How I Use Agent Handoffs to Work 4 Hours a Day

By Belkis Cruz

June 10, 2026 · Build With Me

One of my agents (Monroe) is supposed to execute the weekly content batch. Seven blogs, seven newsletters, seven Instagram briefs. 21 pieces from a template, due every Sunday.

Last month, Monroe got stuck on a blog outline. She had the prompt. She had the pillar. But the topic needed context about Belkis’s actual experience running a $4M remodeling firm, and Monroe didn’t have it. So she stayed stuck for 24 hours waiting for me to clarify.

I realized the problem wasn’t Monroe. It was the handoff.

The handoff was my fault.

I gave her a task. I didn’t give her the mental model.

So I rebuilt the entire content operation around “handoff documents” instead of “task assignments.” Now when Monroe gets a piece, she’s not reading a vague instruction. She’s reading a markdown file that holds the decision tree, the context, the 3-5 edge cases, and the skip conditions.

The handoff document says:

• What does a “Ready for Review” piece actually look like?
• Which pieces require affiliate links (read from Notion), and which don’t?
• If a blog needs a contractor voice, here’s what that means (with examples).
• Here’s what Belkis approved last time, here’s what she killed.
• If you get stuck on X, the answer is usually Y. Here’s how to know when that doesn’t apply.

Monroe executes the handoff document. If something doesn’t fit the document, she logs it and asks. That’s the only case where I jump in.

The shift: I went from “manage the execution” to “maintain the handoff documents.” That’s about 4 hours a week instead of 20.

Why task assignment doesn’t work for knowledge work

Task assignment is designed for work that’s simple and repeatable. You hand someone a task, they do it, they’re done.

“Run payroll.” “Order lumber.” “Call the client.” Those are tasks.

But knowledge work doesn’t work that way. When Monroe needs to write a blog about pricing strategy, she needs:

• The contractor voice, which is different from the personal wellness voice.
• The specific pillar (is this Real Numbers or Build With Me?), which changes the angle.
• Knowledge of what I’ve already published (don’t repeat).
• Knowledge of what I’ve killed before (don’t pitch this again).
• Knowledge of my brand constraints (no these phrases, always these details).
• Access to the affiliate vault if links are needed.
• The decision tree for “when is a blog ready to hand to Belkis” vs “when should I redraft it myself?”

If I give Monroe a task, I’m leaving out half of that. She’ll write something. It’ll be wrong. I’ll have to rewrite it. We both lose.

If I give Monroe a handoff document with the mental model built in, she can make the decisions herself. She only loops me in when the document doesn’t cover it.

What a handoff document actually contains

A handoff is a markdown file that lives where the work lives. It’s not an instruction manual. It’s the decision-making framework.

For content, my handoff document covers:

Notion data sources and how to use them:

• Here’s the Affiliate Link Vault collection ID. When you mention a product, pull from here first. Use the Levanta URL verbatim. If the product isn’t in the vault, note it in your draft and ask.
• Here’s the Content Project Tracker ID. Every piece you write gets a row here with status=Draft and category=(pillar).

Voice requirements by pillar:

• Real Numbers: Contractor-fluent, dollars and percentages, specific examples.
• Build With Me: First person, process-focused, tactical details about how I run the business.
• Truth Bombs: Direct, no varnish, what contractors won’t say out loud.
• Bel’s Edits: Personal and operator-focused, affiliate-anchored, humor allowed.

Examples of “approved” and “killed” work:

• Here’s a blog Belkis loved: (link + why).
• Here’s a blog she killed: (link + why, don’t do that again).
• Here’s a Substack that worked: (link + specific phrases she approved).

Skip conditions (if X, skip drafting and log it):

• If the blog topic duplicates something published in the last 90 days, skip and note.
• If a Substack idea uses a banned phrase (see list), redraft before sharing.
• If an IG brief requires a product photo and the photo isn’t in the vault, note the missing photo and ask before finalizing.

Decision tree (when to ask vs when to decide):

• You have all the context to decide: writing style, pillar, affiliate needs, voice mode.
• You need to ask: if the topic seems important but you’re not sure it fits the pillars, or if Belkis approved something unusual that doesn’t fit the framework.

The approval boundary:

• These are ready for Belkis without her input: drafts with status=Draft, no promises made about publish dates.
• These need explicit Belkis approval before they ship: anything with a publish date, anything with a specific product link, anything that commits budget (paid ads, promotion spend).

When Monroe works from this document, she doesn’t need me to assign her individual pieces. She knows what done looks like. She knows when to ask. She knows what gets skipped.

How this scales

A handoff document is portable. When I hire the next agent or when I shift Monroe to a different revenue stream, the new owner can pick up the handoff document and execute the same way.

It’s also updatable. If I change my voice, or if Belkis decides on a new pillar, I update the handoff document. All future work uses the new framework.

For my operation right now (one person + 7 agents), I have 6 handoff documents:

• Content execution (Monroe)
• Content strategy (also Monroe)
• Design and graphics (Vivienne)
• Operations and scheduling (Athena)
• Paid ads and insight mining (Monroe also, some days)
• List building and email automation (waiting on hire)

Each one is the operating manual for that piece of work. Belkis approves once at the decision boundary. I maintain the documents. Agents execute.

Time investment: writing the handoff document takes 3-4 hours. Maintenance is maybe 1 hour per week. Versus the 20 hours it used to take to manage task assignment, clarify on the fly, and redraft work that was wrong.

The math wins fast.

Why this is different from “better instructions”

Agencies and content shops give their teams “better briefs.” Longer instructions. More examples. More detail.

A handoff document is different. It’s not an instruction. It’s a decision framework. It assumes the agent is capable (my agents are). It’s saying “here’s what done looks like, here’s the decision tree, go.”

A brief says “write a blog about this topic with this structure and these elements.” The team follows it.

A handoff document says “here’s what this role owns, here’s the decision you need to make, here’s how you know you’ve made it right.” The agent makes decisions.

That shift changes everything. Because knowledge work requires judgment, not just execution. And judgment happens faster when the decision framework is clear.

I used to joke that I couldn’t take a day off without the whole operation falling apart. Now I can because the handoff documents are the operation. I just make sure they’re still true.

—