How to Use AI to Be a Great Engineering Leader
10 engineering leaders are sharing their real-world cases on how they are using AI to help them be better engineering leaders!
DevStats (sponsored)
This newsletter is sponsored by DevStats. It’s an easy-to-use dev analytics tool built to help startups and mid-sized teams unlock insights, streamline workflows, and deliver faster.
I’ve checked DevStats thoroughly and I used it in one of my projects as well, so I highly recommend checking them out if you are looking for such a tool.
Here is my personal recommendation:
“DevStats is one of these tools that you think you don’t need, but it’s just so easy to use and gives so many insightful details that you might be missing.
Before using it, my thoughts were that we were having great PR Cycle times as well as Code Review sizes, but with helpful filters and visuals, I could clearly see where we can improve.”
Thanks to DevStats for being a sponsor. Let’s get back to this week’s thought.
Intro
There was a lot of great feedback for the article: How to use AI to increase Software Development productivity. Special thanks to all of the 11 engineering leaders who shared their insights with us in that article.
If you haven’t read the article yet, it’s a must-read for everyone interested in using AI to increase Software Development productivity:
Now, we are doing a similar edition. This time, 10 engineering leaders are sharing their real-world cases on how AI is helping them to be better at their roles.
These engineering leaders are from all different sizes of companies -> from startups, mid-size companies and all the way to Big Tech.
Their roles span from Tech Leads, Staff Engineers to Engineering Managers, Directors, VPs, and CTOs.
Thanks to all of the engineering leaders who have shared their insights:
Ronald van Velzen, ML Tech Lead at Eneco,
Owain Lewis, Director, Software Engineering at Oracle,
Tarik Guney, Principal Engineering Manager at Microsoft,
Miroslaw Stanek, PL Site Leader & Engineering Director at Papaya Global,
Dr Milan Milanović, CTO at Trucking Hub,
Rafa Paez, Senior Engineering Manager,
Tanmoy Bhattacharjee, Senior Engineering Manager at PropertyGuru,
Shivam Anand, Staff ML Engineer at Meta,
Balaji Narayanan, VP of Engineering at Clari,
Joe Disharoon, CTO at OneSource Virtual.
Let’s start with the first insight!
1. Using AI as a thought partner
This was shared by Ronald van Velzen, ML Tech Lead at Eneco. He leads a Generative AI team at a large sustainable energy company in the Netherlands. With a background in mathematics and Data Science, he worked in AI for over six years, focusing on Natural Language Processing.
Before the recent AI wave, he was already building machine translation systems and speech-to-text models. So seeing the current explosion in possibilities feels both exciting and familiar for him.
His team builds AI enablers that make business processes more efficient, especially in customer experience. One key project involves summarizing millions of customer conversations to extract actionable insights. There's an interesting recursion here: they build AI products while using AI to build them.
In his day-to-day work, AI has become an indispensable tool across many areas for him. Since their projects are highly experimental, fast iteration is crucial. GenAI helps by:
Quickly generating initial code for MVPs or analysis scripts, giving a strong starting point to improve upon.
Streamlining documentation with draft READMEs.
Making test writing more efficient with generated test cases that can be expanded.
Beyond coding, he found a lot of value in using GenAI as a thought partner. He regularly uses tools like Claude, ChatGPT, and Perplexity to brainstorm presentations and other communication or to explore alternative approaches to technical challenges.
This often reveals perspectives he might have overlooked. Perhaps most valuable is GenAI's role as an on-demand knowledge base. When encountering unfamiliar technologies or concepts, he can quickly get concise explanations that help him make informed decisions without falling down in too many rabbit holes.
The result has been a more agile, informed leadership style. Exactly what he needs when working at the cutting edge of generative AI applications in enterprise settings.
2. Using AI every single day for writing documents
Shared by Owain Lewis, Director, Software Engineering at Oracle. He uses AI every single day in his role as a manager. While developers have obvious use cases like writing code, managers can benefit just as much.
Writing is a huge part of his job and AI helps him to write executive documents, technical or product one-pagers, emails, team documentation and performance summaries (turning data into a structured template).
An AI tool he’s been using for years is Grammarly. Claude is the best model overall for writing.
