This Generative AI Company Doesn’t Want To Be Called A Generative AI Company

Recently, a couple of venture capitalists knocked on Ben Pham’s door to hear a pitch on his new advertising startup named Obello.

Pham had decades of experience in advertising with Character, a branding agency he co-founded with Oliver Ralph that worked closely with the likes of Apple, Netflix and Facebook. After selling the company to marketing giant Dentsu, Pham and Ralph were working on a new generative AI product.

The first line of the pitch deck read, “Obello: An AI-powered graphic design platform.”

“I visually saw them recoil in their seats and roll their eyes, because they had been beaten over the head by so many companies saying ‘AI-powered’ [or] ‘powered by AI,’” Ralph said.

That’s not surprising. After ChatGPT took the world by storm in November, generative AI startups began popping up left and right. Venture firms were enamored with it — around 10% of global startup funding last year went to AI companies, according to Crunchbase data.

We’re on track to outpace that: So far in 2023, 15% of all startup funding has gone to AI-focused startups.

But the dust is beginning to settle, and some venture firms are getting skeptical of startups that claim to power their platforms using generative AI and large language models.

“The word is hurting us more than helping us pave the way, because the promise hasn’t been delivered,” Pham said.

Genesis of Obello

Pham’s previous startup Character built cohesive design packages for startups and legacy companies, which included tone, color, typography and logos. Character sent this package to its clients’ brand and marketing teams to turn those pieces into emails, landing pages, Instagram stories, Facebook posts and vertical ads.

“One of the challenges that we witnessed for companies of all sizes is that the designers and the marketers are struggling to keep up with the increased demands placed,” Pham said. “Modern brands have to be omnipresent and, as a result, they have to create a lot of marketing content.”

Their new venture Obello takes those brand ingredients, organizes them on its platform, and uses its in-house generative AI to create content based on what platform it will live on, what dimensions it needs to be, and even what tone the company wants to set — it can be “playful” or “understated.”

Traditional workflows, Pham said, “pull designers away from doing other more conceptual, creative work within their organization.”

How AI lost investors

While Obello may sound like a promising concept, it’s simply one of many new generative AI startups to investors.

A ton of generative AI startups build intuitive UI designs that sit on top of APIs from the likes of ChatGPT. These companies don’t have their own proprietary language models or data sets.

“You’re building basically something that is not super defensible and is easy to be replicated by another market entrant,” said Pasha Tinkov, co-founder of Audeo Ventures. He met with the Obello team. “It’s not a great business case to be investing into.”

For those companies, it’s hard to develop long-term business strategies that will net investors multiples on their investments. It also makes them — and those that use them — vulnerable to potential legal ramifications, compliance roadblocks and copyright issues in the future. (AI art tools like Midjourney are already being hit with copyright lawsuits for stealing from artists.)

“I can’t use a part of someone else’s IP because the harm, from a reputational point of view and a monetary point of view, for big brands is going to be huge if they’re discovered using or plagiarizing other people’s IP,” said Eric Hippeau, co-founder of VC firm Lerer Hippeau.

As a result, investors have been more skeptical about the long-term viability of generative AI.

Self-funding a new company

When Ralph and Pham began Obello in April 2022, the economic situation was bleak. Private-market funding was drying up, interest rates soared and private companies were grappling with sky-high valuations.

“What’s required from companies today is they have to be fiscally responsible,” Pham said.  “What’s important for us is to build a product first.”

Startups in most industries aren’t getting the same love as AI. Mistral AI, a startup founded in May, managed to raise $113 million despite not having any sort of product.

“Let’s say we went to a VC and they’re like, ‘Oh my goodness, this is generative AI. You’re training a model. Let me value your company at $25 million,’” Ralph said. “Then we have to actually deliver, to perform as a $25 million company. And that’s a huge burden to carry.”

While most startups are facing a valuation reckoning, and venture firms say they’re course-correcting, it seems like the AI hype is following in the footsteps of crypto and Web3.

“We’re happy that we’re in the right space at the right time,” Ralph said. “But also we don’t want to set ourselves up for failure in the future.”

Illustration: Dom Guzman

Learn More