Writing for People and Machines at the Same Time

Answer the question in the first sentence, says every AI search guide. Sound like nobody else, says every brand guide. The tension is real, but it is not spread evenly across your website.

By Bruna Sotte de Carvalho, published August 24, 2026, 6 minute read

Every guide to AI search gives the same first instruction: answer the question in the opening sentence. Every guide to brand gives the opposite one: sound like nobody else in your category.

Those two instructions pull against each other, and most teams resolve the tension by quietly abandoning one of them. They either publish flat, extractable copy that could belong to any competitor, or they keep the personality and watch machines summarize their page into something generic anyway.

There is a better resolution, and it starts with admitting that the tension is not evenly spread across your website.

The backdrop is the shift described in how AI changed the way we search the internet: people now ask complete questions and expect a complete answer. This piece is about what that does to the writing itself.

Not Every Page Has the Same Job

The mistake is treating one rule as though it governs everything you publish.

Some pages are utility. Pricing, service areas, hours, specifications, how a process works, what is included. People arrive at those pages with a question and leave the moment it is answered. Machines quote them constantly. Personality on a utility page mostly gets in the way.

Other pages are persuasion. The manifesto, the founder story, the case study, the point of view piece. Nobody arrives at those pages needing a fact. They arrive deciding whether to trust you.

Sort your pages into those two piles before you argue about tone. Most of the friction in these conversations comes from two people picturing different pages.

Answer First, Then Earn the Rest

For the utility pile, borrow the structure journalists have used for a century. Lead with the answer. Follow with the detail. End with the context.

In practice that means the first paragraph under a heading states the conclusion plainly, in a sentence that survives being lifted out of the page and read on its own. No throat clearing, no scene setting, no promise that you will explain shortly.

This is the part that makes writers uncomfortable, because it feels like giving away the ending. It is not. It is what earns the rest of the page a reader, and it is what makes the page quotable. Ralph's guide to Answer Engine Optimization covers the structural mechanics in detail.

Voice Does Not Live in the First Sentence

Here is the reframe that resolves most of the argument. Brand voice was never really carried by opening lines. It lives in choices that a direct answer does not prevent you from making:

Which detail you choose. Two companies can answer the same question accurately and pick completely different specifics to include.
What you are willing to say plainly. Naming a tradeoff, a limitation, or a case where a customer should choose someone else is voice, and it is also the kind of statement that gets cited.
The examples you reach for. An Orlando reference, a specific customer situation, a number from your own work.
What you refuse to say. Every category has phrases everyone uses. Declining them is a voice decision.
Rhythm past the opening. Sentence length, the shape of a paragraph, where you stop.

A brand voice framework that survives this shift describes those choices, rather than listing three adjectives and hoping. If yours is still a list of adjectives, the brand voice charter is the template to start from.

Specificity Is Where the Two Goals Overlap

If you only take one idea from this piece, take this one.

Specific writing is simultaneously the most quotable and the most distinctive. "We respond quickly" is generic and unquotable. "We answer emergency calls within two hours across Orange and Seminole counties" is both a differentiator and a sentence a machine can lift with confidence.

Vagueness is the actual enemy, not directness. Most copy that people defend as brand voice is not distinctive at all. It is just indirect.

What Gets Cut

Some habits do not survive this transition, and they are worth naming because they are common:

Openers that describe the state of the world before arriving at the point.
Metaphor-first leads that only make sense after the explanation that follows them.
Superlatives with nothing behind them. A machine cannot verify "industry leading" and a reader stopped believing it years ago.
Burying the price, the timeline, or the qualification criteria several screens down.

None of that is voice. It is delay.

A Page Structure That Serves Both

Abstract advice about tone changes nothing on Monday morning. Here is the skeleton, for a utility page:

Heading: the question, in the customer's words. Not "Our Process" but "How long does a commercial floor coating take?"
First paragraph: the answer, complete and standalone. Three or four sentences. Include the number, the range, the condition. Assume it will be read with no other context, because it will be.
Second block: the qualification. What changes the answer. This is where the expertise shows and where most of the citation-worthy specifics live.
Third block: the proof. A real example, a number from your own work, a named constraint you have hit.
Last: the voice. The point of view, the recommendation, the reason to pick you. By now you have earned it, and a reader who got this far is the one worth persuading.

The persuasion pile inverts this. A case study or a founder story can open on the specific image, because the reader arrived willing to be led. Applying the utility rules to those pages is how brands end up sounding like a manual, and it is the most common overcorrection once a team starts writing for Answer Engine Optimization.

Test the Page Both Ways

Two checks, both quick, and a page should pass both before it ships.

The human test is unchanged. Read it aloud. If you would not say it to a customer standing in front of you, rewrite it.

The machine test is new. Paste the page into an assistant and ask it to summarize what the business does and who it is for. Run at page level here; the site-wide version of the same check is in how to audit what AI says about your business. If the summary is accurate but could describe any competitor, the page is not distinctive enough. If the summary is wrong, the page is not clear enough. Either way you have your revision, and the diagnosis took a minute.

Fold both checks into the editorial workflow rather than treating them as a launch ritual. The content ops calendar is the natural place to put them.

Key Takeaways

Sort pages into utility and persuasion before arguing about tone, because the rules differ.
On utility pages, lead with the answer in a sentence that survives being quoted alone.
Voice lives in chosen detail, plain admissions, and real examples, not in the opening line.
Specific writing is both the most quotable and the most distinctive, which makes vagueness the real problem.
Test every page twice: read it aloud, then ask an assistant to summarize it.
The search shift behind all of this is covered in how AI changed the way we search the internet, with more in chapter sessions on our events calendar and the resource library.

Topics: content strategy, brand voice, aeo, ai search, copywriting