Quick tip from him: Make templates for your common docs and let AI help fill in first drafts.
Even though he’s not coding daily, he uses Cursor for development on side projects. It's super important for managers to keep up their technical skills - AI makes this way easier than before. Plus, you need to know this stuff to help your engineers who might not realize how much AI can boost their productivity.
Big companies have restrictions on how you use AI. Security is a major headache. You don't want sensitive information leaking through public APIs. If you work in a large enterprise company, you’ll likely be highly restricted on the AI tools you’re allowed to use in your work.
AI costs money and companies are still figuring out the budget side. There's surprising resistance from some developers he speaks to. He personally think using AI is non-negotiable for modern developers. You can write code without it, but... why would you?
His personally interested in how to use AI to improve day-to-day processes including operations. The potential here is huge.
He also mentioned that AI Hack Days work really well. Give your team time to play with AI tools, build cool stuff and then share what they have learned. It's a great way for everyone to learn together and get excited about the possibilities.
Once you start using AI in your day-to-day work, it’s hard to imagine how you’d operate without it.
3. 6 areas where AI helps to be more effective
This is from Tarik Guney, Principal Engineering Manager at Microsoft and previously an engineering leader at Adobe and Motorola Solutions.
He found out that one of the most pressing challenges in engineering leadership is scaling your time.
It’s not just about managing more people or tasks → it’s about using leverage to extend your impact in meaningful ways: enabling clarity, accelerating collaboration, and creating progress across teams and stakeholders.
In his role, he is constantly balancing strategy, execution, and people development while collaborating with senior leaders.
As Andy Grove said:
Your output as a manager is the output of your organization plus the output of the people you influence.
These are the 6 areas that AI helps him:
Aligning people faster through better communication
For him, AI has become a key partner in this. As a leader, one of his biggest forms of leverage is communication → whether it's setting direction, aligning stakeholders, or unblocking teams.
He uses AI to review and refine his messages, especially as a way to reduce organizational drag and improve alignment across technical and business stakeholders, before critical updates or tough conversations, so that he can ensure his words land well across diverse audiences.
His process → He feeds the drafted message into an AI assistant and ask it to act as a “perspective simulator”, reviewing the message as if it were an engineer, a peer manager, or a VP of Engineering.
This helps him spot assumptions he might be making, clarify technical or strategic intent, and better align with the recipient's mental model.
Recalling the critical context when people can’t
He also uses AI to review meeting summaries and code snippets when he wants a quick understanding without pulling engineers into ad hoc explanations.
Surprisingly often, AI can recall and summarize past decisions or conversations more accurately than people, especially when it's drawing from structured meeting notes or chat logs.
For instance, when revisiting a design review from a month ago, AI was able to surface the original rationale and trade-offs discussed, something even the attendees had partially forgotten.
This allows him to surface the critical context quickly, make better-informed decisions, and keep the team focused on what truly matters.
Preserving trust and unblocking teams
Another core source of leverage for him is the institutional knowledge he carries and the trust-based relationships he’s built across teams.
Trust isn't just built through personal interactions; it's built and reinforced through consistent preparedness over time, which compounds into what he thinks of as “reputation capital”.
Showing up informed and respectful of others’ time consistently is what sustains that trust. This is where AI becomes a surprising ally.
He uses AI to retrieve past context → decisions, dependencies, or past commitments. It allows him to approach those cross-team conversations with clarity and precision.
Instead of asking others to rehash prior agreements, he comes in already aligned on what was said or decided. That shows respect, reduces friction, and strengthens trust.
For example, if an engineer asks, "Did we ever decide which team owns this integration point?" or "What was the context for this design constraint?", he can feed meeting notes into AI and get a reliable summary, avoiding the need to chase down people or restart threads.
Similarly, when a peer program manager or his own manager asks for a quick refresher on how or why something was decided, he can surface the relevant thread or summary immediately.
This strengthens trust not just laterally but upward as well, showing that he’s informed, prepared, and mindful of others’ time.
Architecture and code reviews
AI also plays a role in architectural decision-making. When they are faced with a new problem, he often uses AI to search through internal documentation, Slack/Teams threads, or meeting transcripts to surface how similar problems were solved in the past, especially under similar constraints.
