How Two Engineers Ship Like a Team of 15 With AI Agents

TL;DR: Today we’re releasing a new episode of our podcast AI & I. Dan Shipper goes in depth with Kieran Klaassen and Nityesh Agarwal, the engineering team building Cora, Every’s AI-powered email assistant. Watch onYouTube or listen on Spotify or Apple Podcasts. Here’s a link to the episode transcript.

Was this newsletter forwarded to you? Sign up to get it in your inbox.

If you’re using AI to just write code, you’re missing out.

Two engineers at Every shipped six features, five bug fixes, and three infrastructure updates in one week—and they did it by designing workflows with AI agents, where each task makes the next one easier, faster, and more reliable.

In this episode of AI & I, Dan Shipper interviewed the pair—Kieran Klaassen, general manager of Cora, our inbox management tool, and Cora engineer Nityesh Agarwal—about how they’re compounding their engineering with AI. They walk Dan through their workflow in Anthropic’s agentic coding tool, Claude Code, and the mental models they’ve developed for making AI agents truly useful. Kieran, our resident AI-agent aficionado, also ranked all the AI coding assistants he’s used. You can check out their full conversation here:

If you want a quick summary, here are some of the themes they touch on:

The workflow you can use to 10x your engineering capabilities

At the heart of Kieran and Nityesh’s workflow with AI is a meta move: They built a prompt that writes prompts. With help from Anthropic’s Prompt Improver, they created a custom command in Claude Code that transforms a rough feature idea into a fleshed-out GitHub issue. Each issue includes a clear explanation of the problem, a proposed solution, the technical details needed to make it happen, and a step-by-step implementation plan. The agent pulls in relevant parts of the existing code and best practices from the web to help guide the approach.

Once the GitHub issue is created, they review it themselves and queue it up to be implemented. By planning the work, writing out the issue, and then reviewing it, they create space to think clearly about the problem before any code is written. Unlike AI code editor Cursor, which is “made to code,” Kieran says Claude Code reduces the friction to think things through before jumping into execution.

Even better, Kieran and Nityesh say they rarely bother with typing. They speak to their computers using an internal voice-to-text tool called Monologue, which is currently being incubated here at Every.

How to write reliable code with AI agents

For Nityesh, an important mental model for coding with AI comes from a decades-old management book, High Output Management by Andy Grove, Intel’s former CEO. The principle is simple: In any production process, fix problems while the stakes are still low. In the context of Claude Code, that means putting extra care into the earliest parts of the workflow, before the agent actually starts coding.

It’s tempting to immediately ask Claude to implement a GitHub issue it just wrote, but that can backfire. Nityesh says, “There are chances that [Claude’s] plan… wasn't the direction that you wanted to go, and you want to catch that before you ask Claude to go and implement the solution.”

Dan admits he tends to rush through the planning phase because reading a wall of text generated by an LLM can be boring. The trick, Kieran says, is to shape the plan so it's both informative and engaging. One way to do this is prompting Claude to ask them questions like a good product manager would. Dan notes a similar approach in a new, yet-unreleased version of Spiral, Every’s content generation tool. General manager Danny Aziz is building a writing agent that interviews you to help you articulate your thoughts better.

Kieran rates coding agents

Kieran has a reputation at Every for testing every coding agent that hits the market, and Dan got him to rank them. Kieran's clear top choice is Claude Code because it goes beyond just coding to handle research, workflows, and everything else. Amp comes in second; Kieran praises its “ergonomics” and notes that it feels like a tool built by people who actually use it. Friday rounds out his top tier because of its opinionated workflows that just work. Even though it doesn’t use Claude 4 yet, he's seen it pull off things he couldn't replicate in Claude Code.

And on the other end of the scale: Windsurf represents the brutal pace of this space—just weeks ago it would've been high up on the list, but without Claude 4, it's quickly fallen behind. GitHub Copilot sits at the bottom of Kieran’s list, though he admits this might be unfair since he hasn’t tested its newer agentic capabilities.

Here’s a link to the episode transcript.

You can check out the episode on Spotify or YouTube. Links are below:

1. Watch on YouTube
2. Listen on Spotify (make sure to follow to help us rank!)

What do you use AI for? Have you found any interesting or surprising use cases? We want to hear from you—and we might even interview you.

Miss an episode? Catch up on Dan’s recent conversations with founding executive editor of Wired Kevin Kelly, star podcaster Dwarkesh Patel, LinkedIn cofounder Reid Hoffman, former a16z Podcast host Steph Smith, economist Tyler Cowen, writer and entrepreneur David Perell, founder and newsletter operator Ben Tossell, and others, and learn how they use AI to think, create, and relate.

If you’re enjoying the podcast, here are a few things I recommend:

1. Subscribe to Every
2. Follow Dan on X
3. Subscribe to Every’s YouTube channel

Rhea Purohit is a contributing writer for Every focused on research-driven storytelling in tech. You can follow her on X at @RheaPurohit1 and on LinkedIn, and Every on X at @every and on LinkedIn.

We build AI tools for readers like you. Automate repeat writing with Spiral. Organize files automatically with Sparkle. Deliver yourself from email with Cora.

We also do AI training, adoption, and innovation for companies. Work with us to bring AI into your organization.

Get paid for sharing Every with your friends. Join our referral program.