to perform code reviews. Any negligence during code reviews can introduce challenges in the pipeline at later stages. AI code review tools speed up the review process for developers with no time to go through the entire code by highlighting high-severity issues and any deviations from the defined coding standards. 

The AI code review tools may even add actionable, descriptive, and helpful comments. Thus, streamlining the SDLC processes, reducing the developer toil, and ultimately enhancing the developer experience. 

5. AI Testing Tools Garner Quick Feedback with Synthetic Test Data

AI-led software test automation tools can–

Auto-generate test scripts.

• Prioritize using machine learning algorithms to know which scripts are important to be executed immediately.

• Self-trigger the testing execution.

• Auto-generate synthetic test data too. 

AI testing tools can be augmented to analyze test results, generate smart reports, and give you suggestions on what needs to be done to improve the quality of software. 

So, embracing AI software testing tools in their workflows helps developers deliver quality code by improving code coverage and proactively detecting defects early in the development phase. Thus, reducing tech debt, rework, and context-switching incidents.

6. AI Code Documentation Tools Help Developers Invest Maximum Time in Cognitive Work and Innovation

Code analysis and content generation capabilities of generative AI-powered code documentation tools improve collaboration among team members and simplify repository maintenance.

7. AI-powered DevOps & SRE Ensures No Red Sirens During Wee Hours

In GitHub’s Developer Experience survey, 27% of developers identified waiting for builds & tests from other teams outside of CI as the top productivity killer. 

AI coding tools can help mitigate this as AI reimagines DevOps and SRE toolchains with enhanced capabilities and effectiveness. These tools help automate deployment, and predictively help keep the infrastructure safe with real-time monitoring, anomaly detection, and issue resolution. Read Hatica’s insights on AI for DevOps for more details.

The Path Forward For Engineering Leaders

AI saves developers from mundane tasks and gifts them the joy of working on cognitive & creative tasks, ultimately making software development more enjoyable and less painful. 

By 2025, as per Gartner, 80% of SDLC workflows will be using AI. 

With an increased proliferation of AI tools in the SDLC workflows, the roles of developers will evolve to be more of the guardians, the orchestrators, and the integration engineers. It becomes crucial for engineering leaders to invest in developer experience to -

• Unlock innovation and operational excellence

• Steer your organization toward being a market leader 

You need to embrace the change and make AI your competitive advantage, or risk falling behind.

However, investing in AI dev tools is just scratching the surface. The real game is to train your workforce to use it effectively, establish policies, regulations, and measures to ensure that AI is not abused within the organization, and not introduce critical vulnerabilities into your digital infrastructure. If not utilized properly, AI can be counterproductive and impede software development. So, it is important to quantify AI’s impact on developer productivity and gain a better understanding of how it impacts developer experience. For this, you can make use of AI-led engineering analytics tools, such as Hatica. Deliver Joyful developer experience with AI-driven insights, and actionable data in your arsenal.