Usually, that comes from a series of perfectly reasonable decisions: a logo when the company started, a website when it needed one, and sales material assembled as the business grew. Each piece solved a problem at the time, but together they can end up describing different versions of the same company.
By then, the product team may have developed its own visual language, the website is still explaining an earlier version of the business, and every new initiative needs another exception. None of those decisions were necessarily wrong. The problem only becomes visible when you look at them together.
Before drawing anything new, I want to understand what has changed and what is still worth keeping. A familiar mark may have years of recognition behind it and simply need a better system around it, while other parts of the identity may have been stretched well beyond what they were designed to do. The amount of change should come from that assessment.
There are usually a few decisions to work through with the team: what they want to be known for, which part of the business should lead the story, and whether a new product really needs its own identity. The questions sound simple until several people answer them and you realize they are describing slightly different companies. Agreeing on those things gives us something concrete to design around, and a better way to judge the work than whether someone likes the color.
For Principles of Intelligence, the challenge was to bring a growing family of initiatives into one recognizable identity. The organization had developed from a research fellowship into a broader platform for AI safety research and programs, so the relationship between those parts needed to be visible without making them all look identical.
I built a family of marks around a single geometric principle that could be taken apart and recombined. Each program could have its own mark while keeping a clear connection to the organization, and we had a way to approach the next addition without starting from scratch. The website and applications could then build on the same thinking.
I still enjoy finding that kind of idea after twenty years of designing logos, especially when something I like as a mark also helps solve a practical problem. In this case, it gave a growing organization a way to present its programs as a family.
What interests me after the presentation is whether the team has something useful to work with when the next product page, presentation, or program launch comes along. There will always be decisions that need a designer, but the team shouldn't have to work out what the company looks like every time it makes something new. I want the identity to make that next decision easier.