AI Agent: Review Instagram Posts Agent – turn metrics into your next month’s plan
By Hannah McKernon, Head of Digital Marketing, Maybe*
Clear insights, sensible next steps, and zero guesswork.
You are staring at Instagram Insights again. Reach is up, saves are down, Stories feel “fine,” and everyone has an opinion about what to post next. You could spend half a day exporting metrics and pasting them into a spreadsheet, squinting at charts, and still end the meeting with, “Let’s just try more Reels.”
Here's what comes back when you ask Maybe* to run it.
You say in Slack, in Teams, or directly in Maybe*: @Maybe analyse our Instagram performance for a specific date range, tell us what worked, what didn’t, and what to do next. You include your account handle, the dates, your goal for the period, and an Insights export or a simple table of post metrics.
What comes back looks like a report you can actually use:
A clean performance summary for the period, with the signals that matter most, like reach, impressions, interactions, saves, and shares, plus how they shifted versus the previous period if you asked for a comparison.
A “Top posts” block that ranks your best performers and names the metric that drove each one, like saves rate for a carousel, shares for a tip post, or completion rate for a multi-frame format.
A patterns section that reads like a creative debrief, not a spreadsheet. It calls out repeatable moves across your winners, format, hook style, topic themes, length, creative approach, and the CTA that actually got people to act.
A next-month test plan with three experiments, each with a clear success metric. Not vague advice, a small set of bets you can run and measure.
Then it pauses at the right moment, if anything important is missing. You get a short checklist of what it needs, why it matters, and the minimum info to continue. You approve, fill the gaps, and it completes the analysis.
That took Maybe* about two minutes to produce once you had the data ready. It would have taken your afternoon, plus a lot of debating.
Platform routing, without the setup work
You do not have to hunt for the right agent, configure a workflow, or translate what you need into “tool language.”
You describe the outcome in plain English, in the channel you already work in. Maybe* routes it to the right AI Agent and brings the results back to the same place, ready to paste into a report, brief, or content plan.
If you want the output delivered somewhere specific, like a shared doc, a campaign planning channel, or a weekly performance thread, you can set that expectation once and keep it consistent.
Scheduling turns a one-off review into a system
Most teams do analysis when something goes wrong, or when someone remembers. A system reviews performance on purpose.
You can schedule this in natural language, like: run an Instagram performance review every Monday morning for the last 7 days, include top posts, Story drop-offs, and three actions for the week.
That is the difference between a tool and a system. A tool helps when you remember to use it. A system quietly keeps your marketing operating rhythm on track, even when the week gets busy.
On-brand, consistent, and governance-first
This agent is built to be dependable under scrutiny. It sticks to analysis and recommendations based on the evidence you provide. It normalises performance across formats, highlights meaningful signals like saves and shares, and turns patterns into decisions without inventing anything.
It also respects brand constraints. If you share your tone, compliance rules, content pillars, or campaign objectives, the recommendations stay aligned, so you get fewer “wait, can we even say that?” moments.
When it pauses, and why that is a good thing
Paused is not failure. It is the safety feature.
Review Instagram Posts pauses when it would otherwise have to guess, like when there is no date range, no analytics export or screenshots, unclear goals, or a benchmarking request without enough context.
When it pauses, you see what is missing, why it matters, and exactly what to provide next. You stay in control, and the output stays accurate.
What it will not do
It will not log into Instagram or bypass permissions.
It will not invent metrics, audience demographics, or benchmark numbers without a source.
It will not use spammy or manipulative growth tactics.
It will not guarantee performance outcomes.
It will not publish content on your behalf without your workflow and approvals.
What to include when you ask
Where you want the result to land, like a report, a content plan, or a weekly summary
Account handle and date range, plus whether you want a comparison period
Your primary KPI, like reach, engagement, clicks, leads, or sales
Your data source, Insights export, screenshots, or a pasted metrics table for posts, Reels, Stories, and ads if relevant
Any constraints, brand voice, compliance rules, content pillars, campaign objectives
Start with one task today.
Send a single request with your date range, goal, and Insights export, then ask for one outcome, like top posts and why, best posting times, or three experiments for next month. Run it once. See what comes back. Then decide whether you want that same on-brand speed waiting for you the next time someone says, “We need to know what’s working on Instagram, today.”
Most teams do not go back to the old way.
