Building for Devs: Lessons from the Trenches of Developer-Focused Products
Spoiler: It’s not just AI hype. There’s real, measurable change happening. Here's what you need to know. Also learn about our upcoming AI Product Management Cohort 2.
This week, we’re zooming into the world of developer-centric products — with a deep dive into how to design, scale, and retain users in this unique category. Our guest is Rishab Jolly, product leader at Microsoft.
Rishab leads product strategy for Azure’s Application Insights—one of the core observability tools used by global enterprises. With over a decade of experience in cloud computing and product management, Rishab is passionate about building products that scale and solve complex technical problems with simplicity. He is also the host of Curious Souls, a podcast where he dives into conversations with leaders across tech and business to uncover their journeys, ideas, and lessons learned.
This episode special - From talking about product-led growth in dev tools, to how AI is reimagining observability, this conversation is packed with takeaways for product builders, founders, and Dev Tool enthusiasts alike.
Let’s get into it.
Brought to you by: Hari Shankar Singhania School of Business.
Website - https://hsb.edu.in/
HSB Admission Link - https://admissions.hsb.edu.in/
🧪 What Makes Developer Products Different?
“Developers hate fluff. They don’t care about marketing spin. They care about tools that just work and fit into their existing workflows.”
This is the first principle Rishab lays out — and it's foundational. If you're building for developers, treat them like power users who value control, speed, and utility.
A few key distinctions Rishab draws between traditional B2B/B2C products and devtools:
The buyer ≠ the user: In an enterprise setting, the CTO might sign the check, but the developer is the daily user. You need to win both.
Meet them where they are: Developers spend most of their time in their IDEs, terminals, or repos. Tools like GitHub Copilot succeed because they integrate natively into the dev environment.
Minimize onboarding friction: Unlike consumer apps, devtools often require SDKs, configuration, and even code changes. A smoother onboarding experience is a huge win.
“We heard developers loved our product, but onboarding was painful. So we shipped auto-instrumentation — a one-click setup that eliminated the need to touch code. It changed everything.”
ANNOUNCEMENT
WIP India and SparkProd are launching their June cohort of AI PM Mastery Course. The last cohort saw Product Managers, Leaders & Founders from across the globe joining and investing in their up-skilling to become AI native product leaders.
Join us for this curtain raiser session on 7th June which will act as your launchpad into how AI is fundamentally reshaping product management. Click the button to learn what our learners had to say about the last cohort and Register for the session.
🚀 Product-Led Growth for Devtools
It’s tempting to assume devtools don’t follow the same PLG playbook — but Rishab disagrees.
“You still need a time-to-value under 4 minutes. Developers will walk away if they don’t see the ‘aha moment’ quickly.”
His framework for successful PLG in developer products:
Discovery: Get embedded where devs already work (e.g. IDE extensions, CLI tools, GitHub integrations).
Onboarding: Offer zero-code or minimal-config onboarding. Think guided flows, auto-setup, sandbox environments.
Instant Value: Deliver meaningful results within minutes. For Application Insights, that meant showing live metrics and tracing errors right after setup.
Daily Habit: Make the tool muscle memory. One team had monitoring dashboards on wall-mounted TVs—that’s deep product stickiness.
“Monitoring isn’t their core job. Coding is. But if your product makes the secondary job seamless, they’ll stick around.”
⚙️ The Role of AI in Dev Products
AI is changing the game in devtools — especially in observability and troubleshooting.
First, a primer:
Observability is the ability to monitor and understand the internal state of a system by tracking outputs like logs, metrics, and traces. It's essential for maintaining performance, especially in high-stakes environments like e-commerce or banking.
“Imagine you’re a dev at Amazon. It’s Black Friday. At 2am, your app crashes. Every minute equals lost millions. You get paged. Your only question: ‘What’s wrong and how do I fix it—fast?’ That’s where observability comes in.”
With AI in the mix, the value compounds:
Signal extraction: AI can sift through mountains of logs and telemetry to surface likely root causes.
NLP-based troubleshooting: Tools like New Relic are experimenting with natural language interfaces. “Why did I get this alert?” can return an actionable diagnosis.
Predictive insights: AI can suggest infra changes before a crash, optimizing code and architecture in real time.
Goal-based agents: In the future, agents might auto-configure entire systems based on your intent: “Set up a compliant banking app infra.” And boom, you’ve got queues, databases, alerts—all scaffolded.
