Generative AI

Generative AI - The Swiss army knife for marketers and creatives

Not a day goes by without reading various reports on the topic of artificial intelligence, primarily generative AI.

Generative AI refers to artificial intelligence models that are able to generate data similar to that with which they have been trained. Instead of simply analyzing data or making classifications, generative models can create their own content, such as text, images, music or even videos. The input can be in text, image or voice form.

Development phases of generative AI

The development of generative AI has made remarkable progress, especially in recent years. Starting with early models such as Variational Autoencoders (VAEs) and Generative Adversarial Networks (GANs), research has rapidly produced new, more powerful models. OpenAI's GPT series, in particular the GPT-3 and GPT-4 models, are examples of the latest developments. Gartner formulates the most important development phases of generative AI as follows:

2010 - Near-perfect natural language translation

Around 2010, AI researchers working on natural language translation discovered that models exposed to large amounts of text performed far better than models using top-down grammar rules.

2014 - Mastering the meaning of words

In 2014, language models began to understand the meaning of words in a natural language by analyzing the context in which the word appeared.

2017 - 2022 - Large linguistic foundation models

The progress from 2017 to 2022 resulted in language models that can serve as a basis for adaptations. Building foundational models is costly, but once created, they can be optimized with a small amount of additional data to achieve peak performance on new tasks without significant investment.

2022 - Conversational large-scale linguistic base models

This year marked the arrival of ChatGPT, which provided users with easy access to a large linguistic base model. The brilliance of ChatGPT is not just in the incredibly advanced model; equally, it is the ability to access this model by speaking to it in natural language. As AI researcher Andrej Karpathy jokingly remarks: "Now the hottest programming language is English!"

Meanwhile, the spread and acceptance of generative AI is enormous. In March 2023 ChatGPT-4 was released, a further developed and improved version that has brought generative AI into the mainstream and further accelerated its rapid spread.

What are the benefits of generative AI?

The benefits of generative AI include faster product development, improved customer experience and increased employee productivity. However, the exact benefits depend on the use case. However, it is always important to bear in mind that generative AI cannot perform magic and that the well-known saying "Shit in, shit out!" applies here in particular.

Generative AI produces artifacts that can be inaccurate or biased, making human review essential and potentially limiting the time saved for employees. Gartner recommends linking use cases to KPIs to ensure that each project either improves operational efficiency or creates net new revenue or better experiences.

In a Gartner survey of more than 2,500 executives, 38% said that customer experience and retention is the primary purpose of their investment in generative AI. This was followed by revenue growth (26%), cost optimization (17%) and business continuity (7%).

What can (generative) AI do now?

We would like to present this using Adobe tools. Current studies show that artificial intelligence can be of great benefit, particularly in the provision of appealing shopping experiences in the course of the growing importance of the digital experience not only in B2C, but also increasingly in the B2B environment. Adobe provides a wide range of tools for this purpose, which are available under the so-called Experience Cloud in which AI now plays a key role.

The following video shows a first insight into Adobe's own AI called "Sensei", which forms the foundation for all of Adobe's AI tools and features and is part of the Experience Cloud:

However, AI has not only been relevant for Adobe since the current hype. Back in March 2018, Adobe and Nvidia announced an important partnership in the field of AI at the Adobe Summit. Nvidia is considered the global leader in AI computing. Building on years of collaboration, the companies have been working together to continuously optimize the Adobe Sensei AI and ML framework. Through this collaboration, new and more powerful tools and functionalities can be provided for the Adobe Creative Cloud and the Experience Cloud and are also being introduced here.

AI for analysis purposes

Artificial intelligence (AI) is playing an increasingly important role in data analysis and especially in anomaly detection. The aim here is to identify unusual patterns in large volumes of data that do not correspond to expected behavior. Modern AI systems use machine learning, primarily deep learning, to analyze huge amounts of data and identify complex patterns and correlations.

The Anomaly detection of the Experience Cloud makes it possible in this case to distinguish "real signals" from "noise". It also helps to subsequently determine potential factors that have contributed to these signals or anomalies. In this way, you can determine which statistical fluctuations are relevant and then identify the cause of a real error. It also provides you with a reliable metric prediction (KPI).

