My job is to come up with good brand names: to find interesting, appropriate words for things. It takes time, effort, and is often seen as “the fun part” of what I do, which it is. I love my job. 

Another, more time consuming, and arguably more important part of my job is helping clients decide on which name to select, making sure the process goes smoothly and that clients leave our time together feeling satisfied with the outcome. That part is also (mostly) enjoyable because it’s challenging and I’m in a position to help and guide, even when it goes wrong—which can be more like “type II fun” (like a backbreaking hike to a gorgeous lake that at the time is terrible but I look back on it glowingly because it was worth the effort and/or I learned something about myself). 

Recently, I’m getting the question “do you think AI is going to take your job?” more often. For the past few years, I dismissed the idea and had little worry. Until recently, as I became fully aware of how good AI is becoming.

So, I’m here to find out if AI can take my job. I’m doing what I can to not enter this experiment with bias, but it’s hard not to when a significant portion of our work in 2024 and 2025 has been AI company and product naming. If AI is so good, you’d think the people running these companies wouldn’t need to hire us. 

That being said, I’m keeping an open mind.

The premise for this experiment will be a typical engagement for A Hundred Monkeys:

Three founders, a CEO, a CTO, and a CMO, who are launching a Kubernetes observability platform startup that’s about to come out of stealth. They’re launching in eight weeks so we have three weeks to get them to a group of contender names they really like so their legal team can run their due diligence before the founders make a final decision. Then we’ll hand off to an agency partner who will work on some basic brand building for the launch and continue fleshing out the brand post-launch. 

OK AI (I’ll be using ChatGPT here), show me what you’ve got. 

“Help me name a Kubernetes observability platform being launched as a start up coming out of stealth. We’ll be working with three founders. They need a legally viable name in a month.”

[Begin first ChatGPT response]

Here’s a curated list of brandable, legally viable-sounding names for a Kubernetes observability startup — modern, flexible, and aimed at standing out in a crowded space. I’ve grouped them into strategic themes depending on your positioning (technical, poetic, bold, etc.):

🔍 Clarity & Insight-Oriented

These emphasize the “observability” aspect — insight, clarity, control.

• ScopeStack
• ClearNode
• Lension (Lens + Vision)
• Obsera (Observability + Suffix for brandability)
• KlyrOps (Clear + Kubernetes + Ops)
• Reflekt
• Tracelet
• Prysmic (Prism + dynamic/mystic — light + insight metaphor)

⚙️ Kubernetes-First

These tap into technical/infra-savvy naming that appeals to devs and SREs.

• Kubiscope
• Podview
• Kyntra (K8s + Infra + extra flair)
• Podflare
• CtrlPlane
• Zerna (K8s + observability abstraction; short and brandable)
• K8teye (Eye on your K8s; edgy)

🧠 Intelligence & AI-Forward

If you’re planning to integrate AI/ML for anomaly detection or predictive insights.

• Metronome AI
• AetherIQ
• Vizionary
• Chroniq
• NeuralPlane
• Obsei (Open-source adjacent; but twist it for a unique mark)
• Opthalys (Sounds like a biotech, but makes a striking observability brand)

🌐 DevOps-Friendly / SaaS-Styled

These sound like other successful devtools/SaaS companies.

• Nodeloom
• Grafth
• Loopscale
• Optrix
• Tensyr
• Portlight
• Runvyu

🔬 Invented / Brandable Gibberish

These are more abstract, easier to trademark globally, and open-ended for pivots.

• Vetryl
• Syntara
• Orvyl
• Kynso
• Ovulaa (sound-first naming—scientific, clean, abstract)

⚡ What Next?

1. Favorites Shortlist — I can help vet for domain availability + trademark clearance.
2. Tone Match — Are you targeting platform engineers? DevOps leads? Security-conscious orgs? I can refine the list.
3. Name Stories — Want a name with a clear origin story (e.g. “Prysmic = seeing the truth through many lenses”)? I’ll craft them for pitches/investors/launch.
4. Domain + Trademark Search — Once you pick 2–3 finalists, we can check actual viability.

Would you like me to shortlist based on domain availability or trademark risk next? Or iterate based on a specific vibe you’re aiming for (e.g. bold, poetic, trustworthy, technical, etc.)?

