“Several retailers, including Northstar, operate in the area.”
FOR THE PEOPLE BEHIND THE PRESS COVERAGE
Great coverage.
Now tell the
whole story.
Your media-monitoring provider sends the reports. PR Iris brings those Excel files and coverage batches together, adds sentiment and article labels, and helps you turn all those mentions into a story your client can use.
Earned media · Share of voiceMedia relations · Client reporting
IRIS
01 / COVERAGEThe reports are in.Monitoring reports · Excel batches
02 / CONTEXTWhich mentions matter?Sentiment · article type · relevance
03 / DIRECTIONReady for the client.Charts · PowerPoint · next steps
THE MESSENGER BETWEEN DATA & DECISION
ONE CONNECTED REPORTING WORKFLOW
Follow a mention.
All the way to the client conversation.
Monitoring companies supply the Excel reports. PR Iris connects the batches, adds context, and helps your team decide what the coverage means. Choose a branch to follow the work.
THE PR IRIS MAPClick a node · follow the connections
01 / REPORTS ARRIVE
From your media-monitoring providers02 / BRING IT TOGETHER
03 / MAKE SENSE OF IT
04 / YOUR JUDGMENT
Need more evidence?Return to the question.
05 / TAKE IT FORWARD
Your team turns findings into recommendationsProvider reports → project context → analysis → review → outputsWorkflow illustration · no live client data
ZOOM IN / THE RELEVANCE BRANCH
A passing mention?
Or the main story?
Sentiment and relevance labels help you distinguish a passing reference from coverage worth discussing.
“Northstar’s new refill programme wins praise from shoppers.”
Earned media·Share of voice·Competitive coverage·Client-ready insights
EXPLORE THE COVERAGE
More coverage.
But is it better coverage?
That’s the kind of question a client will ask. Try switching between article volume and estimated reach below. The answer changes when you look a little closer.
COVERAGE EXPLORERILLUSTRATIVE DATA · NOT CLIENT RESULTS
SHARE OF VOICE / ARTICLE VOLUMEClick a brand to investigate
Current periodPrevious year
OBSERVATION
HYPOTHESIS
The publication mix or the topics covered may explain the difference. Validate against the underlying articles.
POSSIBLE ACTION
Review the outlets and themes driving the shift before changing the next media plan.
This is a scripted demonstration of the analysis approach, not a live connection to Media Insights.
MEET THE AI ASSISTANT
Ask about the whole picture.
Then keep going.
Once your monitoring reports are in the project, ask “What changed this quarter?” Follow up by brand, publication, or uploaded batch. The assistant works from that reporting context, and you can return to a saved conversation.
MEDIA INSIGHTSILLUSTRATIVE SCREEN
RETAIL / PROJECT FILES
A place for every batch.
Reports from your monitoring provider. Choose the batches for this question.
Coverage_Q3_2025.xlsxPrevious year · selected
✓Coverage_Q2_2026.xlsxAvailable in this project
○Coverage_Q3_2026.xlsxCurrent quarter · selected
✓↳
The files stay together.
The question sets the scope.
YOUR AI ASSISTANTSCRIPTED DEMO
Retail project · Q3 2025 + Q3 2026
IRIS / BASED ON THE SELECTED BATCHES
What is this answer based on?
TRY A FOLLOW-UP
Illustrative screens with fictional data, built to explain the workflow. These are not captures of the live application.
THE FAIR QUESTION
“Why not just drop
all the files into Claude?”
For a few files and a one-off question, that can work well. But a growing archive of monitoring reports needs more than a bigger chat window.
A ONE-OFF FILE CHAT
A useful place to start.
Uploading a file doesn’t mean every row stays in the model’s working context. Context windows are finite, and recall can weaken as the amount of text grows. A long conversation alone is not a dependable record of an entire reporting archive.
YOUR RECURRING PR WORKFLOW
The archive lives outside the chat
Monitoring batches and structured reporting data stay in the project database, beyond the model’s working context.
Analysis tools do the data work
The assistant uses project analysis tools and scoped results, rather than relying on the full archive being pasted into a prompt.
Context you can check
Select the files, brands, and periods behind a comparison. Keep sentiment, article type, and relevance labels available for repeated reporting.
Connected outputs
Move from AI questions to charts and Excel or PowerPoint reports.
Store the archive. Analyze the selected data. Give the AI the relevant evidence. That is the advantage—not a promise of unlimited AI memory or error-free answers.
Why context limits matter—and what still needs review
Claude and other AI products can use retrieval, code, and context management to work with larger datasets. The distinction here is the built-in PR data and reporting workflow, not an absolute limit on what another system can do. Read Anthropic’s explanation of context and recall ↗
Separate measured facts from hypotheses. Check the selected brands, files, and dates; review classifications and recommendations against the sources. Estimated reach is not unique audience, and coverage alone does not prove revenue.
Export scope: project and comparison exports are available in Excel and PowerPoint. Editorialized presentations from selected Chat exchanges remain on the roadmap.
MAKE THE BUSINESS CASE
What would you do
with those hours back?
Think about the hours spent combining monitoring reports, reviewing mentions, comparing periods, and preparing client decks. Put your own numbers in and see what less manual work could be worth.
START WITH A SCENARIO
Scenario inputs, not pricing or measured product results. Time saved represents capacity; it becomes cash savings only if actual costs fall. Human review remains part of the work.
YOUR ESTIMATED ANNUAL CAPACITY VALUE
Hours freed / month
Annual value after investment
Modeled net ROI
Break-even time reduction
Show the math
Monthly hours freed = reports × hours per report × time reduction.
Annual capacity value = monthly hours freed × hourly cost × 12.
Net value = capacity value − annual investment.
ROI = net value ÷ annual investment × 100.
UNDER THE HOOD
AI helps explain the story.
A real data system backs it up.
The Media Insights application combines stored project data, analysis tools, and an AI assistant. Here’s what each part does for your reporting workflow.
01 / ORGANIZE & ANALYZE
Python + FastAPI
The application backend handles requests and exposes the analysis tools the assistant can use.
pandasopenpyxlSQLAlchemy
Spreadsheet processing and database access.
02 / KEEP THE CONTEXT
PostgreSQL
A database behind the project workspace keeps reporting data available beyond an individual chat exchange.
Project dataUploaded batchesSaved sessions
Return to the work instead of starting from scratch.
03 / ASK & INTERPRET
n8n + DeepSeek
n8n orchestrates the AI workflows. DeepSeek supports classification and grounded Chat over the project's analysis tools.
ClassificationGrounded ChatAnalysis tools
Review the evidence before acting on a recommendation.
Interface Jinja2 + JavaScriptDeployment Docker on HetznerReporting outputs Excel + PowerPoint
This describes the Media Insights application. The interactive examples on this presentation page use fictional data and scripted responses.
WHAT A BETTER REPORTING DAY LOOKS LIKE
Less time building the report.
More confidence in the room.
Walk into the client review with more than a count of mentions: explain the sentiment, distinguish passing references from meaningful coverage, and show the story in charts and a presentation.
Win back thinking time.
Spend less of the week assembling coverage and more of it shaping the next media relations plan.
Explain the performance.
Put sentiment, share of voice, competitor movement, and the mix of media sources into a clear story your client can follow.
Make the next brief better.
Turn reviewed findings into questions, priorities, and recommendations for the next campaign.
The goal: less reporting friction, clearer client conversations, and better-informed PR decisions. Your team still brings the judgment. Sign in to Media Insights to work with your projects.