Generative AI has emerged as one of the most significant technological advancements for marketers in recent years. With large language models from companies like OpenAI, Google, and Meta powering tools such as ChatGPT, Claude, Gemini, MidJourney, and Stable Diffusion, generative AI has transformed how marketing content is created and campaigns are executed. This article explores the current state of generative AI in marketing, examining the various platforms available, their applications, and the challenges marketers face when implementing this technology.
Generative AI is a form of artificial intelligence that uses sophisticated algorithms to create new content, including text, images, and even video or music. These systems analyse existing data to identify patterns and then generate fresh content based on those patterns. The primary goal of generative AI is to create content that is indistinguishable from that produced by humans, offering marketers powerful tools to streamline content creation and personalisation efforts.
ChatGPT has become one of the most widely used generative AI tools, accounting for over 60% of traffic across all AI tools according to a Writerbuddy study. Released to the public in November 2022, it became the fastest-growing platform in history until Meta's Threads surpassed it, attracting a million users in just five days and reaching over 100 million users within two months.
ChatGPT's key benefits include:
Despite Microsoft's $10 billion investment and integration with Bing search, ChatGPT faces challenges including limited customisation options and difficulties in verifying the accuracy of its outputs or identifying source information.
Google's conversational AI tool, initially launched as Bard in February 2023 and renamed Gemini in February 2024, functions similarly to ChatGPT but with unique features tied to Google's ecosystem. Gemini can access data from other Google applications such as Maps, Gmail, Docs, Drive, Flights, and YouTube, allowing users to apply the tool to their own content.
A key differentiator is that while ChatGPT's responses are based on data available up to its training cutoff, Gemini has the capability to use more current, up-to-the-minute data.
Created by former senior members of OpenAI, Anthropic's Claude operates with a text-based chat interface similar to ChatGPT. While it doesn't search the web to form answers, it can understand text at URLs provided within prompts. Claude supports multiple programming languages as well as ten human languages, offering flexibility in the inputs it can process.
DALL-E, developed by OpenAI, generates unique images from text prompts using a combination of machine learning algorithms and neural networks. Popular in the creative industry, DALL-E creates high-quality, unique images based on natural language requests.
Its benefits include:
These tools use deep learning models to generate images in various styles, including abstract, impressionist, and expressionist. Their benefits include producing high-quality images that are often indistinguishable from human-produced work and offering user-friendly interfaces.
A significant challenge with these platforms involves determining the source of imagery and potential copyright infringement. In January 2023, three artists filed a lawsuit against Stability AI, Midjourney, and DeviantArt, claiming that these AI tools were trained on images without creator consent.
As a more specialised tool, Secta.ai generates AI headshots based on existing photos. Users provide 25-30 images, select background settings and facial expressions, and the system generates new headshots meeting these requirements. Unlike broader image generators, Secta focuses solely on headshot creation.
Video generation requires significantly more processing power than text or image generation, as a video essentially comprises numerous still images. A 30-second video at 24 frames per second requires generating the equivalent of 720 static images.
Capcut offers video manipulation features including background removal, video resizing, speech-to-text for captioning, and text-to-speech functionality. Available as both web and mobile applications, Capcut prioritises ease of use but may have limitations in customisation.
Synthesia transforms text-based prompts into videos featuring AI avatars, supporting over 130 languages. This platform makes customised video creation nearly as simple as creating a PowerPoint presentation, which can be particularly useful for sales teams seeking personalised content or marketing teams needing engaging alternatives to static presentations.
Though not yet publicly available at the time of writing, OpenAI's Sora has been previewed as a potential game-changer in generating photorealistic, high-definition video. Sora uses visual "patches" to approximate full-motion portions of video, creating what appears to be footage from an animated world.
These platforms offer multiple output types from single prompts, capable of creating text, images, and helping with campaign planning and execution.
Tailwind describes itself as a marketing "copilot," assisting with writing for multiple channels and generating images to support marketing campaigns. It also helps create marketing plans with scheduling recommendations for different tactics.
Targeted specifically at marketers, Jasper helps teams write content more quickly, translate it into different languages, and generate new ideas. Rather than replacing human writers, it positions itself as an augmentation to existing marketing teams.
This specialised tool transforms audio or video content such as podcasts or recorded speeches into various content types, including shorter video clips, blog posts, and social media content. Unlike many generative AI tools, SwellAI primarily works with non-text inputs, though users can provide guidance through text inputs and menus.
These platforms cater to organisations with large teams, extensive data privacy needs, and greater risk concerns regarding publicly released content.
Beyond its consumer-facing versions, OpenAI offers an enterprise-grade ChatGPT that protects information from being used to train its models. It allows users to create customised GPTs with specific tones and word choices, trainable from existing writing samples.
Positioned exclusively for enterprise companies, Writer enables content creation with governance constraints (privacy, legal, regulatory) and effectiveness analysis within a single platform. It uses its proprietary large language model, Palmyra, and focuses on reducing "hallucinations" by training on company information without sharing that data with other organisations.
Many existing platforms that marketers already rely on have integrated generative AI capabilities:
Adobe's Creative Cloud now incorporates its AI tool called Firefly into products like Photoshop, Illustrator, and Adobe Express. These features allow users to generate images from text prompts, remove backgrounds, isolate image portions, and more. The primary benefit is maintaining existing workflows for teams already using Adobe products.
A popular Adobe alternative, Canva has released its Magic Design™ tool to simplify creating design variations across different marketing channels including social media, emails, ads, and website landing pages.
As a leading CRM provider, Salesforce has incorporated Einstein GPT into its products, enabling marketers to send more personalised emails and allowing salespeople to automatically summarise customer interactions.
HubSpot, popular among small and medium-sized businesses for CRM and marketing automation, offers generative AI capabilities for creating emails, landing pages, and customer service knowledge bases using text prompts. HubSpot currently uses OpenAI's large language models for its generative AI functionality.
Despite its relative immaturity compared to other marketing technologies, generative AI offers several powerful applications:
Enterprises using generative AI must manage several critical concerns:
Retrieval Augmented Generation (RAG) offers a potential solution to these challenges by limiting generative AI inputs to only what an enterprise specifies. This approach provides:
Generative AI in marketing is still in its early stages, but it offers significant efficiencies when human review of text and image content is maintained. Marketers should proceed with caution, being mindful of potential pitfalls such as unauthorised use of copyrighted material, nonsensical outputs, or results based on false or offensive source material.
With major software providers rapidly adopting AI features in legacy tools, marketers should balance experimenting with new platforms against ensuring AI implementations work well within existing workflows. Adopting too many disconnected AI platforms can create challenges in determining authorship, ensuring ethical data usage, and addressing other potential issues.
Successful implementation requires balancing these concerns, with guidance from legal and technology teams where appropriate. As generative AI continues to evolve, marketers who thoughtfully integrate these tools into their workflows stand to gain significant competitive advantages in content creation, personalisation, and campaign execution.