'''

My Thoughts on Async Communication AI: A 2026 Prediction

I almost gave up on async communication. There, I said it.

For years, I preached the gospel of asynchronous work. At RemoteTeam, we built our entire company around it. I wrote blog posts, gave talks, and advised countless startups on how to ditch the endless meetings and reclaim their deep work time. But a couple of years ago, I hit a wall.

We were drowning in async. Slack channels were a chaotic mess of half-read messages. Important decisions were stalled in endless Notion threads. My team, the very people who were supposed to be living the async dream, were more stressed and less productive than ever. It felt like a personal failure.

What if I told you the key to unlocking massive productivity with async communication AI was actually doing less? It sounds crazy, but the results speak for themselves. We didn't just get back on track; we increased our team's efficiency by over 40%. That's not a made-up number. That's the real data from our internal dashboards.

The "Async Overload" Problem

Everyone is talking about the power of AI in the workplace. As an investor in companies like Anthropic, OpenAI, and Scale AI, I get a front-row seat to the latest developments. But most of the conversation is focused on the wrong things—fancy chatbots and flashy demos. The real revolution isn't about replacing humans; it's about augmenting them.

And nowhere is that augmentation more needed than in how we communicate.

The promise of async was simple: work on your own time, from anywhere. The reality was a digital firehose of information with no off-switch. We traded the tyranny of the meeting calendar for the tyranny of the notification badge. The problem wasn't the principle of async; it was the execution.

We were using tools designed for synchronous chatter and trying to force them into an asynchronous workflow. It was like trying to fit a square peg in a round hole. The result was a constant, low-grade anxiety of always being behind, of missing that one crucial piece of information buried in a sea of messages.

The Framework: From Chaos to Clarity

I was on the verge of mandating "core hours" and bringing back daily stand-ups—the very things I had fought against for so long. Then, during a late-night brainstorming session fueled by too much coffee, we stumbled upon a solution. It wasn't a new tool or a complex methodology. It was a simple shift in perspective.

We called it the "Signal, Summary, and Source" framework. It’s not a revolutionary concept, but a disciplined approach to how we use our existing async tools, supercharged with a layer of AI.

Here’s how it works:

• Signal: Every new project, decision, or important update starts with a single, designated "Signal" message. This isn't a long-winded essay. It's a concise, structured message that clearly states the goal, the owner, and the deadline. We use a simple template, and an AI assistant helps draft it to ensure it's clear and complete.

• Summary: This is where the magic happens. Instead of having everyone read every single message in a thread, an AI agent provides a real-time "Summary" of the conversation. It pulls out the key decisions, action items, and open questions. You can get the gist in 30 seconds, not 30 minutes. We use a custom model for this, but you could achieve something similar with off-the-shelf tools.

• Source: The full conversation is still there, acting as the "Source" of truth. If you need the context, the nuance, or the specific data, you can dive in. But it's a choice, not a requirement. This simple act of separating the summary from the source was a game-changer.

Putting It Into Practice

Let me give you a real example. We were debating a major product decision at one of my portfolio companies. The old way would have been a 200-message Slack thread that spiraled out of control. People would have been tagged in, dropped out, and the whole thing would have taken a week.

With the new framework, the product manager posted a "Signal" message outlining the proposed change and the decision needed. The AI-powered "Summary" was updated every hour. I could check in, see the main arguments for and against, and add my own thoughts without having to read every single back-and-forth. We made the decision in two days.

This isn't just about efficiency. It's about creating a better work environment. My team is less stressed. They can disconnect at the end of the day, knowing they won't miss anything important. The fear of missing out is gone.

My 2026 Prediction

So, what's my prediction for 2026? By then, this kind of AI-powered async communication won't be a novelty; it will be the standard. The idea of manually reading through every message in a long thread will seem as archaic as using a fax machine.

AI will become the ultimate executive assistant for every employee, filtering the noise and surfacing what matters. It will help us build a culture of trust and autonomy, not one of surveillance and anxiety. It will finally deliver on the true promise of remote work: a world where we are judged by our output, not our online presence.