Below are a few examples of anomalies that are detected by the AI in Adobe Analytics and that may require intervention:

• Significant differences or deviations in average order value
• Peaks in orders with low turnover
• Spikes or significant deviations in registrations
• Significant outliers in landing page views
• Peaks in video buffer events
• Peaks in low video bit rates

Adobe's generative AI can also perform the following analysis tasks, among others:

• AI Assistant - Asks ad-hoc questions in natural language to query data. This allows insights to be gained quickly without the need for help from the analytics team.
• Intelligent Labeling - Allows insights to be generated in natural language so internal teams can take advantage of data storytelling.
• Trend detection - Makes trends visible in line visualizations.
• Text-based insights - Provides contextual information from which even inexperienced users can derive insights

AI in content and collaboration

AI-powered content intelligence allows you to optimize cross-team content workflows, repurpose content for maximum engagement and deliver effective personalized experiences that inspire action.

With Adobe GenStudio provides employees across departments and divisions with natively embedded generative AI that enables a seamless content supply chain Content supply chain. For this purpose, well-known Adobe tools such as the Creative Cloud with Adobe Workfront as a project management tool and the Experience Cloud with tools such as theAdobe Experience Manager Assets as DAM and Adobe Analytics as an analysis tool.

Adobe Experience Manager Assets uses a self-learning algorithm to create descriptive tags that help you find the right asset in just a few clicks. Intelligent tagging - Adobe calls this Asset Intelligence - is based on Sensei, the Adobe framework for artificial intelligence and machine learning. Sensei can be trained to recognize and automatically assign standard and company-specific tags to images, video and text-based assets.

Adobe Smart Cropping - a feature in Adobe Experience Manager Assets - saves users hours of editing time with the intelligent AI-driven cropping tool. The software automatically recognizes the focus point in each image or video and crops them appropriately. This means that images can be automatically generated for a wide range of output channels and sizes with the appropriate image section.

AI in e-commerce

Since according to a Deloitte study 69% of consumers are more likely to make a purchase with brands that offer a personalized experience, it is important to take this into account in your commerce offerings.

AI-supported product recommendations and live search results as well as real-time views of customer search behavior make it possible in Adobe Commerceto provide a relevant and personalized customer experience for B2C, B2B and hybrid commerce that drives conversion and customer loyalty. AI-supported merchandising tools increase the productivity of e-commerce employees.

Adobe has also recently introduced "Intelligent Category Merchandising", which is integrated into Adobe Commerce as standard. This feature uses AI to automatically rearrange products on e-commerce websites in real time based on user preferences. This way, the most relevant products for each shopper are presented as they browse.

Analysis shows that almost 60% of store administrators spend at least 20 hours/week on manual merchandising activities. The new feature drastically reduces this effort and still offers customers a personalized shopping experience.

AI to optimize the customer journey

With the advanced predictive insights provided by the Adobe Experience Cloud can be used to develop end-to-end customer journeys based on customer behavior and interaction preferences, from email campaigns to social media surveys.

AI helps to identify and acquire exactly the right customers, retain existing customers and optimize the delivery of experiences to improve campaign ROI.

With the help of AI-based, predictive lead and account scoring in the Adobe Real-Time CDP it is possible to predict how leads and accounts are likely to move through the customer journey. This makes it easier for B2B marketers to focus on the interested customers with the greatest potential.

Adobe Sensei GenAI for Real-Time CDP makes smarter, faster work easy. The AI can perform the following tasks:

• Audience creation and activation: recognizes missed segments and automatically creates new audiences.
• Generative playbooks: Extends templates for use cases by simulating customer journeys based on past campaign performance and profile preferences.
• Segment refinement: Uses insights from dialogs to continuously integrate and improve target group definitions and results.

In Adobe Marketo Engage supports marketing teams in directly engaging, inspiring and retaining customers by using advanced lead nurturing and trigger functions for the right content.

AI for creativity

With Firefly adobe introduced its own generative AI in spring 2023, which has developed and spread enormously. Firefly can be used to generate images based on text descriptions and other style specifications, remove objects from images with a brush or add elements using a text description, apply styles or textures to text using a text description or generate color variants of vector graphics using text input.

Further features such as "3D to image" or "text to vector" are currently under development and will be available soon. Firefly will be continuously developed and successively integrated into existing Adobe tools such as Photoshop or Illustrator in order to increase efficiency in daily creative work.