This helps them to avoid reinventing the wheel, make faster decisions, and surface institutional memory during solution design.
Similarly, during code reviews, AI assists him in identifying more appropriate or idiomatic usage of APIs, classes, or methods, often surfacing examples from other parts of our codebase or adjacent repositories.
This is particularly helpful when reviewing unfamiliar modules or cross-team contributions.
Automating daily tasks
He regularly uses AI to help with small but recurring technical tasks, like writing shell scripts, generating one-off commands, escaping/unescaping tricky JSON payloads, or reviewing long stack traces from exception messages to extract key insights quickly and catch details that are easy to miss in regular reviews.
These small tasks often consumed more time and attention than they should have, frequently requiring third-party tools. Over time, AI has become his go-to assistant for these needs, progressively reducing his dependence on external tools.
Distilling signals from noise
AI also helps him to reduce the cost of context switching by distilling signals from noise.
One of the most effective ways it does this is by summarizing long threads, whether in Microsoft Teams channels, email chains, or internal docs, so he can get up to speed quickly.
Instead of spending his mental energy parsing scattered updates or half-followed discussions, he can instantly grasp what decisions were made, what’s still open, and where attention is needed.
This helps him act faster and minimize the need to interrupt others for clarification.
The tools he uses
Communication, writing, and information recall
For the workflows described above, he uses a mix of ChatGPT, Microsoft Copilot, and Claude Sonnet. Since he works at Microsoft, Copilot is his primary tool for work-related tasks, especially when dealing with internal systems and documents.
He also relies heavily on Microsoft Teams, which has a fantastic Copilot integration. It enables many of the things he described above, like summarizing threads, recalling meeting context, and reducing the burden of context switching.
Outside of that, he frequently uses ChatGPT where appropriate, and occasionally turns to Claude Sonnet for more structured, exploratory tasks.
If you're exploring similar setups, be sure to review your company’s policies around data handling to ensure your tool choices align with security and compliance standards.
Coding & developer tooling
he mentioned that Cursor has been particularly helpful for tasks like code generation, refactoring, and navigating unfamiliar codebases.
At work, GitHub Copilot in VS Code is deeply integrated into his workflow and supports him throughout the day. He’s also experimenting with JetBrains Rider’s AI Assistant, which brings context-aware suggestions tailored to .NET development.
For command-line work, He’s been impressed with Warp’s AI-enhanced terminal, which streamlines writing and explaining shell commands.
He also mentioned the following:
He treats AI like a co-pilot for thinking, understanding, and translating across roles. It helps him protect his time and invest it where it matters most: people, decisions, and long-term outcomes.
Used intentionally, AI isn’t just a productivity boost, it’s a leadership multiplier.
Most importantly, it makes him significantly more productive → giving him more space for deep thinking, personal growth, and meaningful time with his family.
That’s the real win: leading better at work without sacrificing what matters most outside of it.
4. AI helps him to accomplish things in hours that would have taken days or weeks
Shared by Miroslaw Stanek, PL Site Leader & Engineering Director at Papaya Global. His work today looks different than it did a few years ago when GenAI was non-existent in the mainstream market.
AI helps him think faster. He accomplishes things within hours that would have taken days or weeks just a few years ago.
Here's the most important case. Many engineering leaders, including him, constantly jump between numerous initiatives. In contrast to individual contributors, where deep focus is critical, managers drive many things simultaneously.
Practically, this means that coding tasks that take days are usually an ineffective allocation of the time for engineering leaders. Most likely, we won't finish them on time and become a bottleneck for the team.
Using AI for writing code
Today, with GenAI, he can write code a few times faster. It means he could return to regular software development. In his dual role as an engineering director and site leader, he can still deliver value even if he spends only a few hours per week on coding.
Being a hands-on mid-level manager is a game changer on many levels: he communicates with developers, his understanding of the tech stack and DevEx, and then he can translate that back to business and product.
In his work, he uses VS Code and Copilot for auto-completion, and ChatGPT, Gemini, or Claude to generate more structured code, usually after a longer conversation.