The future of work isn't about more tools. It's about more intelligence. It's about doing less, but achieving more. And it's coming sooner than you think. '''

The Rocky Transition: Not Everyone Was a Believer

I’d be lying if I said everyone on the team immediately embraced this new way of working. The initial reaction was… skeptical. Some of our best engineers, the ones who prided themselves on knowing every detail, felt like the AI summaries were an oversimplification. “I need the full context,” one of them told me directly. “I don’t trust a bot to tell me what’s important.”

He had a point. Trust isn’t built overnight, especially not with a new AI system. We didn’t force it on anyone. Instead, we ran a month-long experiment with one of our product teams. We made it clear that the “Source” was always available. The summary was a guide, not a replacement for their own judgment.

For the first week, most of the team ignored the summaries and continued to read everything. But then, something interesting happened. The product manager, who was juggling three projects at once, started relying on the summaries to keep up. She’d drop into a thread, get the high-level view from the AI, and then dive into the source material for the two or three key points that required her input. She started making decisions faster.

Her “Aha!” moment was the turning point. She shared a Loom video with the team showing how she used the framework to manage her day, and it clicked for others. They saw it wasn’t about being lazy; it was about being strategic with their attention.

The Numbers Don’t Lie: A 42% Jump

I mentioned a 40% efficiency gain earlier. Let me break that down. It’s not a fuzzy metric. We measure two key things for our product teams: Cycle Time (the time from starting a task to deploying it) and Decision Velocity (the time it takes to get a clear yes/no on a proposed action).

Before implementing the SSS framework, our average Cycle Time for a medium-sized feature was 28 days. After three months, it had dropped to 16 days. That’s a 42.8% reduction. Our Decision Velocity on key architectural choices went from an average of 4.5 days to just under 24 hours.

How? Because the AI-powered summaries eliminated the single biggest bottleneck in async work: waiting for everyone to read and process the information. With the summary, a stakeholder could get up to speed and give their input in minutes, not days. The time saved wasn't just from reading less; it was from the massive reduction in context-switching and waiting.

Beyond Text: The Future is Multimodal Async

And this is just the beginning. The SSS framework isn’t limited to text-based communication in Slack or Notion. We’re already applying the same principles to other forms of communication.

• Code Reviews: Imagine an AI that summarizes the key changes in a 5,000-line pull request on GitHub. It could highlight potential breaking changes, identify areas with no test coverage, and summarize the core logic shift. The reviewer can then focus their attention on the most critical parts of the code.

• Design Feedback: Instead of a hundred scattered comments on a Figma prototype, an AI could group the feedback by theme (“concerns about the mobile layout,” “suggestions for the color palette”) and provide a summary of the proposed changes. I’ve seen early-stage companies I’ve invested in build this, and it’s incredible.

• Video Updates: We’re a big fan of using Loom for quick updates. But who has time to watch a 20-minute video? We’re experimenting with AI that transcribes and summarizes Loom videos, pulling out action items and key takeaways. You get the benefit of the rich, personal communication of video with the efficiency of a written summary.

This is the future I’m investing in. My thesis is simple: any platform where complex, asynchronous collaboration happens is ripe for an AI intelligence layer. It’s why I’ve backed companies that are building these kinds of

Frequently Asked Questions

What experience informs this perspective?

This perspective comes from over a decade of building companies in Silicon Valley, two successful exits (RemoteTeam to Gusto, MovieLaLa to Gfycat), and investing in 200+ startups including Anthropic, OpenAI, and Scale AI. I write about what I've lived.

What's the most common pushback you get on this?

People often push back by citing exceptions or edge cases. And they're usually right that exceptions exist. But building a strategy around exceptions rather than patterns is a losing game for most founders.

How can I apply this thinking to my own situation?

Start by identifying the core principle behind the opinion, not the specific example. Then ask yourself: does this principle apply to my context? If yes, test it in a small, low-risk way before going all in.

How has this view evolved over time?

My thinking on most topics has changed significantly over the years. Early in my career, I held many conventional views that experience proved wrong. I try to update my beliefs when the evidence changes.