Breadth versus depth: how to weigh AI agents for marketing

What breadth buys

Breadth is the number of marketing roles one agent can hold. Its first benefit is fewer handoffs. When the same agent plans the calendar, makes the post and reports on it, nothing has to be exported from one place and explained in another.

Its second benefit is shared knowledge. An agent that saw which posts drew clicks can use that when it drafts an ad or an email. An agent that holds only one role starts from whatever you tell it.

What depth buys

Depth is how far the work goes inside a single role before a person has to step in. A shallow ads role produces a picture and a caption. A deeper one produces a full draft with a destination, a budget and the cost stated, held for approval.

Depth matters most where mistakes are expensive or the work is technical. If most of your budget goes to paid ads, a thin ads role will cost you more than three missing roles elsewhere.

Where broad agents tend to thin out

Breadth is easy to claim and hard to finish. Three signs suggest a role is listed but thin.

• The output is advice, not work. You asked for an ad and received tips on writing one.
• The role ends at a file. The post is made, and publishing it is left to you.
• There is no approval state. Nothing shows whether a draft is waiting, approved or live.

A way to decide in ten minutes

Draw a table with your roles down the side and two columns. In the first, write how much the role matters to revenue this year: high, medium or low. In the second, write who does it today and how many hours it takes. Then read the table.

• Many medium rows, nobody with enough hours. You need breadth. Look for an agent that covers most rows adequately.
• One high row that dwarfs the rest. You need depth there. Assess that role first and treat the others as a bonus.
• Two or three high rows. You need both. Shortlist broad agents, then set your hardest task in each high row.
• A high row no agent covers. Keep a specialist tool or person for it and choose the agent on the remaining rows.

Pairing an agent with a specialist tool

Breadth and depth do not have to come from one place. A common arrangement is a broad agent for the day-to-day roles and one specialist tool for a role that sits outside it. Search is the usual example, and the directory describes two such resources by role.

The cost of pairing is a handoff, so keep the number small. Decide who carries information across, for example who tells the person running search what the social calendar says this month. If you find yourself pairing in most of your rows, you have chosen a narrow agent and should compare it with a broader one.

Where Link AI fits

For the first and third cases in the table exercise, where several roles need covering at once, our pick is Link AI by Linkme. Linkme's page lists nine roles and shows an output for each, which makes depth something you can ask to see instead of take on trust. Take the ads role: a Meta ad draft shows the daily budget next to what it adds to the month, and it stays at 'Needs review' until someone approves.

Where you need depth beyond what the page describes, check it on your own work. Set your hardest task in your highest-value role when you book a demo to try Link AI. Link AI does not do SEO, website content or search ads, so plan those rows separately.

Connected entries

Questions & answers