BLUE LLC

BLUE LLC is an M&A advisory firm working in TransTech, transportation, logistics and supply chain, representing both the operators being sold and the private equity firms and strategic acquirers buying them. Its site has exactly one job: convince a qualified prospect, a founder considering selling the business they built, or an acquirer with capital to deploy, to book a confidential consultation.

The engagement was a redesign of bluellc.io, run through an AI-based workflow: research, art direction, design and build as one continuous activity by one person, with Claude Code as the working layer and the product and research foundation kept as a structured, sourced knowledge base in the repository.

The existing site was a hand-written static page that undersold the firm: template visual language, 14px body type, and a contact form that posted to a script endpoint in a mode that reports success whether or not the submission was ever delivered. For a business whose entire site exists to produce one booked consultation, an unverifiable lead path is not a technical footnote, it is the whole risk, and it had gone unnoticed precisely because it never visibly failed.

The design problem was differentiation under constraint. The copy was to be reused verbatim, this was a visual and structural redesign, not a messaging rewrite, and the structured data and content had to carry over. Within that, the site had to escape the category's default look without tipping into the opposite failure, the over-designed agency page that makes a discreet advisory firm look like it is trying too hard.

The art direction is the heart of the piece: ink on paper. Warm cream stock, deep marine ink, a single marine-blue accent drawn from the logo, and engravings of the client's actual subject matter, ships, rail, harbours, run at real scale in full ink, bleeding to the viewport edge. Engraving is a print medium, so the page behaves like a printed page: no gradients anywhere, one accent, no glow, no glass. Rhythm comes from two full-bleed inverted bands, paper-on-ink, so the register flip does the dramatic work that competitor sites ask colour effects to do. Typography is a transitional serif for display over a neutral grotesque for body, self-hosted, with body type raised from the original 14px to 17px.

The direction was not the first answer. An earlier same-day pass settled on a contained navy-and-blue option and was rejected on sight for reading as a restructure of the old site rather than a redesign, which the options document had named in advance as exactly that option's risk. The rejection, the reason, and the surviving direction are all recorded in the decision log, and a later first build pass failed in a documented way too: the engravings were initially shrunk into bordered boxes at low contrast, where they read as faint technical diagrams, and the fix, real scale, full ink, edge bleed, is written down so it does not regress.

In a category where every site looks the same, the art direction is the conversion strategy. The audience for an M&A advisory does not compare feature lists; they infer judgement from every surface they touch, and a template signals its absence. Committing to a printed-page discipline, and enforcing it with outright bans rather than preferences, is what let the site claim a position no competitor occupies. The bans mattered more than the palette: a direction is only distinctive while nothing on the page contradicts it.

Cheap variations are the hidden hazard of an AI workflow, and the decision log is the counterweight. When another design pass costs minutes, nothing structural stops a project from cycling through directions indefinitely, and the work degrades into taste churn. Writing decisions down with their reasons, and preserving rejections rather than cleaning them up, is what made decisions stay decided. The rejected contained-blue option earning its predicted failure on sight was the log paying for itself: the risk was named before it was demonstrated.