Social data is more accessible than ever. We released Insights at Buffer, and with a button, I can export my data very easily (I would’ve probably cried with joy to have this feature as a social media manager in 2014).
Now, in addition to all of our social data, we also have the power of AI at our fingertips. One way that I’ve been able to make my social strategy more data-informed is by creating a workflow where I gather all of the context about my recent posts via Insights → have my AI tool of choice analyze it to look for patterns and gaps → and then check that against my overall social goals to help me stay on track.
I’ve been tracking my LinkedIn metrics for a while, and I have specific goals, but I’d never connected the context I could gather from my own posts to having AI coach me toward the output I actually want until downloading my data became this easy.
So here’s the system I built and the exact prompts I used, so you can do the same (if you’d like).
1. Download your data in Buffer (it’s free)
Getting all of your social post data from Buffer is easier than you might think, and it’s free. It’s also an important step in this process because successful systems work best when they are tailored, so we need the context.
Get this directly from your Buffer account
From Buffer, navigate to Insights. From there, you’ll see an Export button in the top right corner. Make sure that you’ve selected the correct dates and then Export as CSV.
Or get this data via Buffer’s API or MCP
If you are already connected to Buffer’s API via your AI tool of choice, you can simply ask AI to get the data for you.
I connected my Buffer account to my Claude account and asked Claude to grab them for me. (Possible on all Buffer plans, even free.)
Here’s what I asked Claude (you can use any AI tool; Claude is just my preference):
Via Buffer, pull all of my posts on [social network] from the last [time period] — the caption, the date posted, and all of the associated analytics.
→ Here’s a full guide to Buffer’s API integrations and how to connect them, or watch the video below to get set up with Claude.
2. Build the analysis most useful to you
This is where it gets fun — and where you get to be specific. The focus here isn’t one perfect prompt; it’s deciding what you most want to understand about your own posting, then asking for exactly that. A few options for you:
Content pillars: what you post about most, and how the mix breaks downYour voice: your real writing style, sentence patterns, the words and phrases you lean onWhat is performing: your most- and least-engaged posts, and what they have in commonFormats: whether text, photo, video, or link posts perform differently for youHooks: which opening lines you use most often, which are getting engagement, and which aren’tTiming: the days and times you post most and whether that aligns with the best time to post dataConversion: whether the posts meant to drive something (subscribers, sales, calls) are getting traction
Pick the few that matter most to you and drop them into the prompt:
Pull all my [social network] posts from [time frame] with their text, dates, and metrics (reactions, comments, impressions, reach). Build me an artifact analyzing [the 3–4 things you picked above]. Don’t generalize from social media best practices — only use my data.
Two things mine surfaced that I wouldn’t have guessed: my pillars were lopsided (I post about systems and marketing constantly, and about career far less — even though that’s something I’ve highlighted as a content pillar for myself), and my most personal posts do the most work (a quick one about taking my birthday off work pulled 104 reactions and 30 comments).
With all of this context, it’s time to reflect.
A simple Keep, Start, Stop is always effective. Based on what you found, what do you want to Keep doing, what do you want to Start doing, and what do you want to Stop doing.
A few more pointed questions you could reflect on:
Is there a gap between what I want to post about and what I am posting about?Have I selected the right days and times to send my posts out to reach my audience?Are the formats I’m using performing well, or do I need to experiment?
3. Use this analysis + AI to have a social coach in your pocket
The analysis is a goldmine of context, and now we use it. If you have specific goals as a creator or as someone managing social profiles, you can use this analysis and your goals to get some really valuable coaching and advice on what to adjust in your strategy or what to post next.
Here’s what I said to Claude:
Here are my goals for [social network] this year: [your goals]. Use the analysis of what I’ve actually been posting that you’ve already conducted. Act as my content coach. Based on the analysis, what do I need to do differently to reach my goals?
To make the coaching sharper, I also handed Claude a few Buffer resources as context, so we know its advice comes from real guidance and not just vibes:
How to Grow on Social Media in 2026: A Data-Backed GuideYour Complete Guide to Social Media Marketing: Platforms, Strategy, and Tips for GrowthThe Best Time to Post on Social Media in 2026: Times for Every Major PlatformAlgorithm information for: LinkedIn, TikTok, Instagram, Facebook, X, YouTube Shorts, and Threads.
From there, you can access this coach as often as you’d like: I’d recommend at least monthly.
I set up a monthly Claude Cowork scheduled task to run this analysis fresh and compare it to my goals again to coach me on adjusting my content for the following month.
You could also set a reminder in your to-do list to check in as frequently as you need.
Fit this into your existing workflows for the highest chance of success.
If you take one thing from this workflow, let it be this: you already have the best dataset for getting better on social media through your own content. Next up, it’s about using it. I’d love to hear how it goes for you!
