3 Big Business Problems Engineering Leads Can Solve With AI
In general, there’s a big divide in how engineering leaders are approaching AI adoption.
Some teams are investing heavily in AI, leveraging it as a competitive advantage and reaping its rewards.
This report from mid-2023 showed that software development teams adopting AI have already seen a 2.5x speed increase in their software development lifecycle.
In other words, assuming an engineer works 8 hours daily, they reclaim 3 hours by using AI.
These teams expect a further speed increase of 3.5x in the next year. That’s 5 hours back, every day, per engineer.
There's a noticeable split in the approach of engineering leaders towards AI adoption.
Some teams are investing substantially in AI, using it as a strategic edge and enjoying the benefits. Others are choosing to avoid AI. According to this mid-2023 report, software development teams that have embraced AI are already experiencing a significant boost in efficiency, with a 2.5 times increase in their software development lifecycle's speed.
If an engineer typically works an 8-hour day, incorporating AI gives them back 3 hours daily. These teams are optimistic about the future too, anticipating a further acceleration in speed, up to 3.5 times within the next year. This could mean an additional 5 hours saved each day for every engineer.
AI that operates at this level is young, but for some crucial areas, it’s been mature enough to turn substantial results for some time. In particular, I think there are three areas AI is mature enough to help us make substantial gains within our engineering departments.
1. Stagnant engineering delivery
2023 put great pressure on engineering leaders to do more with less.
Devs complete many tasks that have always been considered necessary time sinks, but today AI means that the typical time spent on these tasks before is now wasted time.
I’m talking about tasks like code reviews, writing documentation, writing test suites.
When I speak to engineering leaders, they often worry that these kinds of tasks are not inherently value-generating, not to mention arduous for devs. And that’s usually completely true. But until recently, they’ve required exclusively human effort.
Today, AI creates shortcuts and augmentations for various tasks which form parts of the software development lifecycle.
For many teams, this has already begun with GitHub Copilot. But I already consider Copilot somewhat basic and already, to an extent, an outdated application of AI for the SDLC.
I think the most impactful applications today, for engineering productivity, include these:
- Writing and testing code. AI tools can now do much more than Copilot can. Increasingly sophisticated AI can write entire files, write smart test suites, make cross-file changes, help devs understand their codebase, translate languages and much more besides.
- Writing documentation. Code- and context-aware AI can write comprehensive docs and keep your docs up-to-date without having to take many annoying hours out of devs’ weeks.
Try Mintlify Writer.
- Code reviews. One of the most annoying bottlenecks for devs can now be streamlined with AI that can auto-approve clear-cut cases, make automated suggestions, write good pull request summaries and more.
2. We mimic alignment, but nobody’s actually aligned
Alignment comes in all shapes and sizes. We use the word to mean all kinds of things.
Teams struggle to get internally aligned on a day-by-day basis. Engineers struggle to align with PMs and product teams. Teams struggle to get aligned with business objectives.
Humans aren’t really wired for real-time alignment. Without AI, over decades, we’ve built systems – or perhaps better, coping mechanisms – to help us mimic real alignment.
We use synchronous methods like meetings, asynchronous methods like emails and Slack, and systems of record like GitHub and Jira.
But these coping mechanisms generally leave us out-of-date, with incomplete information, without context and without real visibility. And therefore, without alignment.
Visibility is the sibling of alignment – when we lack visibility into the progress and challenges of our projects, the changing needs of our customers, or our business goals, alignment isn’t really possible.
AI can change a lot when it comes to visibility and alignment.
The biggest (and pretty universal) point of failure with the systems we use to mimic alignment (our coping mechanisms) is that they lack context, commentary and concision.
To put it another way, we want a system that tells us exactly what we need to know, as fast as possible. Data points alone don’t do that. Neither do graphs, nor do updates from our team.
AI makes this possible, when before, we could only mimic it.
This is exactly what Stepsize AI does.
Stepsize AI monitors activities in your Jira software. It uses this to craft concise, actionable weekly reports brimming with the ideal blend of context and detail.
Every week, these reports highlight key metrics through easy-to-scan graphs and insights, enriched with crucial context and commentary.
The tool gives you:
- Metrics with automatic commentary. Metrics without commentary are meaningless. The AI does your data storytelling for you.
- Progress at a glance. Understand the flow of progress with rich, actionable graphs and insights, with sources, so you can investigate if you need to.
- Project-level AI insights. The AI uses your epic structures and its own intelligent classifications to identify themes among all your loose tasks.
- Security first. You’re in control of your data, and it’s never used to train AI.
All this means you effectively have instant insights, with zero-setup and complete visibility so you can get aligned on what matters.
3. Shaping data for decision-making
Getting the data needed to make the right decisions, into the right format, is hard work. And the tools we use to do it come with non-monetary costs and limitations that often go unnoticed until they become pain points.
We use tools to herd our data into the right shape, such as native analytics (like Jira’s) and BI tools (like Tableau).
But these tools are resource-heavy endeavours. They demand significant setup, and significant maintenance.
BI dashboards, even the most sophisticated ones, inherently lack what I call 'data storytelling'. They fall short in providing the necessary context and commentary that brings data to life. They might indicate improvements in velocity or flag a stalling project, but they don't provide the 'why' or the 'what's next.'
This lack of context diminishes the value in decision-making. And when the context is added, it becomes a laborious, time-consuming task. In fact, the total cost of report generation can ADD LINK add up to over $40,000 per year per report. This is at odds with the principles of Agile, which emphasise efficiency and value.
How do we make this investment count?
Again, Stepsize AI is the ideal artificial intelligence tool to help with this problem. As we’ve said, Stepsize AI develops a deep understanding of your business, team and project context, and selectively surfaces the metrics that actually matter in easy-to-read formats, with powerful AI-generated commentary.
And since set-up is a 2-minute job – the AI does the work – there’s no need to bring in data teams or consultancies to set up and maintain streams of data.