Deep-tech brand and product: making an invisible product visible

Some products can’t be photographed. Mesh WiFi is radio waves and cloud software. A foundation model for chip design is math running on servers. You can show a box or a screen, but the thing that actually creates value is invisible.

That’s the core problem of deep-tech brand and product design. People can’t remember what they can’t picture, and they can’t trust what they don’t remember. The brand has to supply a metaphor that makes the product graspable, and the product has to keep the promise that metaphor makes. I’ve worked on this twice as the person owning both sides: at Plume, starting in 2016, and today as Head of Design at Cognichip.

The invisible-product problem

When Plume launched its consumer product, the category itself needed explaining. The technology behind Plume, pods coordinated by cloud intelligence that continuously picked the best channels for every device, was genuinely new, and almost none of it was visible.

Cognichip has a harder version of the same problem. The company is building Artificial Chip Intelligence (ACI®), described on its site as AI that learns, understands and solves chip design problems, built on a physics-informed foundation model purpose-built for chip design. Its $60M Series A, led by Seligman Ventures, also brought Intel’s CEO, Lip-Bu Tan, onto its board. The audience is semiconductor engineers and the companies that employ them: technical, skeptical, and allergic to hype.

In both cases, describing the technology accurately wasn’t enough. People needed something to hold onto.

Finding the metaphor

A good deep-tech metaphor does three jobs: it makes the product memorable, it signals the right emotion, and it doesn’t overpromise.

At Plume, the metaphor was effortlessness. Early marketing used a floating pod motif, hardware that seemed weightless, as if it simply took care of itself. The app’s home screen showed a pod that breathed, signaling a system that was alive and quietly working. The message wasn’t "advanced radio engineering." It was "you’ll never think about WiFi again."

At Cognichip, the brand direction is liquid hardware: imagery where silicon and circuitry feel fluid and alive, suggesting a design process that adapts rather than grinds. It’s a way of saying that chip design is becoming something more intelligent and responsive, without claiming specifics the product hasn’t earned. The Cognichip case study shows the brand reel and launch poster.

The test I use for a metaphor: can an engineer hear it without rolling their eyes, and can a non-engineer repeat it the next day?

Directing a launch film

For Cognichip’s launch, I directed a film built with generative AI, end to end. I won’t go into production specifics here, but the general lesson about directing with generative tools is worth sharing.

Generative tools can produce endless beautiful frames. That’s the problem. Without a clear point of view, a film becomes a reel of impressive but unrelated images. Direction is what makes it one piece: a single story, a consistent visual logic, and the discipline to cut anything that doesn’t serve the metaphor. What I direct is the story, the world and every selection. What the tools contribute is speed and range. The film is only as coherent as the decisions behind it.

Keeping the metaphor honest inside the product

The fastest way to break trust is a brand that promises one thing and a product that feels like another. If the launch film says "fluid and intelligent," the product interface had better not feel like a dense legacy tool.

At Cognichip, I own both product and brand from zero to one, so the same design engine produces the website, the film and the product’s agent interfaces. When identity and product share one system, every new product surface still looks and sounds like the company. At Plume, the same principle meant the calm the advertising promised had to show up in the app, the setup flow and even support emails. I’ve written more about that in Product and brand as one system.

For agent products specifically, the metaphor extends to behavior. An agent that overclaims breaks the brand as surely as an off-brand screen. An agent that is honest about uncertainty keeps the promise.

Brand for two audiences at once

Deep-tech companies rarely sell to one audience. Plume spoke to homeowners and, from 2018, to the service providers who offered Plume to their subscribers. Cognichip speaks to the engineers who would use the product, the executives who would buy it, and the investors and talent watching the category. Each group needs something different, and the brand can’t split into separate voices without losing its center.

The approach that worked at Plume, and that I use now, keeps one core promise and adjusts density by audience. The engineer gets precision and proof. The buyer gets outcomes. The broader market gets the metaphor and the story. All three should recognize the same company in a single glance at the site, the product and the film.

Practically, that means a messaging hierarchy that starts with the shared promise and branches by audience, a visual system that holds across a technical product interface and a cinematic launch, and a habit of reviewing every new piece against the core metaphor rather than against the audience alone. The metaphor is what stays constant while everything else adapts.

Naming and claims discipline

Deep-tech brands live or die on credibility, so claims discipline is part of the design job:

  • Use the company’s own marks consistently. Follow how the company writes its product names and trademark symbols, and don’t mix variants across pages.
  • Public facts only. In public work, stick to what the company has announced: its site, its press releases, its published materials.
  • No borrowed metrics. A company’s performance claims are not a designer’s personal results. Keep them separate.
  • Let the product earn the adjectives. "Revolutionary" means nothing. A specific, verifiable capability means a lot.

A brief template for a deep-tech identity

When I start a deep-tech brand, the brief covers:

  1. The invisible thing. What creates value that people can’t see?
  2. The audience’s skepticism. What will make them distrust us?
  3. The metaphor. One image or idea that makes the product graspable without overpromising.
  4. The emotional register. Calm, bold, precise, warm? Pick one primary.
  5. Proof points. Public, verifiable facts the brand can lean on.
  6. Product continuity. How the metaphor shows up in the interface and the product’s voice.
  7. What we won’t say. Claims, comparisons and words that are off-limits.

Picture first, promise kept

When the product is invisible, brand supplies the picture and product keeps the promise. Find a metaphor engineers can respect, direct every asset toward it, and make sure the product feels like the story. For another take on making identity and product cohere, see the Full Node Design case study and the Plume brand work.

Related Product and brand as one system: what eight years at Plume taught meDesigning data-dense dashboards: role-based views that surface the next actionConnected hardware: designing the box, the app and the brand as one launch

See more workGet in touch