AI + PR
How AI is reshaping PR workflows from monitoring to pitch personalization
In August 2025, I did everything by hand. Finding journalist contacts meant clicking through publication websites, copying names into a spreadsheet, and Googling for email addresses one at a time. A single media list took an entire day.
In This Article
Key Takeaway
AI did not replace my PR work. It eliminated most of the 23 hours a week I was losing to manual spreadsheet tasks, freeing me up for what actually matters: relationships and stories. Claude Code, custom scripts, and smart automation turned a solo operation into something that competes with agencies five times the size.
Six months later, the same processes take minutes — not because I hired a team, but because I started using AI as a coding partner in my terminal.
The PR industry is full of vague “AI-powered” claims that never explain what the AI actually does. What follows is the specific, unglamorous list of what changed in my workflow, and what it means for the work our clients get.
What I Was Doing Manually (August 2025)
A real snapshot of my weekly workflow, six months ago:
- Journalist research: Visiting publication websites one by one, clicking through author pages, copy-pasting names and beats into a Google Sheet. About 8 hours a week.
- Email finding: Searching LinkedIn, Twitter bios, and personal websites for journalist emails. Trying different email patterns and verifying them by hand. About 4 hours a week.
- Data analysis: Downloading CSVs, opening them in Excel, writing formulas, manually checking for errors. About 6 hours a week.
- Contact deduplication: Comparing spreadsheets side by side, hunting for duplicate names across sources. Excruciating. About 3 hours a week.
- Campaign tracking: Updating a spreadsheet every time a placement appeared, checking for links, noting domain ratings. About 2 hours a week.
Total: roughly 23 hours a week on operational tasks that had nothing to do with strategy, ideation, or journalist relationships. More than half my working week, gone before the real PR work even started.
What Changed: AI as a Coding Partner
I started using Claude, Anthropic’s AI assistant, directly in my terminal, not as a chatbot for writing emails but as a coding partner that helps me build and run automation scripts.
The distinction that matters: I am not using AI to write pitches or generate content. I am using it to build tools that make the human work faster and better.
Example 1: Journalist Database Building
Before (August): Manually visiting 60 publication websites, finding author pages, copying names. About 3 full days.
Now: I describe what I need in plain language, something like, “scrape the sitemaps of these 60 publications and extract all author bylines,” and the AI writes a script that does it in one run. 715 journalist records extracted in under an hour, including names, publications, and article counts.
The script catches edge cases I would have missed by hand: different sitemap formats, duplicate names across publications, Unicode characters in international bylines.
Example 2: Email Pattern Engineering
Before: Guessing each publication’s email format by trial and error: john.smith@, then j.smith@, then johnsmith@, for every new domain.
Now: We built a pattern map covering 265 publication domains. Add a new journalist and the system generates the most likely email address automatically, based on known patterns at their publication. The AI helped write both the pattern-matching logic and the validation pipeline. What used to take 20 minutes per journalist now takes seconds.
Example 3: Competitor Backlink Analysis
Before: Downloading Ahrefs exports, opening them in Excel, manually scanning referring pages for journalist names, copying them into my contact list.
Now: Export from Ahrefs, run a script. It extracts journalist names from article bylines, cross-references them against our existing database, scores new contacts by publication authority and relevance, and outputs a prioritized list. Processing 35 competitor domains went from a week-long project to an afternoon.
Example 4: Dashboard Building
Before: Everything tracked in scattered Google Sheets. No central view, no way to see which journalists had been pitched, which had responded, which placements had landed across campaigns.
Now: We run a connected ecosystem of dashboards: a Pitch CRM for journalist relationships, a campaign tracker, a content pipeline, and a financial overview, all built with AI assistance. They read from the same data sources and sync automatically. Log a placement in the CRM and it shows up in the client results dashboard and the financial tracker at the same time.
Building this from scratch without an AI coding partner would have meant hiring a developer. Instead I described what I needed and we built it together, iterating in real time.
What AI Does NOT Do in Our Workflow
Further Reading
This part is important, so let me be explicit:
- AI does not write our pitches. Every pitch is written by a human. The personalization, the tone, and the judgment call on which angle to lead with are all human work. AI cannot read the nuance of a journalist’s recent coverage or the timing of a news cycle.
