CLEVER°FRANKE’s identity for ADC is a living, AI-driven system

The Brand Identity: Plus signs, crosses, dots, dashes, triangles and diamonds are the working vocabulary of ADC, an Amsterdam-based AI transformation consultancy whose teams spend their days inside formulas, embeddings and data models. When CLEVER°FRANKE took on the rebrand of their parent company, those mathematical operators became the building blocks. The resulting identity generates a different expressive pattern every time anyone at the company needs one, drawn from a custom AI model trained on ADC itself. ADC’s previous identity had drifted out of step with the business.

A palette of light brown, beige and olive green sat alongside a typeface used in its lightest weights, reading as soft and mild when the consultancy’s promise to clients was anything but. “The old one wasn’t reflecting the boldness and transformative promise ADC was making,” explains Bart Simons, Creative Coder at CLEVER°FRANKE, relaying the brief from ADC’s then-CEO Rik van der Woerdt.

The new system needed to hold together a difficult pair of qualities: technical credibility and analytical rigour on one side, warmth and human character on the other. “ADC is also a collective, and we knew precision alone would feel cold and sterile,” adds Visual Designer Letizia Agosta. “The breakthrough was building a system that’s rigorous in its logic but intentionally loose in execution.” That principle runs through the pattern system CLEVER°FRANKE built, where mathematical symbols morph and cluster according to proximity, density and context.

Quadtrees handle the spatial logic beneath the surface, a generative tool written in JavaScript runs on 23 trained parameters within a broader set of more than 50, and the compositions that emerge feel organic despite being entirely computational. Teaching the AI model what to do with context was the part that took the most time. CLEVER°FRANKE manually mapped more than 100 data points covering ADC’s topics – HR, culture, values, fields of operation, clients, deep tech operations – and visualised the embedding space to confirm that those points were spread widely enough to avoid collapsing into a handful of repeating looks.

Hundreds of patterns were then generated and matched to topics by design feeling, with deep tech subjects associated with densely nested, highly complex compositions and cultural topics pulled in different directions. Those associations are what let the model now translate a written prompt into a specific visual. For the users at ADC, the tool itself strips away the complexity sitting beneath it.

Someone describes what they need a pattern for – a Q1 finance report cover, a LinkedIn post or a cultural update, for example – and the tool places that input inside ADC’s embedding space, offering three variations drawn from slightly different neighbours. Picking a favourite feeds back into the model, sharpening it over time. There are no granular sliders to adjust, which changes how people work with it.

Article truncated for readability. Read the full piece →

The introduction of an AI-driven identity system for a brand is a significant advancement in design strategy, showcasing a novel approach that is highly relevant for brand professionals looking to enhance user engagement.