Honeyjar launches an AI PR agent that pitches reporters, but says when you're not ready
The San Francisco startup's AI agent drafts pitches and builds media lists, but its real pitch is that it also knows when to hold one back.

Honeyjar, a San Francisco startup building an AI agent for public relations, showed its product to the public for the first time this month, appearing in TechCrunch's Startup Battlefield 200 at Disrupt 2026 and opening a waitlist that drew roughly 10,000 signups in its first 24 hours.
What the agent actually does
Honeyjar drafts announcements, pitch emails and social posts in a company's own voice, then goes looking for the reporters most likely to care about them. From there it works the rest of the job a PR hire would normally carry:
- Builds and maintains media lists
- Scans news and social feeds for relevant coverage
- Flags speaking and event opportunities
- Tracks how a pitch or story lands once it runs
The company trains the system through what it calls a Comms Lab, feeding it the judgment calls of working PR professionals rather than just their finished copy — which matters only once the tool starts deciding what not to send, not just what to write. That's the part Honeyjar advertises most: its founders say it's built to tell a founder when they aren't ready for coverage yet, and that a human still has to send or publish anything it drafts. Co-founder and CEO Michelle Masek put it this way:
We never want a world where the human is removed from the loop.
Honeyjar is one of roughly 200 companies TechCrunch selected for Startup Battlefield 200 at this year's Disrupt conference, a track built for early-stage startups rather than funded platforms with existing enterprise contracts. That puts it closer to the stage most software marketers actually shop at: a product you can try before it has a sales team, not one you need procurement to evaluate.
Who it's built for, and who's already testing it
Honeyjar raised $2 million in pre-seed funding in December, led by Heather Hartnett of Human Ventures and Kevin Mahaffey of SNR Ventures, and spent the months since building toward this public launch. The angel roster that joined that round skews toward people who've actually run comms and media operations rather than pure financial backers: former a16z CMO Margit Wennmachers, ex-TechCrunch editor Josh Constine, Project Runway creator and Full Picture founder Desiree Gruber, and former Outcast Agency CEO Becky Porter are all listed as investors. Mahaffey, explaining the bet he made, said Honeyjar "automates the essential tasks and workflows so more focus can go to the story, strategy, and stakeholders."
Masek founded the company after running marketing and communications at the venture firm Obvious Ventures; co-founder Nadia Jamshidi came out of agency and in-house comms work as well. The beta list TechCrunch named points at who Honeyjar is chasing first: neighborhood network Nextdoor, delivery robotics company Coco Robotics, venture firm Bloomberg Beta, and the agencies Mission North and Full Picture. None of those is an enterprise with a dozen-person comms department. That's the same profile as most software marketers who end up owning press outreach alongside demand gen, content and everything else, usually without the training or the reporter relationships a dedicated comms hire would bring in.
What a software marketer should do differently
If you're the one writing your own press releases between campaigns, a tool that drafts pitches and builds reporter lists solves a real time problem. The open question isn't whether it can write a passable pitch — most AI tools can — it's whether it knows when not to send one, and whether a reporter can tell the difference. That judgment is the entire premise Honeyjar is selling, and outside a few named beta accounts and a 24-hour waitlist spike, it's still unproven at the scale most software companies would actually use it.
It also lands at a moment when most software marketing teams already carry PR as a side responsibility rather than a staffed function. Masek frames that as the whole market opportunity, describing "millions and millions of unassigned communicators responsible for the work of comms." Whether that gap gets filled by a single AI agent or by better outsourcing to an agency is a real open question — but the fact that a well-connected angel list, including people who've run newsrooms and agencies themselves, is betting on the agent version is itself worth noting for anyone budgeting for PR help next year.
Treat it the way you'd treat any AI tool that touches a relationship you don't fully control, rather than one that only touches your own content:
- Run a small, low-stakes pitch through it first — a minor feature update, not your next funding announcement
- Check whether it actually declines to pitch something that isn't ready yet, not just whether it can write around the gap
- Ask a reporter you already know if a drafted pitch read as machine-written before you send a cold one
- Keep the final send-or-don't-send call with the person who owns the reporter relationship, not the software
If you already own PR at your company, the risk here isn't the drafting step — it's handing off who-to-pitch-and-when decisions to software that hasn't built your relationships with reporters yet. Pilot it on work you'd be comfortable seeing in print even if the judgment call turned out wrong.
Tools & AnalyticsSalesforce buys Loops to make Tableau explain why your numbers moved
Sienna Mcpherson · 4 min read
AIAnthropic bans Claude-powered content networks built to sway AI answers
Sienna Mcpherson · 4 min read
AI