How a Vice Mayor Juggling Four Jobs Built an AI Executive Assistant

Jun 15, 2026·7 min·By Nicolas Zeeb

Customer Stories

How a Vice Mayor Juggling Four Jobs Built an AI Executive Assistant

Chris Caputo does not have one job. He is the Vice Mayor of Wilton Manors. He is the CTO at UNCS, where the team is building AI product-sourcing tools. He is the founder of AffiniteSeas, a travel tech startup. And he sits on several nonprofit boards he cares deeply about.

Any one of those would fill a calendar. Held together, they fill four, each with its own schedule, its own inbox, its own people who need an answer today. For a long time that pile-up had a cost Chris could name out loud: he felt like he was doing a mediocre job at all of them.

Not because he was bad at any one role. Because no person can hold four contexts in their head at once and give each the attention it deserves. Something always slips. Usually it was the personal stuff, the things that have no deadline until they become an emergency.

Quick overview

Chris Caputo runs four roles at once, Vice Mayor of Wilton Manors, CTO at UNCS, startup founder, and nonprofit board member, and used to feel he was doing a mediocre job at all of them.
He built one AI executive assistant to hold every role, with a consolidated morning brief and nightly recap that synthesize what he needs to be aware of, prepare for, or act on across all of them.
He found Vellum after frustrating OpenClaw installs: what took three or four days to configure happened in under thirty minutes.
The payoff is doing each job well instead of four jobs poorly, with his hours going to the decisions only he can make in each role.
Where it is heading: an assistant that responds to a share of his thousands of unread messages and preps the rest exactly as he would, holding for his approval.

Four jobs, one brain

The trap of running multiple roles is that the work multiplies but the day does not. Every role wants a morning. Every role generates email, meetings, decisions, follow-ups, and the quiet maintenance that keeps a thing from falling apart. A city does not pause because a product launch is live. A startup does not wait because a board meeting ran long.

So you triage by whatever is loudest. The roles with the most urgent fires get your hours, and the rest run on fumes. Chris wanted the opposite: to manage every schedule from one place, to spend his attention on what actually mattered in each role, and to automate the parts that did not need him at all. Most of all he wanted to stop letting his personal life fall through the cracks while he kept four professional plates spinning.

The solution turned out to be an AI executive assistant he could build himself.

Why Chris switched to Vellum

Chris did not arrive at Vellum first. He arrived frustrated.

He had already seen the value of an always-on assistant, so he tried OpenClaw, both self-hosted and hosted. The idea was right and the experience was painful. The setup and training time was, in his words, ridiculous, and the thing crashed constantly. He was sold on the concept and worn down by the tool.

One frustrated afternoon he searched for "better version of OpenClaw" and landed on Vellum. What took him three or four days to configure on OpenClaw happened in under thirty minutes. He was immediately sold. Then something rarer happened: the dev team, and the CEO himself, acted on his feedback right away. That was the moment he knew he had found something that would last.

For someone running four roles, the criteria were simple, and Vellum met them:

Runs as a native Mac app on your own machine or in Vellum Cloud, always on and not crashing out from under you
Setup measured in minutes instead of days, so the tool earns its place before it costs you a week
One shared memory across Mac, iOS, web app, voice, email, Telegram, and Slack, so every role lives in the same context instead of scattered across apps
A skills system to automate the recurring work and leave the judgment calls to him

The deciding factor was not a feature, though. It was that the people behind the product listened when he told them what was broken.

The morning brief that holds all four roles

Here is the shape of it. Chris built his assistant to run two synchronized passes over his entire life, every day.

In the morning it produces a consolidated brief. The assistant reads across all four roles at once, the city work, the CTO work, the startup, the boards, and pulls together everything he needs to be aware of, prepare for, or act on that day. Instead of opening four calendars and four inboxes and assembling the picture himself, he opens one brief that has already done the synthesis. The city meeting at noon, the product decision waiting on him, the board document due, the personal thing he would otherwise forget. One view, every hat.

At night it runs the mirror image: a recap that closes the loop on the day and stages what is coming. What moved, what stalled, what needs him tomorrow. The nightly pass is how the personal life stops slipping, because nothing that mattered gets left in a single role's silo where he might not see it until it is too late.

The point is not the brief itself. The point is that one assistant holds all four contexts so Chris does not have to. The prioritizing happens before he reads a word, and the parts that do not need a human get handled without one.

Time back to do four jobs well

The payoff is straightforward. When the synthesis and the triage happen automatically, the mediocre-at-everything feeling goes away, because Chris's actual hours go to the decisions only he can make in each role, including the nonprofit work he took on because he cares about it, not because it pays.

He is honest that he is still early. He is working through some bugs, or operator error, on app and skill deployment, and he says the potential is already obvious. The clearest picture of where this goes is his communication backlog, which is staggering and which he describes without embarrassment: he has never had fewer than a thousand unread texts and fifteen thousand unread emails, and his voicemail has been full for as long as he can remember. No human can dig out of that.

