CONTENT CREATION

Read A/B test results and auto-generate a new variant round around the winner

On a schedule, pulls variant click-through data from Postgres, identifies the top performer.

CategoryContent Creation
Enginepaperclip
Difficultyadvanced
Triggerschedule
Steps6
Setup~25 min

How it runs

The automated pipeline, trigger to output.

  • TriggerWeekly schedule fires
  • ActionQuery prior-round variant performance from PostgresPostgreSQLPostgres
  • LogicAgent picks winner and decides next variations
  • ActionRender new variant round with ReplicateReplicateReplicate
  • ActionUpload new renders to R2CloudflareCloudflare R2
  • OutputPost new round and rationale to SlackSlack

What it does

Closes the A/B test loop. It reads how each image variant performed, picks the winner, and generates the next round of variants that build on what worked, instead of starting from scratch. The new round is rendered, stored, and posted back to the review board with notes on why each direction was chosen.

When to use it

Use it when you run continuous creative testing and want each round informed by the last. It suits growth teams iterating toward a high-performing creative over several weeks, where the judgment of what to vary next is the real work.

How it works

  1. 1A weekly schedule fires the workflow.
  2. 2It queries Postgres for per-variant impressions and click-through rates from the prior round.
  3. 3An agent reviews the results, identifies the winning treatment, and reasons about what to vary next.
  4. 4The agent writes a new set of variant prompts seeded from the winner.
  5. 5Replicate renders each new variant.
  6. 6Renders are uploaded to R2.
  7. 7The agent posts the new round and its rationale to a Slack channel for the creative lead.

Set it up

What you configure once, before turning it on.

  1. 1
    Connect PostgresAny Postgres URL — query, write, migrate.
  2. 2
    Connect ReplicateImage, video, and model inference.
  3. 3
    Connect Cloudflare R2Object storage, S3-compatible.
  4. 4
    Connect SlackChannels, DMs, threads, mentions.
  5. 5
    Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
  6. 6
    Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
  7. 7
    Test, then turn it onRun once against a sample, confirm the output, then enable the trigger.

Run this workflow in your colony.

14-day trial. No DevOps. No Sales call. Provisioned in under a minute.