← Back to blog

Reviewing Last Month's Campaign Results with an AI Researcher (Coming Soon to 8frame)

How to turn last month's ad results into next month's creative brief: a review routine you can run on a shared board today, and how an AI Researcher agent that reads the report and the ads on the board will help when it ships on 8frame.

Coming soon on 8frame. An AI Researcher that reads your results report and your ads on the board, and lays out what worked and what to try next, is not released yet. This page covers a monthly creative review you can run on a shared board today, and how the Researcher and documents on the board will fit into it when they ship. No date is announced.

The monthly review is where creative learns from media. The numbers arrive in an export from the ad platform, the ads themselves sit in a folder, and the discussion happens in a meeting. Too often the outcome is a slide that says "video outperformed static" and a next brief that ignores it. The point of the review is narrower and more useful: which creative choices moved results, and what the next round should test.

It matters because creative is a large share of what drives results. Nielsen wrote in 2017 that creative "remains the undisputed champ in terms of sales drivers", even as the effect of media on sales had grown "to 36% from 15%" over the previous 11 years.

A review routine that works

  1. Put the ads and the numbers side by side. Not in two files. Each ad next to its spend, its main result and its audience.
  2. Describe each ad in creative terms. Hook in the first seconds, format, presence of a person, product framing, offer, colour. These are the variables the next brief can change.
  3. Separate signal from noise. An ad with very little spend tells you little. Note the sample before you note the winner.
  4. Write "what worked" and "what to try" as two short lists. Each item should name a creative choice, not a channel.
  5. Draft the next brief from those lists. Audience, message, two or three things to test, formats.
  6. Agree in writing, so the next round starts from a decision and not from memory.

For the client-facing side of this, see how to report creative performance to a client; for the meeting, how to run a creative review meeting.

Running it on 8frame today

The canvas does not take documents yet (uploads are images, video and audio), so the numbers travel as comments for now:

The board then holds the review and the start of the next round in one place. For structuring the tests themselves, see the AI ad variant testing workflow.

What is coming: the Researcher on the board

The home page describes this use case in one line: last month's results become next month's brief. Here is how the parts will fit together.

The report goes on the board

You will be able to place documents on the board, such as a PDF report or a CSV export of the numbers, next to the ads themselves.

A Researcher reads it

The Researcher will be an agent node: a generic Agent node you name and give a role, or a ready-made Researcher role. You will wire in the report and the numbers, and hand it the ads by selecting them and choosing to ask an agent. It will read only what is wired in, and its node will list what it read.

You will be able to @Researcher in the board chat, like a teammate: "What worked last month, and what should we try?" The answer will come back into the chat under its name, and it will lay out its findings on the board inside its own frame, next to your ads, where the team can comment on them.

Reading costs no approval; making does

Reading, commenting and proposing spend no credits, so the Researcher will be able to do them straight away. Generating the next round's stills is different: any step that spends credits will wait behind an approval card with Approve and Not now. See approving what an AI agent does before it spends credits.

From the agreement to the brief

Once the team agrees in comments, you will be able to give a text node the agreed points as its task, with the report wired in, and get a draft of the next brief that you can turn into editable text on the board. See reading briefs and documents with text models on the canvas.

What an AI reading of results can and cannot do

Useful: summarising a long report, lining each ad up with its numbers, listing the creative features the winners share, and drafting the brief. Not a substitute for: a proper test with enough spend per variant, your platform's own attribution, or the judgement of the people who know the audience. Treat the Researcher's list as a proposal for the team to argue with, which is why it will land on the board rather than in a private chat.

FAQ

Can 8frame read my ad platform report today?

Not yet. Documents on the board and the Researcher agent are coming soon. Today, put the key numbers in a comment beside each ad.

Will the Researcher change my ads or generate new ones on its own?

No. It is planned to read, comment and propose. Anything that spends credits will wait for someone on the board to approve it.

Which files will the Researcher read?

The documents wired into it, such as a PDF report or a CSV export, plus any ads you hand it as a selection. Its node will list what it read.

Who sees the Researcher's findings?

Everyone on the board. Findings will appear in the board chat and on the board in the agent's frame.

Is there a release date?

No date is announced. This page will be updated when the features are live.

Sources


Try it without signing up: the 8frame canvas opens on a real board. The canvas is free and unlimited; generation is paid from $19/month.

Written with AI assistance. Corrections are noted on the page. Editorial policy

Related articles

use caseRolling One Approved Visual Out to Every Market and Format with an AI Agent (Coming Soon to 8frame)use caseBriefing AI with the Same Context as Your Team: From Private Chats to One Shared Board (Coming Soon to 8frame)use caseAI Agents as Teammates on a Creative Canvas (Coming Soon to 8frame)

Make it
move.

Stay in the loop

Be the first to hear about our launch and get product updates