He is exploring agentic modes, but as an occasional coder, they are still too unreliable.
Because he works with the full tech stack (DevOps, data, backend, frontend, mobile), before he bothers engineers with random questions, he uses AI to explain things for him. A typical recent case? "Explain to me this Terragrunt IaC config" (his background is frontend/mobile).
Using CodeAudits to review the codebase
He also developed his own AI tool CodeAudits.ai, which helps him review the entire codebase. It parses the repository and uses custom prompts to verify architecture, some coding principles, quality, etc.
With the tool, he gets for example, "Top 10 hints and code snippets to refactor this project to Clean Code architecture" or "Critical issues to fix in order to make this codebase follow our internal logging standards."
Under the hood, it uses Google Gemini because of its 1-million context window (most of the repositories he works with didn't fit into ChatGPT or Claude request limits).
Using AI for research and writing
He also uses GenAI for general research and some writing. Like, for example, article summaries or video-call transcripts to bullet points - these are great cases he uses regularly. For example, before he watches a one-hour conference talk, he checks with Gemini "What are the key highlights of this YouTube video?"
However, writing an email, an article, or bringing some empirical data → the results are quite disappointing for him. The content he receives, even though nicely formatted, is wordy, full of fluff, and contains data that is difficult to validate.
Even "Deep Research" from ChatGPT or Perplexity is far from perfect. Sure, you get references, but in many cases, those are websites that don't disclose where their data comes from.
What works for him here is bringing his data in and then refining the output with AI. For example, he puts DORA or any other research paper into NotebookLM to get what he needs.
Or he writes content on his own and makes some adjustments and grammar fixes with Grammarly. When he needs to conclude from data sets (e.g., CSV files, logs, JIRA or Slack API), he still prefers asking ChatGPT for Python Notebook code to do the analysis himself instead of asking AI to interpret the data.
Using MCP to get overviews or suggestions
The light at the end of the tunnel that he’s been recently researching with his team is MCP (Model Context Protocol). He uses it already in a limited way.
For example, Claude tells him what happened in the last 48 hours of the company's Github Pull Requests. Or, the Copilot agent suggests changes to the entire codebase based on their internal coding standards. On his list to explore are integrations with Figma, some design systems, OpenAPI specs, and IDP platforms, to name a few.
In some cases, he can see some limited wins ("write API client for service X based on its API docs"); in others, results are still far from what's promised.
Working with product teams
When it comes to working with product teams, here's an interesting case that works nicely.
Let PMs prototype their apps with v0, Bolt, Lovable, or other "FullStack AI."
While integrating generated apps with a company's infrastructure is a daunting task in most cases, they can use the same prompts that Product Managers wrote in Cursor or VS Code.
The results of such a process are surprisingly good, so this is the direction he will keep exploring.
5. Using AI to streamline communication, improve team operations and make data-driven decisions
Shared by Dr Milan Milanović, CTO at Trucking Hub. As a CTO leading teams on complex projects, he integrated AI tools into his daily workflow in three key areas:
1. Streamlining communication
Writing consumes a significant portion of his day, from technical documentation to meeting notes to feedback for team members. Since English isn't his first language, he regularly uses ChatGPT to polish his writing before sharing it.
An example of a prompt he uses: "Polish for clarity and a direct, friendly tone. Keep technical terms intact. Reply with the improved text and a 3-bullet executive summary."
When drafting important team announcements, he'll ask AI to "check for grammar mistakes and suggest a more concise version." For technical specifications, he requests "simplify this explanation for non-technical stakeholders" to bridge understanding gaps between teams.
This saves him time and ensures his communication lands correctly with different audiences.
He also built his custom GPT for clear communication, which he uses daily.
2. Improving team operations
When joining a new project recently, he needed to understand a large, unfamiliar codebase quickly. Instead of spending days piecing together the architecture, he pasted key files into an AI tool and asked for an explanation of component relationships. This gave him a clear mental map in hours rather than days.
Based on this, he even built a small tool to review codebases that he uses nowadays. And also, they have integrated an AI tool for code reviews as well.
For his 1:1s with engineers, he created a simple system where he feeds AI each team member's current projects and challenges, then ask it to generate personalized discussion points. This ensures his conversations are productive and tailored to their specific needs.
