Meet Your Agents: Understanding Cowork's Autonomous Workers

This is Part 7 of a 16-part series on building your AI VP of Marketing with Claude Cowork. Previous: Skill: Brand Voice Checker | Next: Agent: Campaign Strategist

# Marketing VP -- Acme Widget Co.

This is the operating brief for all marketing work in this project.

## Business Overview
Acme Widget Co. sells industrial-grade fastening systems to manufacturing
plants across the Midwest. Founded 2019, 12 employees, $2.4M annual revenue.

## Target Audience
Operations managers and procurement leads at mid-size manufacturing plants
(50-500 employees). They care about reliability, lead times, and total cost
of ownership.

## Brand Voice
Direct, technical, no-nonsense, helpful. We sound like an experienced
engineer explaining a product -- not a salesperson pitching one.

# Brand Voice
## Tone
Direct, technical, no-nonsense, helpful, specific.
## Words We Never Use
game-changing, revolutionary, synergy, leverage, empower
## Sentence Structure
Lead with the fact or the benefit. Active voice always. Short paragraphs.

What Are Agents?

A Skill is a form your VP fills out -- you designed the fields, and the output follows a predictable structure every time. An Agent is a senior team member you hand a project brief. The brief says what you need and why. The team member figures out the research plan, the analysis framework, and the presentation format. They come back with a recommendation, not a filled-in template.

Until now, your VP has worked in two modes: conversations (Articles 1-4) and Skills (Articles 5-6). Both require you to drive. Agents are different. You hand them a goal, and they figure out how to get it done.

When you give a task that requires research, analysis, and synthesis -- the kind of work that would take a human team member an afternoon -- Cowork spawns agents to handle it. They break the task into steps, execute each step, and combine the results into a finished deliverable. You tell them what you need. They determine how to get there.

Some marketing tasks are repeatable and structured -- generating content briefs, checking brand voice, formatting social posts. Those are Skill work. Other tasks require judgment and synthesis -- competitive analysis, campaign strategy, market positioning. Those are Agent work. You cannot template your way through a competitive field because the competitors change, the market shifts, and the right framing depends on context you cannot predict in advance.

Your job shifts from driving the work to reviewing it -- the same shift that happens when you hire a capable human and stop micromanaging them.

Skills vs. Agents

You have built Skills. You are about to use Agents. The difference between them determines which tool you reach for on any given task.

Use Skills when you know exactly what you want and the format it should take. Content briefs, voice checks, social media posts. You designed the template in Articles 5 and 6 because you already know what a good content brief looks like. The Skill ensures consistency. Every brief follows the same structure, hits the same quality bar, and takes the same amount of time.

Use Agents when you need research, analysis, comparison, or synthesis. Strategy work. Market analysis. Campaign planning. You cannot template these tasks because the right answer depends on information you do not have yet. You need someone to go find that information, make sense of it, and come back with a structured recommendation.

Skills and Agents are not competing approaches. They complement each other. An Agent running a competitive analysis might use your Brand Voice Checker Skill to ensure the final deliverable matches your tone. An Agent building a campaign strategy might pull from your Content Brief Generator Skill to structure individual content pieces within the campaign. The Skills you built in Articles 5-6 become tools your Agents can use, the same way a senior team member uses your company templates without needing to be told they exist.

The simplest test: if you can draw the output on a whiteboard before the task starts, use a Skill. If you need someone to go figure out what the output should look like, use an Agent.

Your First Agent Task: Competitor Analysis

Your VP has context. It knows your business, your audience, your voice, and your standards. It has Skills for generating content briefs and checking brand voice. But it has not done any independent research for you yet.

That changes now.

Competitive analysis is one of the highest-value tasks a marketing VP handles. Understanding how your competitors position themselves -- their messaging, their channels, their pricing -- tells you where the openings are. Most small business owners skip this work because it takes hours of research and the output is hard to structure. It is exactly the kind of task Agents handle well: multi-step, research-heavy, and requiring synthesis across multiple sources.

Copy the following prompt and paste it into your Cowork project:

Research my top competitors and produce a positioning matrix.

Using the business context from my briefing document:
1. Identify 3-5 direct competitors in my space
2. For each competitor, analyze: their core offering, target audience, key messaging themes, pricing model (if public), and primary marketing channels
3. Compare their positioning to ours
4. Produce a structured comparison table with our business in the first column
5. End with 2-3 positioning opportunities — gaps in competitor messaging that we could own

Be thorough. Take your time researching each competitor before synthesizing.

That is it. No template. No fields to fill in. A goal and a set of expectations.

Watch It Work

When you run that prompt, something different happens. Unlike the Content Brief Skill, which returned a formatted document in seconds, this task takes time. Your VP is doing real work -- not generating from a template, but researching, comparing, and synthesizing.

You can watch the progress in Cowork. Your VP shows what it is doing at each stage: identifying competitors, pulling information about each one, analyzing their messaging, comparing positioning across the set, building the comparison matrix. The work unfolds in a visible sequence.

Nobody told your VP to research Competitor A first, then Competitor B, then Competitor C. Nobody specified the analysis framework or the comparison dimensions. It broke the task into steps on its own, decided the order, and chose how to structure the comparison.

