KEY SUMMARY
Buyers now let AI shortlist vendors for them, then look closely at the few that make the cut. Design that only looks decent no longer earns that look. The brands that do earn it are clear about what they stand for before they spend money on how they look. They keep their logo, colours and typeface steady long enough for buyers to remember them, let everything else change freely, and choose on purpose whether to blend in with their category or stand out from it.
I used to give ten vendor websites thirty seconds each. Now AI gives me a top three. That's exactly what is happening with your buyers as well.
Those thirty seconds were where the visual brand did most of its work. It was the first impression that decided who got a second look. Now a machine makes the first cut, and the decision-maker with this shortlist is looking much more closely.
Brand design has been through shifts like this before. It started as a logo and a style guide. Over time it became the look and feel of everything a company puts in front of buyers. Now it is being reshaped again, by AI on both sides. First in how buyers find you, and second in how design itself gets made.
The hiring playbook that refused to be boring
From the Outbound Hiring Playbook (Belong.co, Bloomsbury). Illustrations by Saptarshi Das, Art Direction by Roshni Chinubhai.
A few years ago, Belong.co needed an outbound hiring playbook. Most teams would have produced the usual version of graphs, charts, and a gated PDF nobody remembers by Friday.
Roshni, Saptarshi and the team told a story instead. They cast recruiters as explorers, heading out to scout talent and bring the right people home. The playbook stood out from everything else in its category, and people in hiring still remember it. I still remember it.
That's worth keeping in mind if you sell to businesses. Your buying committee is made up of people, and people respond to stories. A familiar metaphor, applied to a B2B problem, can land harder than a deck full of charts. Every interaction a buyer has with you, in words or visuals, adds up to a story they either trust enough to buy from, or don't.
"B2B businesses are eventually made up of humans only. The same stories will work there as well."
Saptarshi Das, on the GTM & AI Operators Show
This guide is for operators on either side of a brand decision. The founder or marketing leader deciding whether a brand is worth money right now. The design operator trying to get a usable brief out of that founder. By the end you should know how to test whether your brand is weak, when to invest, what to lock down, when to break from your category, and what AI has changed about all of it.
Most of what follows comes out of my own ten years in B2B marketing, my conversation with Saptarshi Das on the GTM & AI Operators Show, and a discussion with Puja Kumar. Saptarshi runs brand at Adobe, scaled brand identity at Sense from Series C into the enterprise AI era, and also runs his own brand practice, Sketchboat. Puja runs her own visual design practice, Studio Bauno, and previously led visual brand at SpotDraft and Wingman.
You can watch my conversation with Saptarshi Das on visual design in the AI era here.
Run this test on your brand today
Imagine your brand as a person. If the brand is Harley Davidson, you get an image instantly. Muscly biker dude, leather jacket, beard. Run it on IBM and someone in a suit shows up before you finish the sentence.
The Test - Do people get a vivid image when they imagine your brand as a person?
Picture your company as a person walking into a room.
Write their description in one sentence.
Ask honestly whether that sentence could describe anyone else in your category.
This takes 30 seconds and zero tools. If your description is generic, keep reading.
If the person who shows up could describe any brand in your category, so could your brand.
The test works because brand design is not a logo project. It is the creative infrastructure of your company. Walk into a McDonald's and you know where you are even if you never see the yellow, because every part of the picture tells one singular story. Your aim should be to build an identity that permeates everything, so your audience gets the story before anyone says a word.
If you didn't do so well in this first test, the sections below are roughly the order in which to fix it.
Invest when conviction arrives, not before
If you cannot state exactly what your company is and who it is for, do not spend on brand yet.
Without that conviction, no identity survives contact with your audience. You will see some other brand on Tuesday and want to look like that. A different one on Thursday. Your designer cannot fix this, because an identity can only translate and not manufacture conviction.
Cred was the opposite in its early days. Its founder Kunal could give you an exact statement of what Cred is and who it is for, every single time, in the same words. That clarity is what a designer builds on. Waiting until yours arrives is a legitimate strategy. A cheap logo you replace in a year cost you nothing if it got you to market. Your ability to do that depends on the context of your market. The luxury watchmaker cannot afford a shoddy launch. The founder shipping Thursday can.
Puja doesn't expect founders to walk in with every answer, though. At Studio Bauno, every project starts with Clarity, the first step of her C.A.S.E. framework (Clarity, Align, System, Execution). What she asks of founders is to work through that step with her and be willing to decide by the end of it. So a lack of conviction isn't a reason to stall. It's a reason to make clarity the first deliverable.
Once conviction exists, though, most B2B teams underinvest, because they treat brand as less of a differentiator than the offering and the sales deck. The 95-5 rule says why that math is off. Only around 5% of your buyers are in-market in any given quarter. The deck reaches only the 5%. The brand is what the other 95% remember.
Lock the core, let the rest move
The old dogma said brand identity must stay consistent above all. In 2026 that is half right. Your product will change under you. Ten new offerings ship in your org before your next quarterly review, and the identity has to flex with what you sell.
But that does not mean everything gets to move. One way to look at this is:
Your logo should survive the next three to four years of the company's plan.
