As of 13 August 2026, AI can only partly create social media analytics reports.

What to actually do

1. Hand it to a person

   The route this page recommends

   A person who owns the outcome does this end to end, worth it when the failure is dear.

2. Use a tool built for this

   Second choice

3. Do it yourself

   The distant third

   A chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.

   How to actually do it

   1. Open the analytics areas for each relevant social platform and export the same reporting period and comparison period, retaining the metric names and definitions.
   2. Create one folder or workbook containing the exports, the reporting objective, campaign names, business targets and any known tracking limitations.
   3. Remove passwords, private messages, personal contact details and unnecessary customer-level identifiers before pasting or uploading the material to an AI tool.
   4. Paste the supplied prompt into a chatbot, replace each bracketed slot with your objective and context, and add the labelled exports or tables.
   5. Ask the model to produce the report and calculation notes, then compare every total, rate and period change against the original platform exports.
   6. Ask a marketing colleague to challenge the explanations and recommendations, especially any claim about causation, audience quality or budget allocation.
   7. Move only the checked findings into your reporting template, label estimates or missing data clearly, and send the final report to its intended audience.

   Prompt

   Create a social media analytics report from the data below. Treat the data as the only source of truth and do not invent figures, dates, platforms, campaigns or explanations. First state the reporting period, platforms, account names if supplied, currencies if supplied, and any missing or inconsistent fields. Then produce: 1) an executive summary, 2) a platform-by-platform table of reach, impressions, engagement, engagement rate, clicks, video views, followers and conversions where those fields exist, 3) period-on-period changes only where both periods are supplied, 4) the strongest and weakest content or campaigns based on the available metrics, 5) three evidence-based observations, clearly separated from interpretation, 6) practical recommendations tied to the evidence, with the expected objective and metric for each, and 7) a list of questions that require a human decision or more data. Show the calculation for every percentage or rate you calculate, use 'not supplied' where a value is missing, and flag definitions that differ between platforms. Do not claim that one channel caused a result unless the data proves causation. Write in plain British English for a marketing team. Use these inputs:
   
   Reporting objective: [objective]
   Reporting period: [period]
   Comparison period: [comparison period or none]
   Business context and targets: [context and targets]
   Data exports or pasted tables: [data]

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.