So, before each 1-on-1, he prompts: "Since 2025-03-01, list Alex's shipped tickets, top blocker, and one win."
AI also transforms their hiring process. Before interviewing candidates for a specialized position outside of his expertise, AI analyzes the job requirements and generates technical questions with sample answers. This helps him evaluate candidates more effectively while providing a consistent interview experience.
Also, he uses AI assistants to streamline communication by summarizing meetings, generating documentation, and even answering routine questions via chatbots → saving hours each week for him and his team.
3. Making data-driven decisions
The most impactful use case has been extracting insights from their project data. He created scripts using AI assistance that pull data from Jira and GitHub, normalizing it to calculate their DORA metrics automatically. What used to be a manual weekly process taking hours now runs in minutes, giving him constant visibility into team performance.
He schedules it with a cron job and pipes the summary straight to their "#eng-metrics" channel.
When planning their latest roadmap, he fed their previous quarter's data into an AI tool and asked it to identify bottlenecks and suggest process improvements. The analysis revealed patterns he hadn't noticed - their QA cycle had specific friction points that, once addressed, improved their delivery time by 15%.
So, by relying on objective, data-driven evaluations, he reduces bias and it helps with greater transparency + trust within his teams.
6. AI has become a second brain in his day‑to‑day work
Rafa Paez, Senior Engineering Manager, has shared that Generative‑AI tools and LLMs such as Gemini, Notion AI, Slack AI, Cursor, and their internal chatbot don’t just save time; they help him lead more effectively.
Here’s how he puts AI to work:
Summarizing conversations & threads
With AI now built into many of the tools we use daily, he can quickly digest long email chains, chat threads, and docs, pulling out the important points without wading through everything.
Example: He was recently a part of a long email thread and needed to respond. AI gave him a quick summary, saving nearly an hour of reading.
Drafting content faster
Whether it’s a strategy doc, performance review, or team update, AI gets him from blank page to solid first draft fast. Then he just fine-tunes the message.
Example: For an engineering proposal, he used AI to brainstorm and outline ideas. Then he rewrote the final version in his own words, asking AI to help polish the last text in English.
Coaching and feedback
AI pulls together feedback, wins, review notes, and 1:1 summaries so he can walk into conversations with clearer insights and better context.
Example: During performance reviews, he gathered highlights from the past 1:1s, weekly updates, and feedback into one doc. The self-reflection took a fraction of the time it used to.
Navigating unfamiliar code
As a manager of managers, he codes little, but AI can explain snippets, flag risks, and help him review unfamiliar code with confidence.
Example: He had to approve a merge request in a codebase he wasn’t familiar with. AI helped him to understand what was going on and spot potential issues before giving the green light.
Meeting support & note-taking
AI-generated summaries and actions help him stay on top of meetings, even the ones he misses. For 1:1s, though, he turns note-taking off to keep the space private and safe.
Example: He missed an important meeting recently but could catch up through the summary and insights; no need to watch the full recording.
Searching internal knowledge
When you have thousands of documents and conversations, finding anything can be a pain. AI makes internal knowledge searchable and easy to surface.
Example: He needed something specific from months ago. Instead of searching manually, he asked AI and had the answer in seconds.
Decision support
He uses AI a lot as a thought partner. AI helps him stress-test ideas, weigh trade-offs, and navigate tough decisions before bringing them to the team.
Example: He was stuck between two conflicting priorities and asked AI for a second opinion and ended up with a better solution than the one I’d initially considered.
Offloading routine tasks to AI frees up hours each week, letting him coach more, align better across functions, and focus on high‑leverage work that boosts delivery, morale, and overall team performance.
7. Using AI to help with a number of different cases
This is what Tanmoy Bhattacharjee, Senior Engineering Manager at PropertyGuru has shared. He works at the biggest e-commerce real estate portal in Southeast Asia and this is how AI helps him:
Drafting, summarizing and overall writing
MS Copilot and ChatGPT helps him with drafting emails and overall writing. Both also help with summarizing email threads, and keeping up with Teams and Slack.
Confluence AI acts as his digital assistant for searching through docs.