He can see a future where his assistant personally responds to fifty to sixty percent of those messages, and preps another twenty percent exactly the way he would handle them, reaching out for his approval before it acts. He had a human personal assistant for years, and he can already see how a Vellum assistant could be trained to do anything they could. Except, he notes, handle his Amazon returns.

How to build this yourself

You do not need four jobs to use this pattern. You need more inputs than one head can track. Here is the assembly:

Pick the roles or areas that compete for your attention. For Chris it was four jobs; for you it might be a job, a side project, a board seat, and a family. List them honestly.
Connect each role's sources to one assistant: the calendars, the inboxes, the channels where the work actually shows up. The goal is a single place that can see everything.
Build the morning brief. Have the assistant read across all of it and produce one synthesis of what you need to be aware of, prepare for, or act on today, sorted by what actually matters.
Build the nightly recap. Same read, run in reverse: what moved, what stalled, what needs you tomorrow. This is the pass that keeps the low-urgency, high-importance things from disappearing.
Hand off the recurring, no-judgment tasks to skills, and keep yourself on the decisions. Set the rule that anything consequential comes to you for approval before the assistant acts.

Done this way it is less a software project and more a daily rhythm you assemble in an afternoon, then let run.

Main takeaway

The pattern travels well beyond a Vice Mayor's office. Anyone holding several roles at once hits the same wall: the work scales, the day does not, and attention gets rationed to whatever is on fire. An AI executive assistant changes the math by holding every context for you and surfacing only what needs a human, so each role gets your judgment instead of your leftovers.

What Chris got back was the feeling that he is doing each job well, not four jobs poorly. That is the quiet promise of an assistant that never loses the thread across any of them.

Chris found this in Vellum. If you are running more roles than one person can track, hatch your own at vellum.ai.

What is an AI executive assistant?

An AI executive assistant is software that handles the coordination work of a senior role: reading across your calendars and inboxes, briefing you on what matters, prioritizing decisions, and automating the repetitive tasks that do not need your judgment. Unlike a scheduling bot, a capable one like Vellum holds the full context of your work and acts within rules you set.

Can one AI assistant manage multiple jobs or schedules at once?

Yes, and that is where it earns its keep. By connecting every role's calendar and inbox to one assistant, you get a single synthesized view across all of them instead of switching between apps. Chris Caputo runs four roles, Vice Mayor, CTO, startup founder, and board member, through one assistant that briefs him across all four every morning.

How is this different from a human executive assistant?

A human assistant is limited by hours; an AI one runs continuously and scales to volume no person could touch, like a backlog of thousands of unread messages. The trade is judgment and relationships, which stay human. Many people, Chris included, see the two working together, with the AI handling synthesis and routine response and the human staying on the relationships that matter.

Is it hard to set up an AI assistant for executive work?

It does not have to be. Chris had spent three or four days configuring a self-hosted alternative that kept crashing; setting up Vellum took him under thirty minutes. The practical test is whether the tool earns its place in the first sitting rather than costing you a week of configuration.

What can an AI executive assistant actually do day to day?

The core loop is a consolidated morning brief and a nightly recap that synthesize everything across your roles: what to be aware of, what to prepare for, and what to act on. From there it automates recurring tasks, drafts responses for your approval, and keeps low-urgency items from slipping. The work it does grows as you train it.

Can it respond to my emails and texts for me?

It can draft and, with your permission, send. A reasonable setup has the assistant handle a share of routine messages directly and prepare the rest the way you would write them, holding for your approval before anything goes out. That approval gate is what makes delegating high-volume communication safe.

Will it act without my approval?

Only if you let it. The standard pattern keeps a human in the loop for anything consequential: the assistant preps the action exactly as you would take it, then waits for your yes. You decide which tasks are automatic and which always route through you.

Does my data stay private with an AI assistant?

It depends on the tool, so check before you commit. With Vellum, the Mac app runs on your own machine and Vellum never has access to your data on any deployment path. For someone holding sensitive work across a city office, a company, and a startup, that boundary is the difference between using the tool and not.

What kinds of professionals benefit most from an AI executive assistant?

Anyone whose roles outnumber their hours: founders, executives, elected officials, board members, and people stacking a job with serious side commitments. The common thread is too many contexts to hold at once, which is exactly the problem synthesis and triage solve.

How much does an AI executive assistant cost?

Vellum has a free Base plan to start. Pro is from $50 a month with pay-as-you-go credits, configurable compute and storage, and your assistant's own email and subdomain. The honest way to weigh it is against the hours you lose to coordination, or the cost of a human assistant for the same work.

Where do I start if I want to build one?

Start by listing the roles competing for your attention, then connect each one's calendar and inbox to a single assistant, like Vellum. Build a morning brief and a nightly recap first, since those deliver value on day one, then hand off recurring tasks to skills as you learn what you can trust it with.