Your typeface and palette should not look dated in one or two, so be careful when you lock those in.
Let campaign styles, illustration systems and surface treatments move with the moment.
Lock what buyers recognize as yours. Let everything else move with the moment.
When you're deciding on fixing or flexing your identity, do this first. Score each brand element on how many of your buyers recognize it, and how many link it only to you. Whatever clears both bars is your locked core. Everything else is negotiable.
Puja's advice is to commit to that core for longer than most founders want to. She'd rather a founder keep their logo, typeface and palette for five years and let them get stronger than redo them every year. Everything else - campaign visuals, illustrations, layouts, photos, launch pages - is meant to feel fresh for a particular moment.
One more thing scale does to this decision. Big companies converge on white, grey, steel and blue, and the usual read is cowardice. The real reason is operational. A complex identity cannot be executed by thousands of teams across hundreds of geographies, which is why Microsoft looks the same in India as it does in the US, held together by grid systems anyone can run. Cred chose the opposite trade. Both are rational. Quirks are a privilege of being small, and a complex identity is a cost you fund deliberately or not at all.
Match your category, or break from it
Every era of B2B has a default look, and you have watched them cycle. Flat cartoon figures after Facebook's illustration system spread everywhere. Trust-blue for enterprise. Dark gradients once everyone copied Linear. Now black-purple for anything AI, because everyone's mental image of futuristic is still Star Wars.
Every era of B2B has a default look. Matching it should be a choice, not a habit.
The convergence has a mechanism. Pinterest and Dribbble upvote homepages teach every founder the same picture of what modern means, and defaulting to what the market leader already proved is the easy way out of an unfinished decision.
So the real question for you is whether looking like your category is a bug or a feature. If your sector is new and you want to look like you belong at the top of it, tribe-matching is a defensible call.
Puja's test for this is simple: "Everyone else in the category does it" isn't a reason on its own. When ten products look alike, buyers can't remember any of them, and that's the bigger risk. There can be a good reason, though. When customers are very cautious, a familiar look helps them feel safe, so she keeps a few of those signals, like a clean layout and a plain product screenshot. And when a brand does stand out, she picks one or two things to be bold about - the colour, the typeface, or how the brand talks - and keeps everything else simple, so the brand is easy to trust too.
If you do decide to break out, it needs to be deliberate. It also has to come from something true about the product, not just a wish to look different. While the AI category raced toward geometric gradients, Anthropic went hand-drawn with Claude. An imperfect, humanized mark in a category obsessed with looking machine-perfect. That was a deliberate positioning statement about human-centered AI, made visually, and it is why you recognize their pages at a glance.
When you do take the risk, take it where your core offering stays untouched. Workday put Ozzy Osbourne and Joan Jett in a Super Bowl ad telling corporate types to stop calling themselves rockstars. A crazy swing for HR software, and a safe one. The joke rode on culture while the product claim stayed conservative. That is the template for B2B risk-taking.
The brief is where AI helps you most
There is a language gap on every brand project. You say classy. Classy means three different things to the three people at that table. Vibrant and bold could be saffron orange in your head and dark blue in your designer's.
If you are the design operator, closing that gap is your job, not the client's. You are in the business of cognitive understanding. If the rest of the company speaks one shared language, the creative person entering that system learns it, not the other way around. That looks like workshops, mood boards, keyword alignment, and competitive references. Saptarshi's client questionnaire has one line item that outworks the rest, and you should steal it whichever chair you sit in: list everything you do not want. Negatives are easier to articulate than positives, and they fence the territory fast.
If you are the founder, AI just handed you a cheat code, and it is not image generation. Before you brief anyone, prompt for a mood board of the word in your head.
For example, intellectual. Look at what comes back and notice your own reaction. If it's a "no, that is not what I mean by intellectual," find better words to describe the image in your head and prompt again. By the time you reach a designer, you have done the articulation work yourself and you arrive with a document instead of a vague adjective. In my experience the weeks lost on brand projects die exactly here in translation rounds, and this exercise solves that pretty simply.
Use AI to find your own words before you brief a designer.
Earlier, when a founder said "I want it to feel premium," it took a few rounds of back-and-forth to work out what premium meant to them. Now they've already played with prompts and can show the designer, which saves most of the first meeting. But there's a limitation as well. Because those images copy what already exists, they look polished but familiar, the same as every other brand that prompted its way there.
"AI can show you what looks premium. It can't tell you why, or whether it's right for your business."
Puja Kumar, Founder, Studio Bauno
Treat the founder's AI mood board as a conversation starter, not the answer. The designer should ask the founder what about it feels right, then build on that idea and push it further until it fits the business. Deciding what a visual direction will signal to your buyers is still a human judgment call, and that judgment is what you pay a designer for.
Where AI stops
Every second essay on my feed says taste is the new moat. Fine. Here is the more specific version I took away from Saptarshi, who uses AI daily inside his brand practice.