Pay attention to the synthesis step. The positioning matrix is not five separate competitor profiles stapled together. It is a comparative analysis -- each dimension shows how competitors relate to each other and to your business. One competitor owns the budget positioning. Another dominates on ease of use. A third is pushing hard on a feature set nobody else mentions. These patterns emerge because the agent holds all the research in context simultaneously -- something that would not be visible from looking at any single competitor in isolation.

Review: Competitor Positioning Matrix

Your VP delivers a positioning matrix. Before you use it, you need to evaluate it the same way you would evaluate work from any team member -- with specific criteria, not just a gut check.

Here is what the deliverable might look like. This example uses Tideway Bookkeeping, the fictional business from Article 1:

Your matrix will look different. The structure might vary too -- your VP may add rows you did not expect, or organize the comparison differently. The value is in the comparison, not the format.

Now evaluate the work.

Are the competitors real? Your VP researched these autonomously. Verify they exist and are actually direct competitors, not adjacent businesses serving a different market. A bookkeeping service for enterprise companies is not a competitor to a freelancer-focused product, even if both do bookkeeping.

Is the analysis accurate? Spot-check one or two claims against what you already know or can quickly verify. Does QuickCount actually use per-transaction pricing? Is FreelanceBooks really freemium? Your VP works from available information, and that information can be incomplete or outdated. One quick verification per competitor is enough to calibrate your trust in the rest.

Do the positioning opportunities make sense? Your VP identified gaps -- spaces in competitor messaging that nobody owns. Do those gaps align with your actual strengths? Could you realistically claim that positioning? A gap exists between "enterprise security" and "consumer simplicity," but if your product is neither, that gap is not yours to fill.

Is anything missing? You know your market better than your VP does. Did it miss a major competitor? Did it overlook a positioning angle you already use? Did it group competitors in a way that obscures an important distinction?

The positioning matrix is a working document, not a final answer. Use it as a starting point for strategic decisions. Which gaps are worth pursuing? Which competitor moves should you watch? Where is your current messaging strongest, and where does it blend into the noise?

When Agents Get It Wrong

Agents are autonomous, not infallible. The same independence that makes them powerful also means they make judgment calls you might disagree with. Knowing where Agents tend to stumble helps you review their work efficiently.

They may identify competitors you do not actually compete with. Your VP works from your CLAUDE.md and available research. If your briefing document describes your business broadly -- "we do bookkeeping" rather than "we do flat-rate bookkeeping for freelancers earning $75K-$250K" -- the competitor set will be broad too. Specificity in your brief produces specificity in the analysis.

They may oversimplify positioning. Reducing a competitor's entire marketing strategy to one row in a table necessarily loses nuance. The matrix is a starting point for strategic thinking, not a substitute for it. If a competitor's positioning is more complex than a single phrase captures, note that and dig deeper in a follow-up prompt.

They may miss recent changes. If a competitor just pivoted their pricing model last month or launched a new product line last week, your VP may not have that information. Treat the matrix as a snapshot, not a live dashboard. Update it when you learn something new.

None of this makes Agents less useful. A human team member doing the same analysis would make different judgment calls, miss different things, and bring different blind spots. The review process from Article 3 applies here with the same force: you are still the boss. You review the work, correct the errors, and approve the output before acting on it.

The Agent Mindset

Using Agents well requires a shift in how you work with your VP. With Skills, you are the architect. You design the template, choose the fields, and control the output structure. With Agents, you are the executive. You set the objective, define success criteria, and review the deliverable.

That shift changes what makes a good prompt. A good Skill prompt is precise and structured: "Generate a content brief using these inputs." A good Agent prompt is clear about the goal and generous with context: "Research my competitors and produce a positioning matrix. Here is what I need to make strategic decisions."

Two principles for working with Agents:

Define the objective, not the method. "Produce a positioning matrix" is better than "first search for competitors, then look at their websites, then compare their pricing pages." But do provide context that shapes judgment -- "we compete mainly on price" or "our biggest differentiator is customer service" gives your VP a lens for the analysis. Without that lens, it treats all dimensions equally.

Review with criteria, not feelings. "This does not feel right" is hard to act on. "The competitor set is missing our two biggest rivals" or "the pricing analysis is outdated for QuickCount" gives specific corrections that improve the work.

You have been building toward this since Article 1. The CLAUDE.md briefing document gives your Agent context. The Rules from Article 4 give it constraints. The Skills from Articles 5-6 give it tools. Agents pull all of that together and apply it to open-ended work.

What Is Next

You watched Agents work. You saw your VP break a complex task into steps, research autonomously, and synthesize findings into a deliverable you can act on. A significant step up from Skills -- but you did not control how the Agent worked. You gave a prompt and reviewed the output. The Agent decided everything in between.

In Article 8, you build your first custom Agent: a Campaign Strategist. Instead of relying on general-purpose agent behavior, you define exactly what your Campaign Strategist does -- what inputs it expects, what steps it takes, what output it produces. You move from using Agents to building them. The difference is like the difference between hiring a contractor and training a full-time employee. Both get work done. But the employee works your way, every time.

This is Part 7 of 16 in the Your AI VP of Marketing series. Previous: Skill: Brand Voice Checker | Next: Agent: Campaign Strategist