- AI does not choose our story angles. Ideation requires knowing what is trending, what carries emotional weight, what will actually surprise people. That is human intuition built on experience.
- AI does not manage journalist relationships. When a journalist replies, a human reads it, understands the context, and responds. Relationships run on trust, not automation.
- AI does not make editorial decisions. Which data points to highlight, which publications to target, and whether a campaign is ready to pitch are judgment calls that need industry knowledge.
That is the dividing line in our shop: AI handles the plumbing, humans handle the relationships, and that division is not changing.
The Real Impact: Time and Quality
The honest before and after:
| Task | August 2025 | Now (2026) |
|---|---|---|
| Build a 50-person media list | 4 to 6 hours | 20 to 30 minutes |
| Find emails for 100 journalists | 8 to 10 hours | Under 1 hour |
| Analyze a competitor’s backlink profile | 3 to 4 hours | 15 minutes |
| Deduplicate contacts across sources | 2 to 3 hours per merge | 5 minutes per merge |
| Generate a campaign performance report | 1 to 2 hours | Automatic, real-time |
The 23 hours a week of grunt work from the breakdown above is now down to under 5. But the time saving is only half the story — the quality went up too:
- Fewer errors. Scripts do not accidentally skip a spreadsheet row or mistype an email address.
- Better coverage. Automated scraping finds contacts a human would miss after getting tired on page 40 of a sitemap.
- Faster reaction times. When a story is trending, we build a targeted media list in 30 minutes instead of half a day. That speed difference can be the gap between landing a placement and missing the news cycle entirely.
Why I Am Sharing This
Two reasons.
First, transparency builds trust. If you are considering hiring a PR agency, you deserve to know how they actually work, not marketing speak about “proprietary technology,” but real specifics about what the process looks like.
Second, this is the future of small agencies. A one-person operation with AI-assisted tooling can now match the operational capacity of a much larger team. That does not make the team obsolete. It makes the solo practitioner viable in a way that was not possible two years ago.
I am one person. I built a journalist database of 27,000+ contacts, a suite of connected dashboards, an automated enrichment pipeline, and a competitive intelligence system. Not because I am a developer — I am not. Because I have a clear picture of what I need and an AI partner that turns that into working code.
The PR work itself (the ideation, the storytelling, the journalist relationships) is still entirely human. The infrastructure around it is where AI changed everything.
What This Means for Our Clients
Lower overhead, better data, faster turnaround. We do not carry a team of 15 people and a $20,000/month retainer to match. Lean operations and sharp tooling mean our infrastructure costs are a fraction of a traditional agency’s, and our pricing reflects that.
That is why our PR Power Pack costs $3,000 instead of $10,000. Not because we cut corners on the work. Because we cut the operational bloat that inflates most agency pricing.
See the difference lean operations make. Same quality placements. Fraction of the cost. Let’s talk.
Frequently Asked Questions
Can AI replace PR professionals?
No. AI is terrible at the parts of PR that matter most: building genuine journalist relationships, reading cultural nuance, crafting angles that resonate with specific audiences. What AI excels at is eliminating the manual grunt work: data cleaning, contact enrichment, pattern analysis, report generation.
What AI tools are useful for PR?
Claude Code (terminal-based AI) for data processing and automation scripts, ChatGPT for brainstorming angles, and various free tools for email verification and data enrichment. The biggest gains come not from AI writing your pitches but from AI handling your data infrastructure.
How much time does AI save in PR workflows?
In our experience, AI cut data processing and research tasks from about 23 hours a week to under 5. That reclaimed time now goes into relationship building, pitch crafting, and strategic thinking: the work that actually produces results.
You Might Also Like
About the Author
Salvador Jovells
Founder of Presslei. 12+ years in ecommerce SEO across international markets. After a decade of link buying for Hockerty and Sumissura, I reverse-engineered 5,272 earned media placements and founded a reactive PR agency that builds authority through data-driven stories journalists actually want to publish. Based in Zurich.
