AI: a lesson in brand naming from an industry doing it right

Having spent a decade working with brands – as an early-stage VC and co-founder of a brand strategy agency – I’ve noticed one sector putting others to shame when it comes to naming habits: artificial intelligence. So why are AI start-ups consistently nailing their naming, and how do we learn from it?

Regardless of industry, brand names tend to fall within set categories. For example, you have geographical names like Arizona Tile, eponymous names like Boeing and Tesla, and acronyms like IKEA, IBM and ABC. Specific to the technology companies, four exciting categories of names have emerged.

1. Category: Descriptive

Descriptive brand names convey the service or product on offer. By nature they’re typically functional and utilitarian (General Motors). However, AI products have injected excitement into an otherwise unremarkable category.

Trend: Robot Love

As engineer-driven businesses, many AI companies opt for descriptive names based on popular industry terms — think “ “Robot”, “Deep” and “Data”. Take DataRobot, an enterprise AI for developing and deploying Machine Learning who have stuck to the naming script and raised over $430M in the process. Then there are more straightforward company names such as DeepAI, DeepLearning, Deep Genomics – all literal. Are traditional descriptive names boring? Yes. Are robots boring? No. Therefore AI companies are at a distinct advantage.

2. Category: Evocative

On the opposite side of the creative spectrum you’ll find evocative names—names based on metaphors that evoke the kind of feelings, experiences, ideals or other qualities a company wants associated with its brand. In AI, evocative names such as Cruise, Vicarious or my favorite AI-driven game AI Dungeon, are layered and complex in their meaning, providing a rich platform upon which a company can develop a brand that’s larger than the sum of its services.

A useful and often underappreciated upside of the evocative category is the ease—and cost-effectiveness—with which trademarks can often be secured. The downside is the lack of immediate association with industry vocabulary, meaning brand recognition will depend heavily on the amount of marketing investment. The .ai domain extension, however, gives the startups freedom to be evocative that doesn’t exist in other industries, as it turns any brand name into an AI centric one.

Trend: Humanisation

When I own a robot I’ll call him George or Fred, not X Æ A-Xii. Major AI companies think along the same lines: names such as Watson, Alexa and Siri humanize technology and serve as prototypes for AI-inspired movies like Her; a recent Vox article noted a similar trend for banking and money apps.

Trend: Misspelling

Misspelled names have been popular for years, the trend was covered by TechCrunch in 2017 and four years prior to that by The Next Web’ in “why the trend for misspelled words won’t go away” (maybe it never will?). So why is this approach seemingly evergreen? It’s simple – there are just not that many words and free domains out there. Switching vowels or consonants with similar sounding alternatives can help a startup obtain a short, familiar domain without breaking the bank. The AI space is no stranger to the misspelling trend, and many in the industry embrace it.

A few examples are companies like AI-powered end-to-end data analytics platform Alteryx, or our client the API for Conversational Intelligence called Symbl, whose name alludes to their capacity to analyze conversations.

3. Category: Lexical

Lexical names rely on wordplay for memorability and are defined by puns, compound words, clever alliteration or usage of foreign words. The downside of lexical names is they can come off as “too salesy”. One of the most famous examples of such a name is Playboy, successfully renamed from Stag Party last minute.

Trend: Traditional Lexical Names are Back

Long after the frenzied ads of the Mad Men era, AI startups are bringing Lexical names back, adding an overtone of humble humor to the coldness of the technology space. An old dog of the AI world, SoundHound (cheap pun intended), is part of the lexical revival, using cliched naming characteristics (simple rhyme, symmetrical wording) to sell a complex technology. The result is a brand that feels whimsical, relatable and trustworthy – and even better it directly relates to their service as an audio recognition company. CloudMinds and DeepMind (with its gorgeous website) are two other AI companies whose names add romanticism to their technical offering; and then there’s Sift Science who recently acquired an extremely desirable domain and can now be found at

Trend: Hiding in plain sight

“Ai” can be included within a plethora of words making it a word players paradise. Adapting real words to include “ai” adds new meaning to a seemingly ordinary name and this clever approach deserves a special mention. Ladies and gentlemen, please bow to the likes of (open eye), Aeye (a-i), and Clarifai (clarify).

4. Category: Invented

Whenever I heard Xerox as a child the name conjured up images of fantasy legends and mythological times. It turns out I was confusing it with Xerxes, a Persian king – the printing wizards made this name up. When this was pointed out my embarrassment was comparable to that of Xerxes after his failed Greek invasion in 480BC.

Many invented names are modified from Greek or Latin words. Kodak, Xerox and Verizon are all examples of invented brand names that have gained so much traction they are now considered real words.

The upside of invented names is simple: trademarks are easy to come by and domains are typically a breeze to secure. Additionally copycats are simple to spot.

The downsides are similar to those of evocative names: invented words require investment. To build a genuine connection between name and company, time and money need to be spent forging a clear association and giving your brand name a meaning consumers can relate to. The AI industry has taken the invented category one step further by adding a new dimension, numbers, into the mix.

Trend: Numbers

Who if not the tech-driven AI leaders (and Elon Musk) could lead the charge in using numbers in names, setting trends and making domain sourcing simpler in the process? Check out the wisely named (aiming at making AI as popular as water), Deep6 AI (saving people from drowning in data) and the straightforward but no less impressive Oculus360.

Trend: (un)popular letters

As a rule of thumb companies prefer names that start with earlier letters in the alphabet (it optimizes discovery as they appear higher on lists and presentations): Alphabet, AirBNB, Agora, Affirm etc. Yet in the sea of A’s, B’s and C’s, names that start with the last letter of the alphabet shine just as bright. Unsurprisingly the AI industry has been quick to capitalize on this opportunity – step forward Zoox and Zymergen, two large AI businesses who have raised $950M and $574M respectively.

Naming trends must be heeded – who knows when a seemingly fleeting approach may become an instant classic – but not necessarily followed. When naming your own business, turn to the AI space for inspiration. The industry has no problem adapting existing trends and in some cases breaking the rules entirely.

In doing so, new standards are set and brands developed that are both different and authentic. Naming a business is an overwhelming process, but the hardest part is knowing where to start. Before diving into the abyss take a look at these must follow rules – stick to them and, who knows, maybe you’ll come up with the new Backrub, these days known as Google.

Get in touch

Inspired? We’d love to connect and explore ways to collaborate.