The AI EA Flex

Show Notes

Will Ruben spent more than a decade at the companies that taught the internet what attention looks like. He led ranking and recommendations across Instagram during the era when Reels stopped being a feature and started being the entire product. He worked on Coinbase's Web3 Wallet. He scaled consumer products for billions of people. And then he walked away from all of it to solve something almost embarrassingly small in scope: the back and forth of scheduling a meeting.

That choice is the whole story. Will is not building Workmate because scheduling is glamorous. He is building it because scheduling is the gateway drug to giving every knowledge worker the kind of strategic support that used to be reserved for executives with assistants and corner offices. The premise is democratization, the wedge is the calendar, and the long arc is a world where you collaborate with a mix of humans and AI teammates that feel indistinguishable from coworkers.

In conversation with Ryan, Will lays out a thesis that is unusual in this AI moment. While most founders are racing to make their agents louder, faster, and more obviously artificial, Will is doing the opposite. Workmate is engineered to disappear. It has an email address at your domain. It writes the same way every time. It is white-labeled, customizable, and in many cases, the people interacting with it do not know they are talking to AI. Will calls this a flex. The flex is appearing more important than you are.

The conversation winds through the ethics of disclosure, the speed of building when the foundation models change every two months, the difference between sculpting and painting, and a tangent on Instagram Reels. It also lands somewhere unexpected: a quiet, almost paternal argument that the founders who win in this era are the ones who go to bed on time.

Frameworks from This Episode

The Trust Curve in AI Disclosure

Will frames the disclosure question not as a binary but as a function of industry, demographic, and medium.

›Internal team communication: full transparency is the default because users know they are working with the product.
›External client communication: depends on industry norms — some sectors expect executive assistants, where AI fits seamlessly into existing expectations.
›The Workmate position: provide both options and let the customer choose the level of transparency.
›The bet: in two years the question will dissolve entirely because AI teammates will be normalized the way remote work was normalized between 2015 and 2025.

The Three Waves of Instagram

Will identifies three distinct product eras at Instagram, each of which informs how he is building Workmate.

›Wave one: filters on the feed — self-expression.
›Wave two: stories — ephemeral connection.
›Wave three: constant content recommendations and Reels — algorithmic discovery.
›The takeaway for AI: the third wave succeeded because it gave users more control over what they saw, not less. Workmate applies the same principle to scheduling preferences.

The Sculpting versus Painting Distinction

Will and Ryan agree that the founder's job is shifting from execution to taste.

›Painting: the founder hand-crafts the output.
›Sculpting: the founder shapes what AI produces by setting parameters, reviewing direction, and arbitrating quality.
›The implication: management skills, not technical execution, become the bottleneck.
›The catch: agents are not fully autonomous yet, so founders still cannot fully step away.

The Stable Ground Problem

Building a company on an AI stack that changes every quarter forces a different relationship with planning.

›Long term: vision and mission still hold.
›Short term: weekly and monthly sprints still work.
›Medium term: the 6 to 18 month window is genuinely impossible to forecast.
›Will's response: rebuild the AI stack every few months and treat throwing out the old code as a feature, not a failure.

The Pick Your Battles Operating System

Will's personal life philosophy, applied to product decisions.

›The perfect is the enemy of the good.
›The last 20 percent of polish costs 80 percent of the time.
›In an AI world, that final 20 percent still often requires a human.
›The discipline is knowing when the last 20 percent matters and when 90 percent is the right place to stop.

Founder Experiment: Build Your Own Scheduling Agent in a Weekend

Open Cursor or Replit. Spin up a new project. Use the Anthropic API to build a lightweight scheduling assistant that lives in your inbox.

1Connect Claude to a Gmail inbox via an OAuth flow.
2Give it read access to your Google Calendar.
3Write a system prompt that defines your scheduling preferences: preferred meeting windows, time zone, default meeting length, blackout days, recurring blocks for focused work.
4Have it monitor for incoming emails that look like scheduling requests.
5When detected, draft a reply with three proposed times and CC you for approval before sending.
6Add a confirmation layer that writes the event to your calendar once the recipient agrees.

