Stop Building Pitch Decks. AI Just Automated 80% of Your Sales Job.

You’ve probably been here before: It’s 11 PM. The sales rep messages you saying, “Hey, can we get a deck for the client by tomorrow morning?” So you start manually stitching together a PPT, scavenging feature points from internal docs, inventing “customized value propositions” based on a vague industry guess, and wasting two hours tweaking alignment and brand colors. We’ve all done it. And honestly, this repetitive cycle should have been automated years ago.

Anthropic recently open-sourced a Claude Code Skill called create-an-asset. Its promise is simple: feed it your customer info, audience type, and delivery goal, and it spits out a fully branded, client-ready asset. Whether you need an interactive landing page, a linear slide deck, a one-page executive summary, or an animated workflow demo, it handles the heavy lifting.

But this isn’t just a tool that writes copy. It attempts to compress the entire presales workflow—from information gathering to visual construction—into a single conversational pipeline. And after tearing apart its design, one thing is clear: AI isn’t killing sales jobs. It’s killing the illusion that copy-pasting features into a slide deck was ever a skill.

The Skill operates on a 7-stage pipeline, but the real genius lies in its adaptive research logic. If you upload rich context—like meeting transcripts and specific pain points—it skips the heavy lifting and just fills in the gaps. If you only give it a company name, it goes deep: pulling annual reports, analyzing CEO speeches, extracting the target company’s actual brand colors from their website, and mapping industry trends. It then routes the asset based on your audience. Executives get ROI calculators; technical teams get architecture diagrams.

However, there’s a deliberate trade-off baked into the design: a hard limit of two clarification questions. If your input is too vague, it won’t hold an endless interview with you. It makes a reasonable assumption, builds the asset, and flags it as “adjustable.” This guarantees speed, but it exposes a harsh reality for lazy sellers.

If you think you can just type a company name and let the AI read minds, you’re going to get a generic template that screams ‘I didn’t care enough to prepare.’ When production becomes a parameterized pipeline, your ability to read the room becomes your only real moat.

This Skill flawlessly automates about 80% of the presales process: the page structures, the visual hierarchy, the brand color matching, and the content framing. What it leaves behind is the remaining 20%—the unspoken needs you catch during a meeting, the industry nuances you understand, and the trust you build. That 20% is the new differentiator.

As these tools become standard, the baseline for “customized” assets will skyrocket. Automated brand color matching will just be table stakes. The differentiation will return to the most primitive question: How deeply do you actually understand your customer, and can your product solve their problem?

The work is shifting from “making deliverables” to “understanding clients.” The more we automate the mechanical parts, the more valuable the un-automatable judgment becomes. If your entire value was building slides, your career is on borrowed time. But if you know how to close a deal, congratulations—you just got your nights and weekends back.