Perplexity Deep Research helps him with preparing for meetings and overall goal setting.
Coding
GitHub Copilot autocompletes code, explains functions, and also helps with tests. He also uses Claude Code as a “pair programmer”.
Stakeholder Management
Perplexity Pro and Confluence AI helps him to answer stakeholder questions by searching, summarizing, and analyzing files and documentation.
Gemini 2.5 Pro and Claude 3.7 → They are his go-to for deep dives, advanced reasoning, and explaining why “this should be done in this specific way”.
Project Management
Notion AI, and Loop AI helps him as “digital whiteboards” and project trackers, ensuring nothing gets lost and that everyone is on the same page.
8. Using AI for document creation to enhance upward visibility
Shared by Shivam Anand, Staff ML Engineer at Meta. He had the most success in using AI to iterate on docs with upward visibility.
He comes up with an initial draft for proposals on his own, but asks AI to help summarize it. He prompts it to make it shorter while still keeping the content accurate. He is primarily looking for feedback on style and grammar.
Writing documents is about 50% of his job. he estimates it makes this task 10% more efficient. In total, that's a 5% productivity boost.
He uses Meta's internally hosted version of LLAMA.
9. Using ChatGPT as a Swiss army knife
This is what Balaji Narayanan, VP of Engineering at Clari, have shared. He doesn’t need to code daily but he wants to be close to the system.
His primary AI tool has been ChatGPT as a Swiss army knife for all things:
researching a topic,
summarizing notes,
reviewing communication - especially check for understanding on the executive level.
He spent a few hours with ChatGPT to build a complete Android app for a Geography quiz to understand vibe coding better.
For code, he has been playing with Factory.ai with reasonable success. It has helped him to understand codebases better and suggest changes.
10. Connecting the knowledge base to AI helps with report generation, among other things
Shared by Joe Disharoon, CTO at OneSource Virtual. They have connected their knowledge base to AI (Claude, Grok, Copilot in order of usefulness) to be able to discuss new features and requirements with AI.
This helps him and everyone else to design feature requests, reports, UI screens, and occasionally design their own AI agent bots in the Copilot Studio.
He mentioned that in their organization, they also have a dedicated enterprise architecture team to evaluate new tools and every new release of the major models they can use, and to make sure they are evolving their best practices.
Last words
Special thanks to all the engineering leaders for sharing their insights on how they are using AI to make their day-to-day better and more effective.
Hope you enjoyed this edition where I bring insights from various people and share their insights. Let me know in the comments.
I’ve personally learnt a lot from all the insights shared and had a lot of fun writing the article!
We are not over yet!
7 Common Mistakes When Growing From Engineer to a Lead Role
Check out my latest video. I am sharing common mistakes engineers make when they grow to their first lead role.
New video every Sunday. Subscribe to not miss it here:
Liked this article? Make sure to 💙 click the like button.
Feedback or addition? Make sure to 💬 comment.
Know someone that would find this helpful? Make sure to 🔁 share this post.
Whenever you are ready, here is how I can help you further
Join the Cohort course Senior Engineer to Lead: Grow and thrive in the role here.
Interested in sponsoring this newsletter? Check the sponsorship options here.
Take a look at the cool swag in the Engineering Leadership Store here.
Want to work with me? You can see all the options here.
Get in touch
You can find me on LinkedIn, X, YouTube, Bluesky, Instagram or Threads.
If you wish to make a request on particular topic you would like to read, you can send me an email to info@gregorojstersek.com.
This newsletter is funded by paid subscriptions from readers like yourself.
If you aren’t already, consider becoming a paid subscriber to receive the full experience!
You are more than welcome to find whatever interests you here and try it out in your particular case. Let me know how it went! Topics are normally about all things engineering related, leadership, management, developing scalable products, building teams etc.
Really excellent piece. Love the multiple perspectives and usecases. Already doing a lot of this but saw some exciting possibilities for refining and iterating. Thank you all
Thanks for including my perspective, Gregor! It’s an honor to be featured alongside so many thoughtful engineering leaders. I really appreciate the work you're doing to highlight how AI is reshaping leadership. Would love to collaborate on something in the future!