The lesson: Not to compete with Workmate. It is to feel the difference between a scheduling tool that follows rules and a scheduling agent that exercises judgment. You will hit the same problems Will hit. You will discover why consistency in voice matters. You will understand why the trust layer is the actual product.

Key Terms

Workmate: An AI-powered executive assistant that handles scheduling, calendar management, and the back and forth coordination that knowledge workers spend hours on every week.

White Labeling: The practice of removing the underlying brand from a product so it appears to come from the customer. For Workmate, this means assigning the AI an email address at the customer's own domain.

Coordination Debt: The cumulative time cost of all the back and forth required to schedule meetings, resolve conflicts, and reschedule when things change.

Foundation Models: The base large language models — Claude, GPT, Gemini, and others — that AI products are built on top of. Will mentions rebuilding Workmate's AI stack multiple times as foundation models improve.

Computer Use (Open Claw): A capability that allows AI agents to operate a computer directly, clicking buttons and navigating interfaces the way a human would. Will's team uses it to manage paid advertising.

Booking Link: A calendar link — Calendly being the canonical example — that lets others self-schedule a meeting based on shared availability. Will argues this approach is commoditized and inferior to assistant-style scheduling.

Calendar Conflict: When two events are scheduled on top of each other. Workmate proactively flags these before they become a problem.

AI Slop: Low-quality, obviously machine-generated creative content. Will admits AI-generated ads still sometimes feel like slop, even when they perform.

Agentic AI: AI systems that can take actions in the world, not just generate text. A scheduling agent that books a meeting on your behalf is agentic.

Coordination Layer: The infrastructure that handles all the in-between work in a knowledge workflow. Workmate is building the coordination layer for meetings.

Q&A

What is Workmate?

Workmate is an AI-powered executive assistant founded in 2024 by Will Ruben. It handles scheduling, calendar management, and the back and forth coordination that knowledge workers spend hours on every week. Workmate operates over email, Slack, Microsoft Teams, and text, with the option to use a custom name and a custom email at the user's domain.

Who founded Workmate?

Workmate Labs was founded in 2024 by Will Ruben, a former senior product leader at Meta and Instagram who also worked at Coinbase and Uniswap. Will led content ranking and recommendation algorithms at Instagram during the rise of Reels.

How is Workmate different from Calendly?

Calendly and similar tools rely on the user manually constraining their availability and the recipient picking a time from a link. Workmate operates like a human executive assistant, exercising judgment over preferences, following up persistently when meetings stall, proactively flagging calendar conflicts, and adjusting schedules across complex constraints like travel and time zones.

Can Workmate appear as a human to outside contacts?

Yes. Workmate supports full white labeling, including a custom name and an email address at the user's own domain. Some customers prefer this for the implied status of having an executive assistant, while others prefer full transparency. Workmate supports both.

What is Workmate's pricing model?

Workmate offers a two week free trial. Pricing is customized for larger teams and white glove for enterprise customers, with a lower individual tier available. The company states pricing is lower than the cost of a human executive assistant.

What integrations does Workmate support?

Workmate works with Google Calendar, Microsoft Outlook, Slack, Microsoft Teams, Zoom, and SMS. It handles scheduling across multiple calendars and multiple time zones.

How does Workmate handle scheduling conflicts?

Workmate proactively scans the user's calendar for conflicts and sends summaries on Sunday night and Monday morning for the week ahead. Users can also ask the assistant to clear a day, reschedule across a range, or set rules like blocking the days before and after national holidays.

Has Workmate raised funding?

Yes. Workmate Labs has raised a seed round to build the product. The company is headquartered in New York.

Links from This Episode

Workmatehttps://www.workmate.com
Will Ruben on LinkedInhttps://www.linkedin.com/in/wruben
AI for Foundershttps://aiforfounders.co