[End first ChatGPT response]

So, at this stage I can see one, maybe two names I’d feel comfortable presenting to a client: Tracelet and Metronome (drop the AI). The rest are somewhere between bad and terrible: invented slop aimed to please the trademark gods and no one else. The main approach here is what we call wordsmash, so not a lot of variety and not an approach we ever recommend.

It’s also unclear what, if any, amount of trademark screening has been done on these names. (At A Hundred Monkeys run all names presented to clients first through a search engine knockout screening to weed out the obvious conflicts and then by an attorney to search against relevant TM databases (often the USPTO) database and have them assess viability.)

“please review “Tracelet” and “Metronome” for trademark availability and risks in classes 9 and 42”

[Begin second ChatGPT response]

Here’s a practical, “knockout-level” risk check for the two names in Classes 9 (software/downloadables) and 42 (SaaS & tech services). This isn’t a substitute for counsel or a comprehensive search, but it flags the most obvious landmines and next steps.

Tracelet

What we can see quickly

• I can’t find a current U.S. registration for the exact word TRACELET in 9/42. (You should still run a full USPTO/TSDR search before filing.) (USPTO, Trademark Search)
• The term “tracelet” is used in computer-science literature (e.g., HHVM/JIT and program-analysis research). That means it’s not purely fanciful; it has some descriptive/technical flavor in software contexts. Expect a slightly higher chance of descriptiveness pushback if your spec leans into tracing/profiling. (Engineering at Meta, CSAWs)
• Potential indirect noise: Medtronic has marketed a vascular compression device stylized as “TRAcelet.”Different category (medical device), but similar look/sound that a diligent Examiner or third party might surface. Not in 9/42, yet worth noting. (PTCA)

Class 9 (downloadable software)

• No obvious, identical live conflicts found in a quick pass. Given the technical meaning of “tracelet,” an Examiner could ask for disclaimers/clarification if your goods involve code tracing/telemetry. (Nolo)

Class 42 (SaaS / observability services)

• Again, no identical, live 42 registration popped on a quick check, but the term’s technical use could invite a mild descriptiveness inquiry if your SaaS is for tracing/telemetry. Be precise in your ID and avoid generic “tracing software” phrasing. (Nolo)

Bottom line for “Tracelet”

• Preliminary risk: Low–Medium. Looks relatively clear of identical 9/42 marks, but technical usage + proximity to “trace” concepts could draw questions. A professional comprehensive search is still recommended before you invest. (USPTO)

Metronome

What we can see quickly

• Crowded term. Multiple registrations exist across industries. Critically, there’s an active Class 42 registration for METRONOME covering SaaS in the transportation space (Swiftly, Inc.). That alone raises conflict potential for software/SaaS.
• There’s also a well-known billing-platform company operating under Metronome (software/SaaS). Even if their registrations differ by owner or class, their marketplace presence increases the likelihood of confusion/opposition if you enter adjacent software categories. (metronome.com, Tracxn)

Class 9 (downloadable software)

• Given how common the word is, expect prior 9 filings (including music-related goods, audio tools, etc.). Even if the goods differ, the crowded field reduces your scope and heightens opposition risk. (General class guidance cited here; a full database pull is advised.) (Nolo)

Class 42 (SaaS / tech services)

• The existing Class 42 METRONOME registration for SaaS (Swiftly) is a concrete obstacle. If your services are software/SaaS (observability), the USPTO could find relatedness; third-party opposition is also plausible. High risk of refusal or costly coexistence negotiations.

Bottom line for “Metronome”

• Preliminary risk: High (especially in 42, and likely in 9). The word is heavily used, at least one live Class 42 registration exists, and there’s a prominent software business using the same name.

Filing tips (for either name)

• File both 9 and 42 (if you’ll ship both downloadable agents/CLIs and hosted services). Misclassifying downloadable vs. SaaS is a common, painful mistake. (Gerben IP, JPG Legal)
• Craft tight IDs of goods/services (e.g., “downloadable software agents for application performance monitoring and distributed tracing” in 9; “providing temporary use of online non-downloadable software for observability of cloud-native applications” in 42). Don’t rely on vague terms. (USPTO)
• Do a comprehensive clearance search (USPTO/TSDR, state, common-law, domains, socials) before announcing or filing intent-to-use. For ongoing protection, set up a trademark watch to catch later-filed conflicts. (USPTO, Clarivate)

Recommendation summary

• Tracelet: Proceed to a professional, full-scope clearance. If results stay clean, it’s a viable candidate for 9/42, with attention to wording to avoid descriptiveness issues.
• Metronome: Retire it for observability software unless you’re prepared for high risk and potential disputes. The collision surface is too large in 42 (and likely 9).

