Hi all,
Happy Sunday from NYC. If you’ve also spent the majority of your summer grinding through work, congrats. I probably should have taken more breaks (and I know you should have too) but I’m at this point where if I continue to grind it out for the next few years, the next decade will seemingly be way easier.
As a fun fact, at my marketing company Knight Vision, we get all of our companies via inbound, either through referrals, VC partners, outreach from my content/newsletter, or warm intros. This means we see a ton of dealflow that’s already been filtered through.
There’s so much money flowing into AI right now. Every day there’s a new company, new tool, new category, new pitch.
Despite all of this, buyers do not understand what a lot of these startups are selling.
That is a huge, critical error. Here’s how it’s playing out:
AI companies market their products assuming the buyer already understands terms like "agentic AI," "AI workflows," and "digital workers," and what they mean in practice.
From what we’ve observed at Knight Vision, whether its via language that resonates with millions $$ in spend or conversations with customers through our “Customer Creator Program,” it increasingly shows that assumption is false, even in B2B, even with sophisticated buyers.
TLDR: People do not really understand what these AI terms actually mean when they are functioning in live working environments.
AI products have evolved faster than the audience's understanding of them. So companies either speak in technical language buyers can't decode or oversimplify until the product becomes hard to picture, differentiate or believe.
OPEN YOUR EYES:
You can see this everywhere.
There are companies describing themselves in ways that sound impressive but are still too abstract for the customer. If a buyer does not understand what “agentic workflows” means in practice, the message falls apart immediately.
And don’t make the mistake of thinking a more specific ICP means your buyer needs less explanation. Being the right target customer doesn’t mean you automatically undersand a new AI product or category.
You’re going to now walk through a series of landing page screenshots (live today) of companies that have collectively raised $1B at $10B+ combined valuations. The landing pages are a result of this problem:
AI products are complex, so companies invent language like “agents,” “digital workers,” and “AI employees” to explain them. Those terms still don’t mean much to the average buyer.
When companies try to simplify the message, they strip away so much detail that the product becomes hard to picture or trust.
The result = marketing that is too technical to understand and/or too vague to imagine.
You can also see how the market and marketing become repetitive because of this.
There are so many AI employee companies, digital worker companies, workflow tools, and software-building tools that they blur together. When everyone and everything sounds similar, vague positioning is an extreme liabiliy.
Town raised $55M Series A From a16z and Forerunner, announced in June 2026
Viktor raised $75M from Accel to bring an AI coworker to every team in the world, announced in May 2026
Sable raised $45M from 8VC and Sequoia to build the first AI employee that can click, see and explain, announced in July 2026
11x has raised $76.0M across 3 funding rounds. Most recently, it raised $50.0M Series B led by a16z in November 2024 for AI workers.
Dapta has raised $5.2M across 2 rounds, smaller than many competitors, but has totally crushed it with impressive market share in LATAM.
Maven AGI has raised $78M total , most recently $50M series B announced June 2025 to expand from task autiomation to building “Business AGI that powers enterprise end-to-end”, aka completeting complex workflows vs. just isolated tasks
Can we also just LOL a little how everyone is giving their AI agent a human name (Viktor, Aidan, Alice, Julian, etc.) as if that’s a key in marketing now to humanize the agent.
Do you still think it’s a coincidence that all these companies sound the same?
Let’s look at AI video editing and video generation.
$315M most recent raise at $5.3B valuation announced Feb 2026. To their credit, “while the company has historically built a strong customer base in media, entertainment, and advertising — including a recent partnership with Adobe — a spokesperson told TechCrunch that Runway is increasingly seeing adoption in gaming and robotics.” ”Runway released its first world model in December and now views the technology as central to tackling major challenges across fields like medicine, climate, energy, and robotics.”
Now let’s bring it back full circle with creative specific agents.
Luma AI raised $900M Series C announced Nov 2025 to develop what they call "advanced AI systems" that learn from multiple data types, including video, audio, and text.
You have to admit, it’s a little ridiculous.
So, what do you do?
Go vertical-specific: If you are building in a crowded market, broad messaging is not enough (e.g. screenshots above where everything sounds the same). You need to anchor the product in a specific buyer, use case and specific workflow.
Example: beehiiv (one of our portfolio companies at Knight Vision!) has done an incredible job of this. They started off 4 years ago in “newsletter” software for journalists and creators, and now service creators across the entire digital economy via digital products, podcasts, etc.
The first iteration of the beehiiv landing page from 2021, 5 years ago. Only about newsletters and for newsletter operators.
beehiiv’s landing page today. Newsletters, podcasts, community, etc - for brands, creators, businesses and more.
Educate better: Once you narrow in on a specific vertical, do not assume your end-buyer has the same rich vocabulary you do. Of course, there are exceptions, and if you’re talking to the CTO this doesn’t apply in the same way. More technical buyers may need less edu, but no one should have to decode your product. And the average end-user of your software at a company is likely going to be far less sophisticated. A few guiding principles:
Assume the average person does not know what an AI agent even is and that “agentic” has literally 0 meaning.
If the buyer still needs help understanding what the product does, how it works, and why it matters, then teaching is part of the job. Build an onboarding experience that details this.
Example: When Knight Vision onboarded to Melius, we had to integrate platform specific language like “node” and “canvas” into our creative workflow vocabulary.
Demo real customers: These proof points are literally the top thing that matter bc, with so much overlap between products, they make a new category and company feel real. People understand products faster when they can see themselves in the end buyer and relate to the example in front of them.
The companies that win in AI will be the ones that know how to make a complex product understandable, incredibly vertical specific and rich with examples of people using them.
In consumer, we call it a “mimicry effect”.
Examples of how to seed this in a product: Leaderboards where users can see how top players win, templates in Canva that people can adapt with one button, filters on Instagram.
Now it’s your turn. Make it specific and special to your company. Take vertical specific users + add real customer examples + ways to immediately copy them.
Well, How’d I Do?
Build your customers, prominent creators and vertical density into your moat.
Do it before you have to rebuild your product because of lack of differentiation.
If you have any specific questions on your company please reply to this email, happy to share thoughts.
I hope you have an incredible week ahead.
Julia