Firefly is trained with Adobe Stock images, openly licensed content and other domain content for which the copyright has expired. Millions of licensed images from Adobe Stock are among the highest quality on the market and ensure that Firefly does not generate content based on the intellectual property of other people or brands. Therefore, commercial use of generated images is also possible.

But things are moving forward in leaps and bounds! With the Project Stardust adobe is working on using AI to edit existing images as comprehensively and intuitively as possible afterwards.

AI in support

With Adobe Dynamic Chat is a native communication and engagement channel that enables customers and interested parties to help themselves and qualify themselves. Directly in Adobe Marketo Engage and powered by Adobe Sensei GenAI, Dynamic Chat is designed to move buyers into the sales funnel faster. According to recent studies today's B2B buyers want to self-serve before contacting sales, which means your website remains an important demand generation tool and key touchpoint for support requests.

With Dynamic Chat's end-to-end conversation automation, website visitors have multiple ways to engage with your website on their own, book appointments and get answers to their questions. This allows support staff to focus on the really important issues, namely providing the most personalized customer service possible.

AI for document processing

Documents, forms and workflows are becoming increasingly digital. AI makes it even easier to create and manage them. Adobe Sensei supports functions in Adobe Document Cloudto optimize all form-based processes and provide a smooth and enjoyable document experience for your customers.

Customers today expect simple options for Scanning and approval of documents, as well as uncomplicated, smooth digital signing processes. With the Adobe Sensei AI capabilities provided in Adobe Document Cloud, you can meet these demands no matter where your customers are.

The Adobe Experience Cloud tools at a glance

The Adobe Experience Cloud is a set of leading analytics, content, commerce and personalization tools that deliver engaging and seamless customer experiences across all relevant touchpoints. The Experience Cloud tools are regularly recognized by leading analysts.

The individual solutions of the Experience Cloud can be used "standalone" or in a comprehensive and linked manner via intelligent interfaces, allowing even more benefits to be achieved.

With the exception of Adobe Firefly, which is part of the Creative Cloud, and the example of AI in document management, which is part of the Document Cloud, the examples mentioned are all integrated into Experience Cloud applications.

Adobe study on generative AI from a customer and company perspective

To understand how generative AI can change both customer expectations and the creation of unique shopping experiences on the corporate side, Adobe conducted a series of studies between February and May 2023 Studies and surveyed over 2,000 consumers and 498 marketing professionals in the UK.

Over a third (39%) of all consumers surveyed believe that generative AI will enhance their personal creativity. The figure increases as respondents get younger: 55% of Gen Z consumers say generative AI will make them more creative and when it comes to their experience with brands, 62% of consumers surveyed say generative AI will improve their customer experience, with seven in ten Millennials (72%) and Gen Z (70%) even more optimistic.

When it comes to the most important things companies should do when using new generative AI technologies, the issue of responsibility ranks first among consumers. 23% of respondents would like to see measures such as the implementation of guard rails to promote the responsible use of artificial intelligence. 13% of consumers said it was most important to use generative AI to improve customer experiences 7% of respondents said that the most important reason for companies to adopt generative AI is to make the company more financially successful. Only 15% stated that companies should not use generative AI at all.

The Content Authenticity Initiative Content Authenticity Initiative (CAI) is an example of industry-wide guard rails. With more than 1,500 members, the CAI is committed to open global standards and technologies, including content credentials. This is a digital mark of origin for content that allows consumers to see exactly how and where content was created.

The majority of marketing and CX professionals state that they will use generative AI in their future work. Almost eight in ten respondents (79%) say they have already used some kind of generative AI tool, with 57% having tried conversational bots like ChatGPT and 44% having experimented with image generators like Adobe Firefly. Almost all (94%) professionals surveyed believe their organizations will use generative AI in their future work.

Generative AI will increase efficiency and make customer experiences more personalized. Marketing and CX decision-makers believe that generative AI will help them in many ways:

• better work results (88%)
• better and more content (87%)
• get more work done (85%)
• Make better use of creative tools (85%)
• better personalization (89%)
• Identify new target groups (88%)
• Reach the right customers (87%)
• better identification of customer experiences (87%)

The topic of personalization in particular is also discussed in another Adobe study "State of Digital Customer Experience 2023" as a key success factor for the future, where AI can help massively.

