What Does a PR Agency Actually Do for a Startup?

A PR agency does seven jobs for a startup: it sharpens the story into something a reporter can use, builds a list of the right journalists, pitches them, writes announcement materials, prepares founders for interviews, handles bad news when it arrives, and reports on what landed. It does not sell your product, guarantee coverage, or fix a weak story. Good PR earns you third-party credibility in outlets your buyers, investors and future hires already trust.

That's the short version. The longer version matters because most founders hire PR without knowing what a normal month of work looks like, and then judge the result against the wrong yardstick.

I've spent six years placing Web3 and AI founders in outlets like Forbes, CoinDesk, Decrypt and AI Magazine, on both sides of the agency line. So I'll walk you through the actual deliverables, what you should do yourself, and how the agency model compares with a fractional senior operator.

What a PR agency does, deliverable by deliverable

Strip the jargon and a startup PR engagement is a production line. Each stage feeds the next. If one stage is weak, everything after it underperforms.

Narrative and messaging

This is the most underrated deliverable and the one that decides everything else. A reporter doesn't care that you raised money or shipped a feature. They care whether your news tells their readers something new about a trend they already cover.

A good agency will push back on your deck language. "AI-native workflow platform" means nothing to a journalist. "The first audit tool that catches smart contract bugs before deployment, now used by X teams" might.

Media lists and pitching

The list is where most of the hours go. A useful list is not a database export. It's 30 to 80 people who have written about your category in the last 90 days, with notes on what they cover and how they like to be pitched.

Pitching is the visible part. It's short emails, follow-ups, exclusives offered to one outlet at a time, and a lot of rejection. On a typical launch, a strong pitch gets a response from maybe 10 to 20 percent of a well-built list. A weak one gets close to zero.

Press materials

Press releases still exist and still have uses: wire distribution for SEO, a canonical source for reporters checking facts, and an artifact your investors can forward. But a release on its own rarely produces tier-1 coverage. It supports the pitch. It doesn't replace it.

Interview prep and crisis readiness

When a reporter says yes, the agency briefs the founder: who the reporter is, what they've written, likely questions, the three points to land. This is where a lot of earned coverage gets lost, because an unprepared founder rambles and the quote that runs is the weakest thing they said.

Crisis work is the part you hope never to need. For a seed-stage startup it's usually a two-page doc with holding statements and a clear rule about who speaks. Web3 teams need more, because hacks and token price moves happen in public.

Measurement

Expect a monthly report. Clip counts and "estimated reach" numbers are the weakest part of it. Better reports show which placements drove referral traffic, demo requests, investor inbound or new citations in AI search answers.

What a PR agency does not do

This list causes more disappointment than any other part of the relationship, so it's worth being blunt.

• It doesn't guarantee coverage. Any agency promising guaranteed tier-1 placement is either buying sponsored content or overselling.
• It doesn't generate leads directly. PR builds trust that makes your sales and marketing convert better.
• It doesn't fix a story with no news in it. If there's nothing new, the best pitcher alive won't place it.
• It doesn't run your social media, paid ads, or SEO, unless that's explicitly scoped and priced.
• It doesn't replace founder time. Expect 2 to 4 hours a week from the founder during active campaigns.
• It doesn't produce results in week one. Most first placements land in weeks 4 to 8.

What results realistically look like

Founders tend to anchor on a single headline in a dream outlet. That's a fine goal, but it's the wrong unit of measurement.

For an early-stage startup with a real announcement, a reasonable first-quarter outcome looks like this: one anchor story in a tier-1 or strong trade outlet, three to six supporting pieces in trade or regional press, one or two podcast or newsletter appearances, and a founder who now has a recognizable point of view on one topic.

What that unlocks is less glamorous than the clip itself. Sales decks get a "featured in" bar. Investor diligence turns up third-party coverage. Recruiters have something to send candidates. And increasingly, AI assistants like ChatGPT and Perplexity cite those articles when someone asks about your category.

I've seen what a well-sequenced campaign can do. When we ran the RARI Chain mainnet launch, the embargo lifted into 11 simultaneous tier-1 placements across CoinDesk, The Block, The Defiant, Cointelegraph and CryptoNews, followed by coordinated APAC translations. That outcome came from the boring stages being done properly: narrative locked early, list built carefully, embargo managed tightly. You can read the breakdown on the RARI Chain case study.

Agency vs fractional PR: how the structures differ

The deliverables above are roughly the same whoever does them. What changes is who actually does the work and how much of your budget goes to coordination.

Agencies make sense when you need volume: multiple markets at once, a big consumer launch, constant event support. The senior person sells the account, and a team executes. That's not a scam. It's just how the model funds itself.

A fractional operator suits companies that need senior judgment more than headcount. The person who builds your narrative is the same person writing to the reporter. If you want to see how that setup works in practice, I've laid out the model on the fractional head of PR page.

The DIY share: what you should keep in-house

Even with outside help, some work belongs to you. This is the split I'd use.

• You own the facts: metrics, customer names you're allowed to share, funding details.
• You own approvals: one decision-maker, 24-hour turnaround on drafts.
• You own the founder's voice: opinions, hot takes, the things only you can say.
• You own the calendar: launch dates, investor sign-offs, legal review windows.
• Your PR partner owns the story framing, the list, the pitch, and the relationships.
• Your PR partner owns timing: embargo windows, exclusives, sequencing.
• You both own measurement: agree on three metrics before the work starts.

The biggest single cause of failed campaigns I see isn't bad pitching. It's slow approvals. A reporter interested on Tuesday has moved on by Friday.

How to tell if an agency is doing the job

Within 30 days you should see a sharpened narrative document, a named media list you can actually read, and pitch drafts sent to you for review. By day 60 you should have at least a handful of reporter conversations, even if not all of them turn into stories.

If you're getting weekly status calls with no list, no drafts and vague "we're in conversations" updates, that's a warning sign. I wrote a fuller list of those in PR agency red flags, and if you're in the middle of choosing one, the PR agency evaluation scorecard gives you a structured way to compare proposals.

Ask every agency the same three questions:

1. Who on your team will write and send our pitches, and how many other accounts do they run?
2. Show me a media list you built for a company at our stage in our category (names redacted is fine).
3. What would you tell us is NOT newsworthy about our upcoming announcement?

The third question is the most revealing. An agency that says everything is newsworthy is telling you what you want to hear.

Who should hire a PR agency, and who shouldn't yet

PR earns its cost when you have news coming and a buyer or investor who reads the press. Funding announcements, product launches, category creation and enterprise sales cycles all benefit.

It's usually too early if you don't have a product in users' hands, can't name the reporters who cover your space, or have nothing to announce in the next 90 days. In that case, spend the money on product and founder-led content first. My overview of PR services explains which engagement types fit which stage.

A PR agency sells access, judgment and labor. The question is how much of each you're actually buying.

Not sure whether you need an agency, a fractional lead, or neither yet? Book a 30-minute teardown and I'll tell you straight.