If you want, I can draft specimen-ready goods/services language for 9 and 42 tailored to your product footprint so your counsel has a strong starting point. (Tramatm)

[End second ChatGPT response]

So, we have one name that, on first look, might be legally viable (it’s worth noting that ChatGPT fails to recommend checking homophones, other forms of the name (adjectival, etc), or names close but not exactly the same for trademark issues), and is creatively interesting. Here at A Hundred Monkeys we present 10-12 names per round, so it’s looking like it’ll take a while to go from one name to 10-12.

Assuming we do that work, which I’m sure we could with ChatGPT, can it build a compelling presentation deck, collect feedback, and most importantly, guide three founders through feedback, questions, and disagreements? 

BW: “can you build a presentation deck for Tracelet that places the name in context, builds rationale, highlights the trademark risk, and offers next steps?”

ChatGPT: “I’ve drafted a presentation deck-style document for Tracelet that frames the name’s meaning, rationale, trademark landscape, and concrete next steps. Would you like me to also design it in a slide-ready format (with suggested slide titles and bullet breakdowns) so you can present it directly to the founders and CMO?”

BW: “yes”

[Begin third ChatGPT response]

Tracelet: Brand Context & Trademark Viability — Slide Deck Outline

Slide 1: Title

Tracelet
Kubernetes Observability Platform — Brand & Trademark Review

Slide 2: Brand Context

• Name: Tracelet
• Category: Kubernetes Observability Platform (Software/SaaS)
• Positioning: Technical credibility + approachable brand
• Associations: Traces, observability, modular insights

Slide 3: Why Tracelet Works

• Meaningful yet distinctive → “trace” + “-let” (lightweight, modular)
• Memorable & pronounceable → clean, one-word brand
• Suggestive, not descriptive → ties to observability but not generic
• Story potential → “Each tracelet builds visibility”

Slide 4: Brand Rationale

• Technical resonance: Aligns with tracing, telemetry, spans
• Invented yet intuitive: Differentiates from common terms
• Scalable: Works for both agents (Class 9) & SaaS (Class 42)
• Flexible: Supports long-term product expansion

Slide 5: Trademark Landscape (US)

• Class 9 (Software): No identical active registrations
• Class 42 (SaaS): No identical active registrations
• Historical mark: Cancelled in Class 020 (bracelets) — not blocking
• Risk: Possible descriptiveness concern due to “trace”

Slide 6: Trademark Landscape (EU)

• Active mark: Medtronic holds TRACELET in medical devices
• Different category: Not in software/SaaS
• Risk: Moderate for expansion into EU markets; requires legal clearance
• Opportunity: Likely registrable in tech/software

Slide 7: Risk Summary

• US: Low–Moderate risk
  • Clean classes, but technical connotation could prompt examiner queries.
• EU: Moderate risk
  • Existing medical mark; coexistence strategy may be needed.

Slide 8: Next Steps — Legal

1. Commission comprehensive trademark clearance (US & EU).
2. Draft goods/services descriptions to avoid descriptiveness pitfalls.
3. File Intent-to-Use (ITU) in Classes 9 & 42 (US).
4. Monitor EU situation; consult counsel on expansion.

Slide 9: Next Steps — Brand

1. Build story framework → tracelets as modular insights.
2. Secure digital assets → domain, socials, GitHub.
3. Develop visual identity → balance tech + accessibility.
4. Prepare backup names in case clearance reveals conflicts.

Slide 10: Recommendation

• Tracelet is a viable brand candidate for US Classes 9 & 42.
• Distinctive, defensible, memorable.
• Trademark path is favorable with manageable risks.
• Next step: commission full search + file ITU.