When asked about the most important ways in which companies should use generative AI, marketing and CX managers cite improving the quality of their products and services in first place, followed by providing great customer experiences and working more efficiently in third place.

Generative AI will play a major role in the future, particularly in the area of content. Here, marketing and CX professionals see the following advantages in particular:

• faster generation of content
• Optimization of content
• Generation of more content

In any case, generative AI will be crucial to Content supply chains and streamline content supply chains and help brands worldwide to meet the growing demand from customers for up-to-date and engaging content. Current estimates assume that the demand for content will multiply in the coming years. This requirement can only be met through strategic content supply chain management.

However, there are also legitimate concerns. While most marketers are optimistic about the benefits of generative AI, there are also the following points of criticism:

• Quality of generated content
• Lack of transparency about the data input of the AI models
• potential risks in the area of copyright

Overall, the results of the study indicate that generative AI has a promising future for both consumers and companies. Most customers and decision-makers on the corporate side are ready and eager to use generative AI to improve products, services and experiences. It is now up to companies to harness this technology to meet both the opportunities and the expectations.

Forecasts for generative AI

The outlook for the AI market is extremely promising from a business perspective. Statista provides the following the following forecasts:

• The market size in the Generative AI market will be around € 42.17 billion in 2023.
• The market size is expected to have an annual growth rate (CAGR 2023 - 2030) of 24.81 %, resulting in a forecast market volume of € 198.90 billion in 2030.
• In a global comparison, the largest share of the market size is expected in the USA (€ 15,160.00 million in 2023).

And the analysts at Gartner see the following developments in the coming years:

• By 2024, 40% of enterprise applications will have embedded conversational AI, compared to less than 5% in 2020.
• By 2025, 30% of organizations will have implemented an AI-enhanced development and testing strategy, up from 5% in 2021
• By 2026, generative design AI will automate 60% of the design effort for new websites and mobile apps
• By 2026, over 100 million people will train robot colleagues to simplify their work.
• By 2027, almost 15% of new applications will be generated automatically by AI without a human being involved in the process.

On the customer and consumer side, this enormously dynamic development is being driven on the one hand by the fact that communication has now reached the mainstream. When even the Bild newspaper, with Hey_! has its own AI chatbot based on ChatGPT on its website, you know that the topic has really reached the masses.

On the corporate side, in addition to countless studies, expert opinions and recommendations and the fact that Gartner has included AI in the well-known Hype Cycle the FOMO (Fear of Missing Out) effect also seems to be having an impact here, according to which companies should have addressed the topic early on so as not to miss out. To be fair, it must be emphasized here that generative AI tools now also offer real added value, particularly in the area of digital experience, and are far more than "just" hip.

However, the rapid growth is also due to other factors. Firstly, the algorithms for generative AI are getting better and better. This enables AI systems to create ever more complex and realistic content. Secondly, the costs of generative AI applications are falling. This makes them increasingly affordable for companies and private individuals.

Conclusion

The last decade has seen many technological advances, but the development of generative AI in the past 12 months has been revolutionary. This technology has the potential to transform industries, massively optimize and also streamline workflows and provide innovative solutions to previously unsolved problems.

The applications seem almost limitless and people's imaginations are being reignited by the capabilities of these advanced models. However, despite this impressive potential, it is important to emphasize that companies wishing to venture into this new territory would be well advised to do some homework first.

Before plunging headlong into the world of generative AI, organizational adjustments and targeted personnel development are often necessary. A structured approach and comprehensive preparation will ultimately make the difference as to whether you can really benefit from the advantages of this technology or whether it remains just another unused tool in the growing digital toolbox.

In general, however, companies should always keep the customer in focus, especially in today's world. Artificial intelligence can deliver real added value here, but technology cannot perform magic. The results are only as good as the data used to generate them and the issue of transparency and ethics should always be kept in mind. Not everything that is now technically possible will really help, and the younger generation in particular sometimes acts much more critically than was the case in the past.

In the SPIN Digital Experience Lab, you can not only experience the Adobe tools live in a coherent demo environment, but also try them out for yourself. Further information and free registration are available at https://www.spin-dx.de/.