MARKETING
Score Creative Variants with a Vision Model and Shortlist the Top Three
Reads pending creative variants from Airtable, scores each image on brand fit and clarity with an OpenAI vision model, then flags the top three as the campaign shortlist.
How it runs
The automated pipeline, trigger to output.
- TriggerScheduled run over pending variants
- ActionFetch unscored variants from AirtableAirtable
- ActionScore each image with OpenAI vision rubricOpenAI
- LogicRank scores and select top three
- OutputWrite scores, rationales, and shortlist flags to AirtableAirtable
What it does
Applies a consistent rubric to a batch of candidate creatives. It pulls every unscored variant from Airtable, asks a vision model to rate each on brand alignment, message clarity, and visual stopping power, then writes the scores back and marks the three highest as the shortlist.
When to use it
Run this after you have generated a pile of variants and need an objective first pass before human eyes. It replaces the ad-hoc gut-check with a repeatable, documented score so the shortlist is defensible.
How it works
- 1A schedule kicks off (or you trigger it manually) once variants exist in Airtable.
- 2The flow fetches all Airtable rows with an empty score column.
- 3For each variant, an OpenAI vision call rates the image against the rubric and returns a numeric score plus a one-line rationale.
- 4A logic step ranks all scored variants and tags the top three as Shortlisted.
- 5Scores, rationales, and shortlist flags are written back to the Airtable rows.
Set it up
What you configure once, before turning it on.
- 1Connect AirtableBases, tables, views, automations.
- 2Connect OpenAIModels, embeddings, files.
- 3Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
- 4Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
- 5Test, then turn it onRun once against a sample, confirm the output, then enable the trigger.
More Marketing workflows
Agentic UTM Taxonomy Steward: Detect, Explain, and Propose Fixes
An agent reviews the full backlog of campaign links in Airtable, reasons about taxonomy drift beyond rigid rules, drafts standardized replacements for malformed UTMs.
Slack UTM Check: Paste a Link, Get an Instant Verdict
Lets marketers drop a tracking URL into a Slack channel and get an immediate threaded reply that says whether the UTMs are valid, what is wrong, and a corrected suggestion.
Backlink Outreach Drafter for Verified Mentions
Picks verified, high-authority unlinked mentions from Airtable, drafts a personalized backlink-request email for each.
Source-Authority Verifier for Mention Provenance
Takes a queue of detected brand mentions and runs an agent that verifies whether each source is a legitimate, authoritative publisher versus a scraper or syndication copy.
Unlinked Brand-Mention Detector with Authority Scoring
Scans the web for fresh brand mentions, filters to ones that name you without linking back, scores the publishing domain's authority.
Weekly Brand-Mention Provenance Digest to Notion
Aggregates the week's tracked mentions from Airtable into a single summarized report covering new opportunities, authority distribution, and outreach progress.
Run it inside a business
This workflow drops into a full company template. Import the org, and this is one of the playbooks its agents run.
Run this workflow in your colony.
14-day trial. No DevOps. No Sales call. Provisioned in under a minute.