AI is a mirror, not a sparring partner. It reflects your own inputs back, pulled from everywhere, but it does not show up with a genuinely different opinion. Give it a first-principles question like "should this thing exist at all?" or "should we deliberately make this brand less polished to attract a more humanistic audience?" and it bounces back to the first generic idea. It seeks what the average of the internet makes, and the average pulls quality down.
So the real risk isn't that AI replaces you. It's that AI becomes an echo chamber. When you work only with AI, you stop seeing perspectives other than your own, because it keeps reflecting your inputs back at you. Taste can't grow in that loop. It grows the slow way with years of work, real audiences, and plenty of failures along the way.
So use AI for what it does well today, which is transcribing conversations, giving you a first read on research, and suggesting related ideas as you work. It handles that layer brilliantly. But it's an assistant, not a partner that challenges you. The final call on which identity will connect with your buyers and which won't still belongs to you.
Your buyer now sees three websites, not ten
Like I said at the start of this article, my own buying changed. Earlier I gave ten websites thirty seconds each, and the visual brand decided which ones earned more. Now I ask AI to scan the market and hand me a top three, and I spend ten minutes on each of them.
Fewer brands make the shortlist, and the ones that do get a much closer look.
I'm definitely not the only one who does that. Adobe tracked AI-referred visitors to retail sites converting 54% better than everyone else by mid-2026. These visitors (myself included) arrive late in the decision, pre-qualified by the machine, ready to judge.
This has two consequences for your brand.
First, you now design for a filter and then a human, in that order. Getting into the AI's top three is an evidence problem, not a volume problem. The Princeton GEO research found that citations, quotations and statistics lift a source's visibility in AI answers by up to 40%, while keyword stuffing performs worse than doing nothing. Do not jam your site with text. Make it dense with proof.
Second, the human at the end judges harder, and how hard depends on stakes. You will let an agent order your socks. You will not let it pick what you wear to your sister's wedding. The higher the value of the purchase, the more certain a human shows up at the end, now backed by AI, comparing you against two finalists instead of nine also-rans.
The sharpest line from my conversation with Saptarshi is the one to run your next brand review against.
"AI has effectively removed the layer of mediocre performance."
Saptarshi Das, on the GTM & AI Operators Show
Puja put the same idea in commercial terms. Anyone can now get a decent-looking logo, website, or deck in a few minutes, so "good enough" has no value left. What's still worth paying for is the thinking behind the design.
The market never rewarded passable design in the first place. McKinsey tracked 300 companies over five years. The companies in the top 25% for design grew revenue 32 percentage points faster than their industry peers and delivered 56 points more in shareholder returns. Among everyone else, the bottom 75%, the differences were small. In other words, being good at design but not great was already worth almost nothing.
That's the shift underneath everything in this piece. When buyers gave you thirty seconds, a brand only had to look good enough to earn a second look. Now they give three finalists ten minutes each, and ten minutes is long enough to notice whether a brand actually stands for something - or just looks like everyone else in its category.
FAQs
What is visual brand design?
It's the creative infrastructure of a company: the identity system of logo, typeface, palette, illustration, and every surface a buyer touches, all telling one consistent story about who the product is for before anyone reads a word.
When should a startup invest in brand design?
After conviction, before scale. If you can state exactly what the company is and who it's for, invest. If you can't yet, ship with something cheap and rebrand when clarity arrives, or make clarity the first step of the brand project. In B2B, your brand reaches the roughly 95% of buyers who aren't in-market this quarter. The sales deck only reaches the 5% who are.
Will AI replace brand designers?
Not the part that matters. AI handles transcription, research first-passes, and concept generation well, but it mirrors its inputs and defaults to the internet's average. Selection, cultural context, and first-principles calls like "should this thing exist at all" still sit with people.
Why do all B2B websites look the same?
Category defaults (trust-blue, dark gradients, AI purple), Dribbble and Pinterest feedback loops, and the risk appetite of founders and creative directors both. At enterprise scale convergence is also operational: complex identity systems can't be executed across thousands of teams and hundreds of geographies.
How do you brief a brand designer well?
Bring a stable description of your offering, a clear target audience, and a list of everything you don't want. Use AI mood boards to discover what your adjectives actually mean before the kickoff, and then expect a good designer to challenge the references you fell in love with.
With AI at my disposal, do I need a designer?
For the thinking, yes. AI can produce a decent-looking logo, site, or deck in minutes, and it's a useful way to show a designer what you like. But it can't tell you why something feels right or whether it fits your business, and it builds from what already exists. A designer works out what's behind your references, decides what to be bold about and what to keep simple, and pushes the direction past the familiar. That thinking is what you're paying for now that good enough is free.
What are the limitations of AI in brand design?
Three big ones. It copies the average, so its output looks polished but familiar, and every brand that relies on it starts to look the same. It can't judge fit, so it can show you what looks premium but not whether premium is right for your buyers. And it doesn't argue with you: it reflects your inputs back instead of offering a genuinely different view, which makes it a strong assistant and a weak creative partner.
AUTHOR
Builds and runs go-to-market motions for B2B companies rather than advising from the sidelines. First hire and zero-to-scale operator across HR, legal, and sales tech, with a track record of taking businesses from early customers to 200-plus in North America.
