by: Brinda Gulati

I’m Dr. Frankenstein, and we’re going to summon a good brief into existence with my non-doctorly methods.

I won’t lie, I’ve gotten to the stage with a lot of my clients where I get a keyword, and the brief builds itself in front of me, effortlessly, like rungs of a ladder appearing beneath my feet each time I fear I may fall off. That comes with familiarity and practice, though.

I’m also in a very different position when I receive a brief as a writer, versus when I make the brief and edit the resulting draft. I’m going to talk about the latter, that is:

How much can (or should) you put into a brief before you’ve left the writer nothing interesting to do?

If I’ve made my angle feel compulsory, a writer may leave out the very evidence that would have strengthened the angle or argument. Then I’ve taught someone to make my own blind spots more convincing.

And if you’ve worked with good editors before — any good teacher, really — you know that there’s a difference between a prescription and a proposition.

Let’s start with headings.

If I see one more heading that says “Why X Actually Matters in 2026,” I’m going to go up in flames.

So where do you find good headings? You reverse-engineer the search engine results pages (SERPs). That’s how I make my briefs for the writers I train.

You don’t need a software subscription for any of this, by the way—we’re going in without gloves.

Go to the People Also Ask (PAA) box on Google. Keep clicking and clicking and clicking the drop-down arrows until you veer away from the intent altogether. Then look at the first question again.

How far off was the top PAA question vs. the last one? That’s a gap you could potentially close. And equally, wandering away from intent is the cue to center yourself and get selecting.

From an initial exam of just this box, mind you, you can make a skeleton:

Across 15 questions, we know people are interested in the definition, cost, types, examples, and implementation.

That’s it — that’s your outline right here:

• What is an AI guardrail?
• Is there any AI without guardrails?
• What is the purpose of guardrails for AI?
• The three jobs AI guardrails do
• How do AI guardrails work?
• How to implement AI guardrails
• What are the limitations of AI guardrails?

The headings by themselves might seem nebbish, I know, but we’re talking about commercial writing here; and so I save the truly new information—what AI folks now call information gain—safely nested under the headings. The job is for someone to find us first, and then be blown away.

And for the headings you can’t accommodate in a measly 2,000 words, there’s always the FAQ section.

• Why does ChatGPT have guardrails?
• What are the different types of AI guardrails?
• How much does a guardrail cost?

On the topic of headings, let AlsoAsked and Reddit guide you toward relevant FAQs that show search intent. Also Asked, in particular, makes some very satisfying branches:

I see an innominate question here about AI safety in general—that’s a good opening to go hunting for privacy regulations and monetary consequences of non-compliance. The European Union is always a safe start.

Give the problem a face, I say, whenever you can. But we probably do not need Stephen Hawking’s warning about aliens in a piece on AI guardrails.

Now for the introduction.

Make or break, if you ask me. I don’t remember a single SaaS post I’ve read if it bored me within the first 150 words. You’re allowed to build a runway to set up tension, but the name of the game is keeping it ARS: accessible, relatable, and specific.

I give directions to my writers to the tune of something like this:

“Use the PAS framework to set up the tension in the first two lines.

What’s the problem? If we have a case study, use it. If you can find an attention-grabbing SME quote, use it.

How is it agitated? Is there evidence of real-life consequences? What’s the numeric cost of getting it wrong?

Do we have a solution? Yes, we do, but let the reader arrive at the conclusion by themselves. No hard plugs yet.”

And now that we’re talking about meat, how much direction is helpful vs. harmful in the body of the brief?

In my opinion and experience, this depends on the topic and the writer. And by writer, I just don’t mean their seniority; I also mean their familiarity with the product and/or the niche.

Once you know both, set the parameters for yourself and the brief accordingly; because while you don’t want to give away too much and basically write the piece for them, you *do* want this one piece to fit in the larger puzzle of the product that you are privy to, but they, in their limited access roles, might not be.

I find BoFU content easier to brief because a lot depends on G2 reviews, product walkthroughs, interface screenshots, and feature lists. That’s largely templatable with a reliable content partner, especially if you have someone who’s willing to go an inch wide and a mile deep.

But ToFU content is notoriously hard to get right. You risk being a summary of a summary, a watered-down version of an IBM glossary page, or simply—and this may be the worst of all—more boring than watching concrete cure.

Below is an example of a brief I did for Userpilot on a company that’s pretty well known, but shall remain nameless.

I think by asking questions in the brief, I give my writers the freedom to answer as they please—so long as it’s coherent.

If I find a good, data-rich source, I give them the raw information within the context, and the rest is up to them. Most of the time, they do well.

You’ll notice I’m quite bossy about certain details—I found that statistic for a reason, scouring the bowels of the internet—but how the writer turns them into a lucid argument is up to them.

They might find a better source. They might discover that my proposed angle collapses on contact with the evidence. I would quite like to know that before publication.

But ultimately, I wish for a writer to come back with something I couldn’t have written myself; it would be a shame to brief that possibility out of existence. ■

---

Brinda Gulati is a fractional content marketer who writes easy-to-understand, informative content for humans. She has two degrees in Creative Writing from the University of Warwick, and believes that, above all, stories are a deeply human endeavor. That’s why Shopify, Jotform, Whatagraph, Userpilot, Dock, and Dash trust her with their content.