Design and Bots

Art. Lebedev Studio has developed technologies that magically bolster human abilities. Dazzling brand identities, seamless wayfinding systems, unforgettable advertising campaigns and intuitive interfaces are now created with the help of bots.

Design at Art. Lebedev Studio is now done by engineers.


A fresh look

Neural networks offer an unlimited number of fresh perspectives and can create thousands of new meanings. Engineers set the parameters and the networks come up with fascinating (and sometimes entirely unexpected) results.

Nikolay Ironov

Art. Lebedev Studio developed an AI designer and for over a year was passing it off as a human. To date, the network has been successfully used in over 20 of the studio’s commercial projects.

Stylish bots

Humanity has already wasted enough millions of hours on brandbooks and “how to correctly place the logo on documents” illustrations. Engineers can now build bots that churn out pitch-perfect mock-ups in the blink of an eye.

Media libraries

Design can be instantly tested on any number of formats. Create a new logo, then click a button to instantly see how it would look on a coffee cup, a billboard, a t-shirt or a car, on Twitter or on a bank card.

Rendering selected options

“Just throw the pictures and logos into the system, it will generate variants of trucks for you and we’ll choose the best one.”

Self-managed mock-ups

The system can modify mock-ups based on the data it receives. It can choose the best layout, the right number of columns and the perfect size of illustrations and typeface. These technologies have long been used on the web, but are unique in graphic design.

Auto-typesetting brandbooks

Brandbooks can now do more with bots behind them. They can be interactive, produce mock-ups, calculate costs and order prints.

High-tech design

Machines took on 90% of designers’ work. And that’s a good thing.

“The automatic typesetter complained that line spacing in the mock-up exceeds word spacing. And even attached a ‘weak composition’ tag.”

“Two text design rules violated.”

“The lower part of the body is usually covered with dirt and the mock-up has text there.”


  • Information planning
  • POI
  • Readability
  • Graphics
  • Media

Strong links

Our advantage in making wayfinding systems is our graph that connects signs, buildings and objects around them into information networks.


“The taxi desk has finally been moved from the second floor to the first. All airport maps and signs will be updated before lunch. We’ll reprint the signs in the evening and start hanging tomorrow”

Technological wayfinding

Maps and wayfinding signs in airports 🛫, parks 🌳, stadiums 🤾‍♀️, railway stations 🚃 and the metro 🚇 are typeset and filled with content by machines.

Retypesetting in bulk

“The Aleksandrovskaya—Ekaterininskaya hub opens July 1. We’ll typeset all signs at the hub from scratch and retypeset them at all other stations.”

Calculation of object importance

“The high rise shouldn’t even be here, it has lower priority than other items”

Merging information from different transport systems

“The city started sending information on ground transport routes that operate near park entrances. Can you add a block to the signs so that bots can populate it?”

Smart wayfinding

“The high rise won’t appear on this sign because the sign knows it’s located so close to the building that you can actually see it.”

Readability checking

“According to our metal friend here you won’t need to have perfect vision to see the sign comfortably from this distance.”

Designers engineer design

Machines resize media, tag content and locate mock-ups by keywords.

“We created mock-ups for a couple of signs manually, but they came back from autotagging with the message ‘Bot not OK.’ I’ll take a look at what exactly it didn’t like and fix it if I have to”

“Look for mentions of number 17 in the system, it’ll give you all the JPEGs that have it”

“We got back the results of the pilot design test, everything looks good. We’ll start autotypesetting tomorrow and will produce schedules and maps for new bus stops”

“All signs inside the Boulevard Ring have finished generating, do you want to have a look at the list of differences?”


  • SKU
  • Product passport
  • Information highway
  • Barcode
  • Marketplace

Designing for retail presents a unique set of challenges. Huge systems need to be kept in working order and a huge number of operations need to be made.

Here we need to update hundreds of package mock-ups, verify endless lists of ingredients and bar codes, translate texts, warn about possible errors and roll out designs to marketplaces.

Online stores

A large online store is a essentially huge bot that helps buy and sell millions of products. It requires a few dozen designers or just one high-tech solution that would generate an insane number of banners, pictures and descriptions.

Wildberries banner generator

Designers used to create each banner manually. Now an engineer designs a banner-producing machine.

Variable design

 and interfaces 

Creating design generators

Improving the appearance of a website with the help of a designer is expensive. It makes much more sense to design, tune and occasionally adjust an automated design system.

Mendeleev Periodic table of elements

It’s a construction kit that helps everyone get their own version of the periodic table with the data they need.


A set of stickers that updates itself once every five minutes. Each time a completely new sticker pack is generated, unseen before and never to be seen again.

Large systems

Maintaining huge systems requires efforts of dozens of people. Interfaces are an example of such systems. The good news is that they have incredible capacity for automation and contain many pages, scripts and paths that can be calculated mechanically.

Designers only do what they do best: design.

Bottom line

Design is now done by engineers

Request a design...