The traditional public relations firm is named after one or two people, and those people are the product. 

They know which section editor answers at eight in the evening. They know which story will run and which will die in a features meeting. What they sell is proximity, and proximity does not scale.

A PR firm can hire bright graduates to answer emails, but it cannot manufacture another founder. So, the trade has been priced on access rather than outcome. The client pays a retainer, waits, and often cannot tell whether the silence means the story was weak or the call was never placed.

Omar Hamdi spent close to seven years inside that world's supply side, presenting and publishing at BBC One Wales, BBC Radio 2, The Independent and elsewhere. He studied cognitive science at Leeds University School of Computing and has watched artificial intelligence since long before it was fashionable. 

His view is that nothing much changed in the theory between 2006 and now. What changed was that the compute arrived, and with it the tools.

He founded Pathos Communications in 2019 on a simple observation. The service a listed company chief executive takes for granted has never reached the sandwich shop, the consultant or the small engineering firm. That is not for want of appetite. 

Ask any owner whether they would like to be in the news, and the answer is yes. The professional apparatus that delivers it was built for someone else.

His analogy is the website. There was a period when a corporate site cost six figures and only serious companies had one. Templates arrived, the price collapsed, and now a business without a booking link looks unserious. 

He believes PR sits about where websites sat two decades ago, and that the tool which collapses the price this time is a language model.

Pathos floated on AIM in December and sells the inversion of the retainer. A client pays when the article appears in a named publication, and not before. A first piece costs $5,000, three cost $15,000, and returning clients can take a $949 annual membership that drops the price to $3,500. The risk that traditional agencies push onto the buyer sits instead with the seller.

Two tools carry the load. PathosMind reads more than 50,000 news sources, builds a profile of a prospect and drafts the angle before anyone picks up a phone. Pressella joins client calls, takes notes and proposes ideas in real time. 

Hamdi calls the method human-led and AI-fed. The point is not to remove people but to give ordinary people expensive capabilities. His best content executive had no prior PR experience, last worked in a juice bar and drove half a million dollars of revenue in a single month.

Early testing puts Pressella's success rate at seven times that of a human colleague in sales development work. Both tools are due to reach general availability in the first half of 2027, at which stage they become products sold outward rather than machinery running inward.

What Pathos has that a rival could not assemble is an institutional memory. Seven years of calls, emails and transcripts covering small business PR, a segment nobody else has worked at this volume. Synthetic data will not substitute for it, and the alternative is to start collecting in 2026 and wait. 

That's what Warren Buffett would refer to as a competitive moat.

The engagement with publishers is, shall we say, 'non-traditional'. It may make some 'fourth estate' purists a little queasy but also speaks to the direction of travel within media in this zero-click era.

Pathos submits work as a freelance journalist would, and an editor at the publication approves or rejects it. It then pays a licence fee, which covers the client's right to reproduce the article and the masthead in their own marketing.

Cavendish, the company's broker, describes the money as a contribution to editorial costs. The Pathos name appears nowhere on the piece. The company argues that editorial control stays where it belongs and that disclosure rules in the UK and US are respected.

Whatever one makes of that, the demand behind it is real and growing. Search traffic is draining from publishers as AI assistants answer questions without sending readers anywhere, which is why newsrooms are taking these calls from Pathos.

On the client side, the same shift has raised the value of being written about. An AI model, or AI-enabled search, cites what has been published about a company. It's no longer about the expensive on-page real estate sold by Google.

A study quoted by Cavendish found more than 80 per cent of links surfaced by AI tools come from earned media. Gartner expects global PR budgets to double by 2027, while the addressable audience is put at 400 small- and medium-sized businesses globally.

Having looked at the business model, it is worth considering how this translates to the profit and loss account and balance sheet.

The first interesting observation is around the pay-on-results, which invites a particular kind of customer failure. That's the client who approves the article, watches it publish and then declines to settle. Bad debts ran at 15 per cent to 25 per cent of revenue. The 2025 figures carry a $2.1million write-off, which is the reason a business with 75 per cent gross margins reported a loss.

The repair was straightforward (while the problem will be familiar to those working in a high-growth environment). 

A financial controller joined in April 2025 and a chief financial officer in July of the same year. Cards are taken and tested before publication, customers are screened at qualification, and commissions now follow cash rather than invoices. 

On contracts signed from the second quarter of 2025, bad debt runs at 3 per cent to 5 per cent. Cavendish assumes 8 per cent across this year.

The rest of the numbers more than pass muster. Revenue reached $13.1million in 2025, up 15 per cent, with adjusted EBITDA of $2.9million against $1.9million the previous year. Net cash was $6.2million.

Repeat revenue has moved from 20 per cent of the total in the first quarter of 2025 to 41 per cent of cash receipts a year later. July brought revenue above $1.8million, a record month, on new client sign-ups running 30 per cent higher after a sales reorganisation. 

The market is looking for $14million of revenue and $4million of adjusted EBITDA this year, with interims due in September. At the current run rate, the benchmark for the year looks eminently achievable.

The share price, down 21 per cent in the last month, belies that financial narrative. At 25p, the company is valued at just over £17million, of which £4.4million is cash. The operating business is therefore priced at just over four times the EBITDA it is expected to earn this year.

Cavendish keeps a 42p target. Brokers attribute the decline to one holder selling into an August market with nobody on the other side. The company has told investors it knows of 'no reason for it'.

The float raised £5million and most of it, as we have noted above, remains intact. That said, Hamdi is itching to deploy those funds in pursuit of growth opportunities.

Organic expansion is repetition of what has begun to work. A reorganised sales floor has lifted new client sign-ups by 30 per cent a month, and an APAC team is in place.

Spanish-language selling has been tested in Latin America, with conversations under way in mainland China. Alongside that sits product breadth. The same writers and the same tools now produce books, podcast appearances and television placements, so a client who has tired of articles has somewhere else to spend.

Acquisitions are the second lever, but Hamdi resists the word roll-up. His is not a traditional media buy-and-build model. Two small deals sit inside the business already: Thought Leadership PR in 2024 and PodcastWise in 2025. 

What he looks for is a company with a good product where the sales process can be supercharged, on the reasoning that outreach is one of the things Pathos has industrialised. He observes that several listed peers now trade at levels where the operating business is close to free.

Hamdi's back-of-the-envelope arithmetic observes that growth of 24 per cent a year for a decade carries revenue from $13million to $100million, with any deals on top. Less dull is where he wants it to happen. He has US resellers, US contacts and a US non-deal roadshow under discussion, and he could point the company at Nasdaq in time. His stated preference is to stay in London and follow the AIM companies that grew into the main market and the FTSE 250. He is not romantic about it. Value has to be realised somewhere, and he has said so himself.

Now for the cautionaries. I am telling the story as told by management and after poring over and interpreting public documentation.

The risks are the ones you would expect of a founder-led small-cap. Hamdi holds the majority of the shares and embodies the strategy. Around 90 per cent of clients are in North America, so a US downturn would arrive before any European or Asian offset.

Sales floors are harder to scale than slides suggest, and software due in 2027 has a habit of arriving in 2028. That said, Hamdi remains confident on the rollout, and the execution thus far has been flawless. So, the recent share price weakness, which has been hard for management to rationalise, may prove to be an opportunity for those new to this story.

For all the breaking tech and media news, go to www.techdefused.com