But even here, balance matters.
“It’s not autopilot—it’s co-pilot. Developers still want control. They’re okay with AI doing the repetitive stuff, but they want the final say.”
🧠 TL;DR – Rishab’s Advice for PMs & Founders Building for Developers
Learn their language. If you don’t understand dev workflows, your product will miss the mark.
Build where they work. Don’t ask developers to change behavior—adapt to them.
Make onboarding invisible. Zero-config > zero-trust.
Nail time-to-value. Under 4 minutes is the bar.
Use AI as augmentation, not automation. Co-pilot, not auto-pilot.
Design for daily use, not occasional value. Stickiness beats novelty.
1. Docs Are Out. Contextual Coding Help Is In.
In the old world, devs Googled error messages, combed through Stack Overflow, or scanned endless documentation. Today? Tools like GitHub Copilot are flipping the script.
When a developer hits an error, they can now ask Copilot within their IDE:
“Why did I get this error?”
Copilot doesn’t just return generic advice — it pulls context from the actual project, programming language, and intent. It gives developers personalized, in-the-moment help based on their actual codebase.
🧠 Insight: Documentation is becoming a background knowledge base, not the starting point. IDEs are now the frontline of support.
2. AI Can Make Coding Fun Again
Fixing bugs in someone else’s code? Not a developer’s idea of a good time. But AI is helping take away the repetitive, frustrating parts of the job — making space for creativity.
GitHub reported a 35% boost in developer efficiency for those using Copilot. That’s a real impact.
“Help developers do what they love — create new things. Let AI handle the boring parts.”
3. Reliability Comes from Transparency + Feedback Loops
Yes, AI-generated code isn’t perfect — and that’s okay. What matters is:
Transparency: Why was this code recommended? On what basis?
Feedback loops: Let developers rate or dismiss suggestions to help models improve.
Without these two, trust breaks down. But when they're in place, developers can control and shape the AI experience — a critical part of adoption.
4. PLG Metrics for Dev Tools: What Really Matters
Building for developers? Your PLG metrics will look different — but the fundamentals stay the same.
Here are three core metrics Rishab uses:
Time to First Value: How fast can a dev get to their “aha!” moment? For some tools, this could be <1 minute.
(Example: Microsoft’s Live Metrics shows telemetry data within seconds of deployment.)Expansion Signals: Are users moving from free to paid tiers? Are they inviting teammates?
Activation Rates: Are users actually completing onboarding and starting to build?
💡 Amazon calls these “controllable input metrics” — internal levers you can move to drive external success like NPS or revenue.
5. B2C Inspiration in Dev Tools? Absolutely.
It might sound strange, but dev tools are starting to borrow from consumer products. Rishab pointed out how AWS’s new Q Developer feature mimics the Instagram feed — except instead of photos, it shows updates and alerts in your cloud monitoring environment.
“At work, they’re developers. After work, they’re Instagram users. Familiar patterns work across contexts.”
6. What Product Managers (and Founders) Must Master
Rishab had two sharp pieces of advice for anyone building dev-focused products:
✅ Learn the Language of Developers
You don’t need to be a senior engineer, but you do need to understand what devs do all day — their tools, workflows, and pain points. If you don’t, you’ll talk past them.
✅ Design for Their Journey, Not Your Feature Set
Don’t start with your product. Start with their day:
What does it look like from start to finish?
Where are the bottlenecks or frustrating moments?
Can your product slip naturally into their flow?
7. One Final Word: Be Thoughtful About AI
There’s pressure to bolt on AI to everything — but Rishab issues a clear warning:
“If AI doesn’t align with your product’s core value, it’s a gimmick. You’ll get applause on launch, but poor retention will follow.”
Adoption and retention require AI that genuinely solves a problem. Product leaders must embrace AI with intention, not just trend-chasing.
Final Takeaway:
Developer-first products are changing fast — but product thinking still wins.
Understand your user. Fit into their workflow. Use AI as a tool, not a headline.
And remember: devs are humans, too. The patterns they love in consumer apps? They want those in their work tools, too.
🎥 Watch the full conversation on YouTube → Will AI Replace Developers? with Rishab Jolly | Product Talk with Malthi
🎙️ Listen to the full episode on Spotify -
📥 Forward this to a teammate who’s building dev tools.
🛎️ Subscribe for more deep conversations like this, straight to your inbox.
Catch you next week,
— Product Talk with Malthi