[End third ChatGPT response]

Kinda decent. This is a very practical and dry approach, as there’s too little on the creative and brand potential—the most important piece at this step—which is unsurprising given what I know about AI and its inability to discern taste or think expansively in a creative sense. We always want our clients to be in an expansive mindset rather than a reductive one here. Trimming names is good, but thinking of possibilities and new angles is more important. Side note: ten slides on one name is overkill.

Now for the hard part, disagreement.

“the founders aren’t agreeing that Tracelet is a good name, how do you build consensus?”

[Begin fourth ChatGPT response]

That’s a common and very human hurdle — naming is subjective, and technical founders, business-minded leaders, and marketers often see it differently. Here’s a structured approach you can use to build consensus around Tracelet (or any candidate name) without it becoming a purely taste-based debate:

1. Align on Criteria Before Debating Names

Make sure everyone agrees on what makes a good name for your startup:

• Legal viability (trademark clearance, URL availability)
• Distinctiveness (not generic like “Cloud Monitor”)
• Memorability & pronunciation (easy to say and recall)
• Flexibility (works if product scope expands beyond observability)
• Storytelling potential (can you explain it to investors/customers in one sentence?)

Once the group signs off on the criteria, it reframes the discussion: “Does Tracelet meet these criteria?” instead of “Do I like Tracelet?”

2. Reframe the Name as a Story, Not a Word

People connect better when a name is positioned as part of a narrative:

• Tracelet = a lightweight, modular building block of observability.
• Each tracelet contributes to a bigger picture, just like spans and traces in Kubernetes.
• The suffix “-let” makes it approachable and product-friendly, like a tool engineers want to adopt.

This gives the name meaning beyond just sound.

3. Show Market & Legal Differentiation

Founders are pragmatic: highlight that Tracelet is:

• Largely clear in Classes 9/42 (low-moderate trademark risk vs. many other names already taken).
• Short, brandable, and dot-com or strong domain variants are more likely available than descriptive names.
• Different from the crowded “-ops / -monitor / -insight” naming space, which makes it easier to stand out.

4. Use a “Shortlist Comparison” Exercise

Put Tracelet on a shortlist with 2–3 other legally viable options (from your earlier explorations, e.g. Vetryl, Orvyl, Prysmic).

• Score them against the agreed criteria.
• Often, Tracelet rises in consensus once it’s seen in relative context instead of isolation.

5. Highlight the Cost of Indecision

Remind the team:

• You need a legally viable name within a month.
• Endless debates waste time and risk missing domain/trademark filings.
• An “80% good name” chosen now is more valuable than a “perfect name” chosen too late.

6. Offer a Path Forward

Consensus doesn’t mean unanimous “love” — it means shared confidence. Try this script in a meeting:

“Let’s agree: no name will be perfect to everyone on day one. The goal is a name that’s viable, protectable, and brandable. If Tracelet meets those criteria better than the alternatives, can we align on it and move forward?”

✅ If consensus still doesn’t form, the fallback is a structured decision: the founders + CMO vote using weighted scoring (legal strength > memorability > personal preference). This takes it out of pure gut feel.

Would you like me to draft a scoring matrix with Tracelet vs. 2–3 other candidates (e.g. Vetryl, Orvyl, Prysmic) that you can use in a founder workshop to drive this discussion?

[End fourth ChatGPT response]

Honestly, not bad. It’s a comprehensive approach, covering all the right angles: timing, brand considerations, trademark, avoiding perfection, etc.

So, will AI take my job? Maybe in some instances. I’m sure there are plenty of cases where people have used AI to name their company, product, feature, offering and have been successful. Is there a subset of those people who might have gotten in touch with us and have the time and resources for a project together? Probably.

On the other hand, we’re about as busy as we’ve ever been—naming several AI companies and features this year and last, along with the usual mix of companies and products like a bakery in the San Juan Islands, medtech, a financial advisory service, top secret CPG products, and a soon-to-be-launched wine.

One split I see here is do you want to work with a human? Do you want to be able to pick up the phone (or the more modern version: schedule a Zoom meeting) and hash it out with me or one of my stellar colleagues? Do you want the advice of someone who’s been helping clients for a decade or more or do you want to go it alone with some amazingly advanced technology? 

I, for one, prefer humans and suspect I’m not alone. 

If you have a project in mind and want to engage in a human-driven process drop us a line.