Recently, I was talking to a founder about her company’s brand evolution and remembered an article I’d written on this a few years ago, before AI and large language models were part of everyday workflows.
When I went back and reread it, what struck me was that the fundamentals still hold, but how founders approach brand work and what they should expect from agencies has changed meaningfully now that AI and LLMs exist.
I see this play out regularly with the companies we work with at Bonfire. Founders are asking a very real question now. How do you justify paying a brand agency when AI can generate a brand for significantly less?
So I rewrote it. This post reflects how I think about picking and managing the right brand agency today, informed by the same principles but updated for an AI enabled world.
Brand Hasn’t Changed, But the Process Has
At Bonfire, we begin investing at the Seed stage, when companies are still forming their identity. Early on, “brand” is usually a logo someone designed quickly, messaging written by the founders, and a website that does the job well enough. That’s not a bug - it’s exactly what should happen early.
As companies raise capital, hire teams, and start selling more seriously, they usually reach a point where their brand needs to evolve. I still prefer the term brand evolution over “rebrand,” because nothing is broken - your company has simply learned more about itself, its customers, and its market.
What’s different now is that founders don’t show up to this moment with a blank page. They show up with:
- AI-generated positioning statements
- Draft website copy written with ChatGPT
- Dozens of narrative variations created in minutes
This is both incredibly powerful and potentially dangerous.
AI can accelerate brand work, but it cannot replace the hard decisions that make a brand differentiated and durable. With that in mind, here’s how I think about picking and managing a brand agency today. These aren’t new rules so much as updated principles, grounded in what still matters, but adjusted for how founders actually work now.
1. Narrative Still Comes First - AI Just Gets You to the Starting Line Faster
In my original post, I said narrative is half the battle. That’s still true.
What’s changed is that AI now makes it easy to generate something that sounds like a narrative. The risk is mistaking speed for clarity.
LLMs are excellent at:
- Producing first drafts of positioning
- Pressure-testing whether messaging is understandable
- Exploring multiple narrative directions quickly and telling you which ones to toss vs keep
Where they fall short is original insight, market nuance, and making the hard tradeoffs about what not to say.
Before engaging a brand agency, founders should absolutely use AI to explore ideas. But they still need to answer the core questions themselves: Why this, Why us, and Why now. AI can help refine those answers, but it can’t generate conviction on your behalf.
At Bonfire, we help founders work through narrative deliberately, often via a structured workshop. AI can support that work, but it can’t substitute for it. If your inputs are generic, AI will simply help you get to a generic answer faster.
A brand agency should not be discovering your story for the first time during design.
Narrative alignment-human alignment-is still the biggest determinant of whether a brand project succeeds or stalls.
2. RFPs Matter More Than Ever in an AI-Enabled World
If anything, AI has raised the importance of a clear RFP. Use this practical Brand RFP template we put together to help founders and marketers scope a brand evolution and engage agencies more efficiently.
Agencies now vary widely in how they work. Some are deeply AI native. Others use AI quietly behind the scenes. Some barely use it at all. None of these are inherently right or wrong, but founders need to understand how an agency works before hiring them.
In addition to traditional scope questions, founders should now be asking:
- How do you use AI in your creative process?
- Which parts of the work are AI assisted versus human led?
- How do you ensure originality and defensibility in the final output?
- How will this brand system scale when our team uses AI to generate content post launch?
AI doesn’t eliminate the need for careful scoping. It makes sloppy scoping more expensive.
Anything assumed to be included in a brand project has a way of surfacing later as added time, cost, or frustration.
3. Be Disciplined About Agency Selection and Budget
A quick note on budgets and agency selection. You should only send an RFP to as many agencies as you would realistically hire. For most founders, that’s two or three at most.
The more agencies you engage, the more work you’re creating for yourself. Each agency will typically want two to three meetings before submitting a proposal, and that time adds up quickly. More importantly, wide RFPs tend to degrade signal. You end up comparing very different scopes, philosophies, and approaches rather than choosing between true alternatives.
There’s also a natural chicken and egg dynamic in these processes. Founders want the highest quality work for the lowest possible price, so they often wait for agencies to put a number on paper first. Agencies, meanwhile, are trying to scope work without knowing what’s realistic.
In practice, sharing a budget range upfront saves everyone time.
It helps agencies quickly assess fit and propose the right scope, and it prevents founders from comparing proposals that were never designed to be comparable. A budget range doesn’t lock you in. It creates efficiency.
4. Tight Decision-Making Is Even More Critical Now
One of the most common failure modes in brand projects has always been too many stakeholders and subjective feedback. AI has made this worse. Now, anyone can generate alternatives: “I prompted ChatGPT and got this version…” Access to tools does not equal accountability for decisions.
Before starting a brand project, it’s still critical to:
- Clearly define who the decision-makers are
- Agree on how feedback is collected and delivered
- Anchor decisions in strategy and research, not taste or trends
AI should expand thinking early, but converge decision-making fast. A small, empowered approval group is still one of the strongest predictors of success.
5. Brand Assets Should Live in Systems, Not Static Files
This is where the biggest practical shift has happened. Today, a brand isn’t just a logo and a brand book, it’s a system that needs to scale across website, sales materials, social content, and now, AI-generated drafts created by non-designers.
So founders should expect agencies to deliver:
- Modular design systems
- Editable templates in tools the team actually uses (Canva, Figma, Google Slides, Webflow)
- Clear rules that make AI-assisted content feel on-brand, not chaotic
A good question to ask agencies now is: “How does this brand system hold up when our team is using AI every day?” If the answer is “you’ll need us every time,” that’s a problem.
Final Thought
A good brand is still good business. AI hasn’t changed that. What it has changed is the pace, the surface area, and the number of people who can now participate in brand creation. That makes clarity, alignment, and strong partners more important, not less.
The best brand agencies in 2025 don’t just deliver beautiful assets. They help founders make hard strategic decisions, translate narrative into systems their teams can actually use, and build brands that hold up when content is being created every day, by humans and by machines.
If your brand only works when a designer is involved, it won’t scale. And if your agency can’t explain how their work survives contact with AI, it’s probably not built for where your company is going.
If you’re thinking through a brand evolution and want to compare notes, or want my short list of vetted agencies across budgets, I’m always happy to share.