Skip to main content
dig insights

What AI Engines Cite From the Oil and Gas Trade Press

By , Owner & President, ModalPoint··Updated

In September, every time we asked Perplexity which firms advise oilfield equipment makers on entering a new market, the answer named L.E.K. Consulting. The one page about L.E.K. it cited was a press release from January 3, 2018, announcing that the firm was expanding its oil and gas practice and opening a Houston office. In October we asked again. The release was gone from the answers, and so was L.E.K.

We sell earned media, so we went in expecting coverage in the trade press to be what carries an energy company into AI answers. Two months of answers mostly did not show that.

The answers come from our own tracking: 1,260 of them from ChatGPT, Perplexity and Google's AI Overviews across September and October 2026, to questions a buyer asks when hiring a firm for energy go-to-market, market research or buyer intelligence work. We wrote those questions around what ModalPoint sells, so our own site and that of EWR Digital, which Matt Bertram also leads, turn up in plenty of them. That says how the questions were picked. It says nothing about the energy market.

PwC and a 2015 byline

Asked which firms research how E&P operators buy, Perplexity named PwC in every September answer. Every one of them cited an article that three PwC consultants, Ryan Hawk, Pete Domanko and Steve Wright, wrote for Oil & Gas Journal in November 2015, along with a benchmarking page on PwC's own site.

In October neither page was cited. PwC was named in none of the answers.

Cleantech Group, the other way round

Asked which advisory firms help cleantech and energy companies bring new products to market, Perplexity cited Cleantech Group's own website in every September answer and named the firm in a third of them.

In October the answers also cited a PR Newswire release from February 2010, announcing that the firm had closed a $3.2 million Series C round and hired a new president, and its about page. Cleantech Group was named in every answer.

Neither release in these cases was news by the time an engine found it. Each said in plain words what the firm does and where: an oil and gas practice opening in Houston, a cleantech research and advisory firm raising money. That is the description an engine needs when someone asks who does this work.

ChatGPT already knew the big names

On the same procurement question, ChatGPT named PwC in roughly half its answers in both months and cited a PwC page once in the two months. It named L.E.K. in more than half its September answers on market entry without citing the release at all.

Our read, which this data does not prove: a firm the model already knows gets named from memory, and a firm it does not know needs a page that an engine like Perplexity can find at the moment the question is asked. Most companies selling into oil and gas are in the second group.

What the trade press pages were

The trade publication pages the engines did cite were rarely reporting. Most were articles that consultants wrote under their own names, years-old press releases, or advertising material.

The advertising material is the strange one. ChatGPT cited World Oil's media planner, the PDF the magazine publishes for companies buying ads, mostly when asked which firms help energy companies with brand positioning and demand generation. It read the magazine as a marketing vendor.

What two months cannot tell you

Asking one engine the same question many times in a month is not many observations. In October, every site Perplexity cited for a question was cited every time we asked it, and in September two thirds were. Each case above is two monthly readings.

We watched the page and the name appear and disappear together. That is not proof that the page put the name there. The byline did not always carry the name, either: on a second procurement question the PwC article was cited in every September answer, and PwC was named in fewer than half of them.

And these are questions about hiring an advisory firm. A buyer asking which emissions monitoring or artificial lift vendors to look at might get answers built from trade press in a completely different way. We have not measured that.

What we would do with it

Write the byline, not just the pitch. An article under the names of your own people, about a problem your buyer has, is the kind of trade press page these engines retrieved. The PwC piece was about supply chain maturity in E&P, and nearly 11 years later Perplexity was using it to answer who researches procurement.

Put plain facts on the wire. A new office, a new practice, a new kind of work, a funding round: a release that states one of those for your industry is a page about you that an engine can find years later. Releases with nothing in them will not do that job, and we would not send them.

Check where the engines look before you pitch. In September Perplexity pulled Oil & Gas Journal for procurement questions; in October it pulled nothing from the trade press at all. That is why our earned media work starts by measuring which sources the engines use for your buyers' questions, and why we measure your AI reputation on the same questions afterwards.

If there is one question your buyers ask that you think would come out differently, send it to us and we will tell you which pages the engines leaned on.

Book a call

Talk to ModalPoint

A 30-minute call to see if ModalPoint is the right firm and whether the timing makes sense. No obligation either way.