The Earned Media Round Up agent runs a full earned and organic media audit — editorial coverage, social conversation, screenshot evidence, and a 16-slide PDF that reads like a PR story, not a data dump.
It grounds the audit in trade press first to make sure it knows what actually ran, pulls editorial coverage from media intelligence platforms (online news, print, broadcast, podcasts) and social conversation from listening tools (X, TikTok, Reddit, YouTube, blogs), and captures article and post screenshots as visual evidence. Before generating the PDF it applies an internal PR Expert review — if the narrative is thin or visual evidence sparse, it revises before shipping.
Before you start
Have your media intelligence connection set up for the brand
Know the campaign window (launch week, last month, full year) and any key talent or partners involved
If competitor benchmarking matters, make sure your competitor list is up to date
How to run an audit
Step 1: Open the agent
From the sidebar, open Agent Library and search for Earned Media Round Up, or go directly to app.boldstream.ai/run?agent=earned-media-round-up. Select your client and click Run Agent.
Step 2: Pick a pathway
The agent opens with three options:
Express (15–20 min) — brand-level audit, last 30 days, coverage scorecard + hero pieces + PDF
Structured (45–60 min) — full 6-phase workflow with PR Expert self-check before the deck
Conversational — brief the moment first and narrow scope before pulling data
Step 3: Confirm scope
The agent surfaces working brand context (category, core territory, visual system, social handles) and an intelligent competitor frame for sense-checking. Then it asks for the audit scope via a structured wizard: window, brand-level vs campaign, key talent/partners/competitors, markets and languages.
One important behaviour: the agent pauses before running trade-press grounding until scope is confirmed. A brand-level audit and a named-campaign audit produce very different searches and final narratives, so it asks first.
Step 4: Watch the data come in
Once scope is locked, the agent runs through:
Coverage criteria design and validation for editorial and social capture
Cost gate — if social conversation is under threshold, treats it as a campaign-landing insight rather than a data gap
Full data capture — editorial and social, with proper pagination
Visual evidence — minimum 12 screenshots for any single-brand campaign report (hero pieces at full bleed, tier-1 articles as tiles, social grids, posts, newsroom page, talent profile)
Step 5: Read the analysis
The audit scores coverage on five dimensions: Reach, Sentiment, Tier-1 density, Message cut-through, Social resonance. More importantly, it explicitly names:
The angle — the single sharpest story hook the coverage hung on (in one sentence)
Cut-through evidence — where coverage broke past brand-owned amplification into influencer pickup, analyst commentary, or industry-narrative territory
Step 6: PR Expert review
Before generating the PDF, the agent adopts the persona of a senior comms director and scores the draft on narrative arc, PR framing, executive summary as edit, hero-mention density, visual evidence density, voice, and methodology discipline. The deck only ships once it passes.
Step 7: Get the PDF
The final deliverable is a 16-slide PDF: title + executive summary + 3 hero pieces + tier-1 pickup + the angle + cut-through evidence + social activation + sentiment + timeline + observations + methodology. Story leads, methodology stays at the back, every data slide is paired with a screenshot or pull-quote.
💡 Tip: When a campaign's social volume comes back low, the agent treats it as a finding ("conversation did not pick the campaign up under that combination"), not a data gap. That's often more useful than a noisy hashtag spike — it tells you the partnership didn't bridge brand and talent in social media's eyes.
What's next?
Re-run the audit at the end of major campaigns, or set up a scheduled routine for monthly brand health checks. Over time you build a defensible earned media record that reads like PR Week rather than a generic dashboard.