How a creative businesses are using A.I. today

The media and trade press constantly talk about how creative services businesses are rapidly adopting AI tools like Midjourney, ChatGPT, and DALL-E to enhance their ideation, design and production processes. Something we thought was missing in all of these write ups, were specific examples of HOW these businesses are integrating generative AI.

This post covers some interesting implementations of AI from agencies around the world (taken from a Creative Boom article) as well as showing how I've integrated AI into the strategy and design workflows of itswhatsnext.

Here's the primer on how creatives are using A.I. today:

Idea Generation and Visualisation

One of the primary uses of AI in creative workflows is for idea generation and visualisation. Tools like Midjourney allow our designers to quickly visualise concepts and communicate intangible ideas to clients.

For example, Poppins (https://poppins.agency/) is using Midjourney to generate 3D scenes they can place designs into, such as a hand holding a mobile phone to Brutalist architecture we'd like to place a car in. Situations like this where a client has product imagery but not the broader scene, generative AI can works wonders.

Image created by Wonder using Midjourney

At IWN, we've used Dalle to generate digital backgrounds, 3D renderings, and product scenes in various styles to accompany design concepts and flesh out brand assets like our website, and new business proposals.

Art Direction and Experimentation

AI image generators have become invaluable for art direction and experimentation. AI can be used to create preliminary visuals, ensuring the client is aligned with the proposed direction. For example, in a project with a bank, we could use AI-generated images to define an art direction capturing aesthetic and personality before having photography be shot.

A core use case for AI at IWN is to create mood boards. We recently kicked off a rebrand with a business and used Dalle to build the foundational aspects of a mood board - we then used traditional tech (Figma, etc), to build the full brand IP. AI can be tapped in again to fill gaps when suitable imagery can't be found (especially for free) through prompts creating styles, photographic effects, and emotional qualities.

AI generated imagery I-AM used to inform the thinking of their mood boards.

Content Summaries and Research

Large language models (LLMs) like ChatGPT are proving incredibly useful for research and content summarisation. The most common use case of any LLM is quickly synthesising complex reports, news articles, etc - enabling more efficient ideation and strategic planning. For example, using Claude.AI to summarise large reports into short synopses for client presentations.

Design Workflow Automation

LLMs have seen a massive shift in how I structure a design sprint. We've used ChatGPT to draft IA that has been converted into wireframes by the tool UIzard. From UIzard we can integrate figma and begin customising individual aspects in the design. These efficiencies can result in up to an 80% reduction in implementation hours, which can be rerouted to ideation and experimentation.

Generative UI from UIzard

Streamlining CRO with AI-Powered Automation

We have recently started putting together proposals outlining how we can use automation and AI to drive down efficiencies in conversion rate optimisation projects.

By integrating tools like HotJar, PPC ad platforms, and Google Analytics with ChatGPT using Zapier, we've created a streamlined workflow for data-driven CRO sprints.

The initial data collection is done in GA, and it's connected to GPT via custom Zaps with user behavior data from HotJar's heatmaps etc manually uploaded into the chat instance. The data is fed into ChatGPT, which generates insights and hypotheses for improving website conversions. If set up correctly, ChatGPT can also assess performance data from PPC campaigns and Google Analytics to identify areas for optimisation.

This approach slashes CRO planning time by about 80% so we can rapidly ideate and prioritise test ideas. ChatGPT can then draft copy variations, which are then reviewed and refined. For design tests, AI image generation can be used to visualise concepts before implementation.

The future of AI in creative service industries

AI is being used for a wide range of tasks in creative agencies, but it's not about how much AI is being integrated, rather, it is whether or not is being integrated in a smart and thoughtful way. Rob Skelly (creative director of Born Ugly) "Like any tool, it's important that this usage is meaningful and grounded with a big idea to avoid it becoming a gimmick. Heinz is a great example of using AI in a really clever way to celebrate the strength of as an icon."

Whilst we don't know what will happen with AI, it's almost a definite that it will grow into a bigger feature of agency life for some years to come. Some rules to adhere to that will help guide you through a translucent era in marketing:

1. AI is a collaborator, not a solver of problems

2. AI will benefit you most when there's a human layer included

3. Try and focus on how AI will create efficiencies, not end results

4. Be conscious of sharing business and personal data

5. Experiment, experiment, experiment