Engineering-led companies typically hire their first designer late, and hire them for the wrong job. The brief is usually to make the product look better, which treats design as decoration applied at the end. What a good designer actually changes is what gets built — which problems are worth solving, what the flow should be, and which of the four features someone requested was the real need. Hired as a decorator, they will decorate, and the company will conclude that design is a nice-to-have.
The signal that you need one is not that the product looks dated. It is that users are confused, support is answering the same how-do-I question repeatedly, onboarding drops people at a consistent point, or the team is arguing about what to build with no way to resolve it except seniority. Those are design problems, and they persist regardless of how good the engineering is.
Involve them before the specification, not after. A designer given a finished list of requirements can only arrange them on screens. A designer present when the problem is being discussed will ask who this is for, what they do today, and what would change if we did nothing — and will frequently reduce the scope, which is worth more than any interface improvement. The most valuable output of design is often a decision not to build something.
For hiring, look at process rather than portfolio images. Ask a candidate to walk through a project from the problem to the outcome: what they were told, what they found out, what they tried, what they discarded and why, what they would do differently. Strong designers have opinions about the decisions and can explain the trade-offs. Weak ones describe the visual result. This is the same distinction as with engineers, and it separates candidates just as reliably.
Be careful about the first hire's shape. Product design, visual and brand design, and user research are related and distinct, and one person rarely does all three well. Most product companies need product design first — the person who works out flows and structures — with brand and research bought in as projects. Hiring a strong visual designer and expecting product thinking, or the reverse, is the commonest mismatch and it is nobody's fault when it disappoints.
Finally, give them access to users. A designer working from second-hand descriptions of what customers want is guessing more elegantly than you were. Even a handful of calls a month, or sitting with the support queue for an afternoon, changes what they produce entirely. Companies that treat customer contact as a sales function, and design as an internal one, get design that is confident and disconnected — which is worse than none, because it is more expensive to argue with.
