We’re exploring how creatives are actually using AI at this snapshot in time, so we sat down with a senior copywriter to unpack the tools and outputs he is incorporating in his copy workflows. We uncovered four specific use cases and a thesis about senior judgment.
Brad Cook has twenty-plus years of marketing copy behind him. He anchors our Databricks account and works across non-tech clients too. He agreed to walk us through where AI shows up in his week, and where he keeps it out. This is the second piece in a series on how individual roles at C5 use AI in real work, following our conversation with ACD Kyle Light on his design workflow.
Net-net, Brad uses AI as an assistant for the cognitive overhead, and his twenty years of taste do the rest. His specific relationship to AI says something interesting about where craft sits in an AI-saturated industry. We found Brad’s perspective to be a refreshing support for our conviction that prioritizing strong talent is too-oft forgotten in the surface-level marketing AI conversation.
The Databricks brief: be credibly technical
Brad spends a big chunk of his day editing Databricks copy. Landing pages, ebooks, and demand-gen briefs arrive from the client already drafted, sometimes by AI. A senior editor then has to make them land. This is some of the most technical work Brad has done, and the audience is hard to win on marketing terms alone.
Why this audience is hard to win
Databricks sells data and AI infrastructure to engineers and data professionals. These buyers make real procurement decisions, and they read very differently from a marketing buyer. They research their way to a technical conclusion. First, they want to know whether the platform can actually do what the copy claims. They also check whether the API documentation matches the marketing pages. Then they ask their own team about the architecture before any of it reaches leadership. As a result, marketing copy has to clear a higher bar on two fronts at once.
The technical claims have to be accurate, so an engineer reading closely cannot find a hole. The voice also has to sound like someone who understands the product, rather than someone briefed the morning the page went live. Copy that hits one bar and misses the other reads as fluff. Moreover, a technical audience has learned to skim past fluff on the way to the documentation.
Brad’s job is to hold both lines at once. The copy has to survive engineer scrutiny, land with the broader buying committee, and clear the Databricks style guide. For a sense of what we ship into this environment, see our Databricks case study. That tightrope is where AI earns its keep, in four specific places.
Decoding technical nuance before he edits it
Brad came up as a writer, and Databricks sells to engineers and data professionals who live inside the gap between those worlds. So when he hits a feature description he does not fully understand, he stops. He asks Google’s AI Overviews to explain the underlying concept in plain language. He treats the output as comprehension support rather than writing support, and it goes nowhere near the final page. What it gives him is confidence. He can then edit aggressively, ask sharper questions in the client review, and land explanations that survive engineer scrutiny.
Brad has noticed a second-order effect here that is worth your attention as a marketer. He now lets AI read the sites that used to answer his questions and synthesize the answers for him. He rarely clicks through to read them himself anymore. That is a real shift in how research gets done at the senior level. It is also the same shift happening to your customers when they evaluate vendors. Gentle reminder: if your brand fails to show up in the AI’s synthesis, you lose audience you used to win on the third Google result.
Compressing dense quotes for event banners
Databricks throws a major annual event, and customer quotes go on banners with hard character counts. Most of those quotes are short enough that Brad trims them himself. Every so often, though, one arrives as a dense block of executive testimony that has to collapse into a single line. Even then, the final line still has to sound like the person it is attributed to.
Those blocks are where Claude earns a first pass. Brad also reaches for it when he gets stuck finessing a quote under the character count on his own. He drops in the original and asks for several versions trimmed to length, then treats the output as starting points rather than finished copy. He figures AI gets him about seventy-five percent of the way there on this kind of cognitive-load task. For the last twenty-five percent, he leans on the lessons his twenty years of copy experience have taught him.
Treating AI as the assistant he couldn’t otherwise afford
Brad’s framing for AI is one we had not heard quite this directly before. He thinks of it as the personal assistant he could never justify hiring on a per-project basis. It is someone to bounce a phrasing off, to explain a piece of dense technical writing, or to hand a messy first-pass cleanup. The assistant runs the errands so Brad can spend his attention on the calls that need judgment.
That posture differs from the “AI replaces writers” reduction that dominates so much industry conversation. In Brad’s version, AI is a contractor working under the principal’s direction. The senior copywriter keeps every meaningful decision, and the AI handles the prep work that used to eat his mornings.
He also uses AI for the occasional tool tutorial, when he gets stuck in Figma or another app he rarely touches. The volume is low, but the time saved versus digging through forum posts is real.
Editing AI output so it stops sounding like AI
A good chunk of the copy Brad sees in a given week was at least partly drafted by AI. The Databricks demand-gen team, for example, now uses AI to fill in landing-page copy on their briefs. So the question for a senior editor is how to make that output stop sounding like itself.
Brad’s answer is to edit by hand, the way he has edited for two decades. He keeps no AI detector in the loop and no checklist of telltale phrases. Instead, he uses his ear, tuned to what the Databricks audience reads as credible and what they read as marketing. Pattern-matching against AI tells is shallow work, and the patterns themselves are unreliable. As Brad put it:
“Em-dashes are something I’ve always used, and apparently now they’re a hallmark of AI. So maybe my writing’s always been a little AI.”
One defense against generic AI copy is to write with enough specific detail, earned authority, and point of view that no detector mistakes the work for generic. That is what you pay a senior copywriter for. The same instinct, applied systematically, is how you keep AI output on-brand across every surface where AI now produces content. Twenty years of practice builds that ear.
The reckoning: AI rewards senior judgment and removes entry-level work
Brad’s read on the moment is direct. AI has raised the value of senior people with experience and discernment. At the same time, it has started removing the entry-level work that produces those people in the first place.
That is the part worth sitting with if you run a creative team. The ladder still exists, but the rungs are getting further apart. Junior copywriters used to build judgment by doing the prep work that AI now absorbs. So the brands that keep building senior judgment, even as AI eats the bottom rungs, are the ones that will hold the talent in five years.
This is the same argument we made in our piece on brand drift, which looked at how easy AI access floods the market with middle-quality content. The brands that once leaned on senior craftspeople to hold the line feel that flood first. For the same reason, we tell clients to protect the senior voice as they grow, which is the core idea in our guide to scaling content without losing your voice. Brad’s vantage from inside the Databricks account adds another angle on how humans and AI actually share the work.
In practice, you want a human at the start of the pipeline and a human at the end, because that is where AI sits in the work right now. The middle is where the assistants live, and the ends are where the judgment lives. That judgment is the part your clients, and the engineers reading their way to a procurement decision, can still tell apart.
Find Brad’s work across our portfolio. And if you want C5’s senior copywriters involved with your brand, reach out.
Let's upgrade your marketing game
Get fresh tips, how-tos, and expert marketing advice every week. (15,000 people already are.)
