Summary: AI tools are increasingly reading and summarizing earnings releases before human analysts. That means businesses should be redesigning releases with bulleted highlights, clear sections, bolded key numbers, and distinct guidance segments to enhance machine readability and narrative control. The companies getting this right are treating the earnings release as both a regulatory document and a piece of owned content competing for attention inside AI-generated answers.

Before a human analyst ever opens your latest earnings release, there’s a good chance an AI tool has already read it, summarized it, and started shaping how your quarter is described. That changes the job of writing an earnings release in ways most investor relations and PR teams haven’t fully reckoned with yet.

I recently conducted a review (with help from AI) of more than 65 Q1 2026 earnings releases from large NASDAQ- and NYSE-listed companies—banks, energy producers, industrials, tech, healthcare, consumer brands—looking specifically at how the format of these documents is evolving, not just the numbers inside them.

The patterns were hard to miss: the companies that get quoted, summarized, and cited most cleanly by AI tools (the right number attached to the right quarter, a CEO quote lifted intact instead of paraphrased into something slightly off, guidance kept distinct from actual results instead of blurred into one generic takeaway, message pull-through and tone aligned with reality, etc.), have quietly restructured their releases to be machine-scannable, not just human-readable.

For companies that have a small number of analysts that follow them, those AI-written reports set the narrative, often before the CFO and CEO even dial the phone for their first post-conference briefing call with a human analyst. While the numbers cited in the AI-generated summaries from traditional (pre-AI) earnings releases were overwhelmingly correct, the nuances of shifting strategies, one-time events, or underlying trends were often missed by the reports written by AI agents without any human intervention.

So, what should communicators do?

How Earnings Release Formats Have Already Changed

A few years ago, an earnings release was a wall of paragraphs: a lead quote, some prose about the quarter, and financial tables at the bottom. That structure is disappearing fast among sophisticated issuers.

In our review, most companies now open with a bulleted “highlights” block, break the body into clearly labeled sections, and insert a meaningful quote from their CEO or CFO that moves beyond the ceremonial quotes of the past. Increasingly, the remark from a senior executive is written to function as the key takeaway of the quarter, because AI summarizers tend to lean heavily on direct quotes. While, for most companies, that was always the goal too often edit by committee created quotes that were not as focused as the AI agents require.

Companies like Intel, UnitedHealth Group, and Morgan Stanley have leaned hardest into this new format, pairing highlights with compact, upfront data tables and tightly labeled sections that make it easy for a tool, or a reporter on deadline, to lift the exact fact they need. Other companies, who tend to issue legacy “notice” releases that just point to a linked PDF, are effectively invisible to most AI systems that only read what’s distributed on the wire.

It is important to note that Intel, United Healthcare, and Morgan Stanley might have made this shift for readability and convenience for human readers, not for AI. But these companies are yielding online dividends in the form of higher quality visibility in AI-generated reports.

Anecdotally, I also spoke with four financial analysts who cover multiple companies and all confessed to using AI tool to take a first pass at earnings announcements and help them with crafting questions to ask leadership and provide content concepts. The final reports are theirs, but AI is becoming an important influencer, if not partner in creating those reports.

Where Most Companies Are Still Behind the Earnings Release Curve

Here’s the part that should get IR and PR leaders’ attention: the basics—bullets, subheads, a usable quote—are becoming table stakes, adopted by the vast majority of large issuers. That means doing them well no longer differentiates you; it just keeps you from looking dated.

The real gap has to do with the format, and comes down to a few specific moves most companies still aren’t making:

  • A true “at-a-glance” summary card. More than just another bulleted highlights list, these summaries are separate, visually distinct callouts near the top that explain “what happened this quarter” in a single glance (much like an AI answer card or knowledge panel). This is one of the fastest growing trends because it might be the clearest ways to signal the information that AI tools and answer engines can use and cite. Often, this takes the form of multiple subheads below the headline and before the lead paragraph which makes most communicators cringe. As one CFO put it, “It looks like crap, but it’s what gets read.”
  • Bolded numbers inside the narrative, not just in bullets or tables. Many releases still bury their most important figures—such as margin expansion, backlog growth or a guidance revision—inside full sentences. This format forces an agent or LLM to parse the content to extract relevant data. Companies that bold those figures directly in body copy see more of their narrative pulled through into the AI report. This is one of the easiest changes available and has an outsized effect on whether your framing of a number gets surfaced, or a model reconstructs its own.
  • A guidance section that stands entirely on its own. By providing a separate guidance section that is not folded into the CEO quote or a closing paragraph. Keeping forecasts/predictions in their own clearly marked section, or “ring fenced” makes it unambiguous to both readers and machines what’s a projection versus a reported result. Significantly, our analysis showed it’s often the single data point most likely to get pulled directly into an AI-generated summary.

None of these moves is hard to execute once a team commits to them. What’s harder and where we spend most of our time with clients on the details:

  • Deciding which numbers earn that treatment
  • Wording a summary card so it reinforces your narrative instead of just repeating your headline
  • Sequencing all of it so the release still reads well for the human analysts and reporters who are, for now (and we hope forever), still very much in the loop.

Why Better Earnings Release Should Be an Strategic IR and PR Conversation

Reformatting earnings releases touches all aspects of your corporate story.

  • Narrative control: Whose words show up when someone asks an AI tool “how did [company] perform this quarter”
  • Analyst and investor perception: Whether your framing of guidance, margins, and segment performance get surfaced, or whether a model fills in the gaps itself
  • Earned media pickup: Is the release written in a way that both journalists and AI agents can easily scan the most salient points.

The companies getting this right are treating the earnings release as both a regulatory document and a piece of owned content competing for attention inside AI-generated answers. That requires the same disciplined thinking PR teams already bring to media relations and thought leadership, but applied to a document IR teams (or finance and legal) have historically owned.

It’s Time to Talk About Your Next Release

If you’re heading into your next earnings cycle or just leaving it and wondering whether your release is/was built for how it’s actually going to be read by analysts, by AI tools, by journalists skimming on deadline that’s a conversation worth having before the draft goes to legal for review, not after.

If you would like a more detailed best practices guide to help your future earnings releases, please email tdruart@piercom.com and we will send you a guide.