Generative AI in Business: Innovation, Creativity, Automation

AlphaFold, which uses gen AI to predict protein folding with unprecedented accuracy, a feat that recently earned the 2023 Nobel Prize in Chemistry. Tools like ChatGPT and DALL·E are enabling businesses to generate images, content, ideas, and products in ways previously unimaginable. But with great power comes risks like deepfakes and bias. How can we balance the potential of generative AI with ethical responsibility in its use?

Generative artificial intelligence (gen AI) has emerged as a transformative force, revolutionising how businesses create, innovate, and operate. These sophisticated algorithms, exemplified by tools like ChatGPT, possess the remarkable ability to generate diverse forms of content, from text and code to images, audio, and complex simulations. 

Beyond text generation, generative AI is also being used to create realistic images and artworks through platforms like DALL·E, compose music and soundtracks using tools like Jukedeck, generate videos and animations with tools like Runway, and even create 3D models and simulations for industries like gaming, architecture, and healthcare. 

While the technology presents certain challenges and risks that demand careful consideration, businesses are increasingly recognising that the potential benefits far outweigh the obstacles. 

This has led to widespread adoption and experimentation across sectors, as organisations uncover unprecedented opportunities to enhance productivity, streamline workflows, and develop innovative products and services.

As generative AI capabilities grow, discussions about its ethical use are becoming more prominent. Companies are focusing on creating responsible AI frameworks to address issues like bias. However, a recent Deloitte study reveals that 56% business executives are unaware or unsure if their organisations have ethical guidelines for using generative AI.

Understanding generative AI and how it works

Generative AI comprises deep learning algorithms capable of generating content across various formats. These models do not possess human-like understanding but are trained on vast datasets to identify patterns, predict likely outcomes, and generate responses based on statistical probability. The most widely used techniques include:

• Generative Adversarial Networks (GANs): Used for image and video generation.
• Transformers (e.g., GPT models): Specialise in text generation and natural language processing.
• Stable Diffusion: Focuses on high-quality image synthesis.

Despite their sophistication, generative AI systems lack human traits such as common sense, emotional intelligence, and intuition. They are unable to grasp abstract concepts or ethical dilemmas in the way humans do. Rather than replacing human intelligence, these systems are best seen as tools for augmenting human capabilities. When used responsibly, they enhance efficiency and creativity without compromising critical judgement.

The operation of generative AI can be summarised through the following stages:

1. Training

Generative AI relies on foundation models, deep learning systems trained on broad datasets to identify patterns. This phase involves high computational demands and financial cost. For example, a large language model may be trained on books, websites, and forums to understand grammar, semantics, and structure.

2. Tuning

After training, models are tuned for specific tasks:

• Fine-Tuning: Involves further training using labelled datasets. For example, a customer service bot may be fine-tuned on past support tickets and responses.
• Reinforcement Learning with Human Feedback (RLHF): Here, human evaluators score generated outputs to improve model relevance and accuracy.

3. Generation and Evaluation

Once deployed, models continuously generate outputs. Developers monitor these for performance and update models periodically (typically every 12–18 months). A method like Retrieval Augmented Generation (RAG) is often used, where the model accesses external databases to generate more accurate or current outputs.

Industry Applications of Generative AI

1. Retail and Consumer Goods

Generative AI supports personalised marketing, dynamic pricing, and improved customer engagement. It enables businesses to automate content creation for emails and advertisements and facilitates virtual try-ons for online shopping. Examples:

• Etsy uses Vertex AI for optimised product recommendations and ad placement.
• Dunelm employs Google Cloud’s AI to improve product discovery and reduce search friction.

2. Automotive and Logistics

In automotive design, generative AI assists in optimising components and developing autonomous systems. Logistics firms use it for route planning, inventory management, and customer service automation.

Example: Volkswagen of America has launched an AI assistant within its app to improve customer interaction.

3. Healthcare and Life Sciences

AI assists in diagnostics, drug discovery, and personalised treatment. It supports medical imaging, synthesises patient records, and aids virtual health assistants.

• Insilico Medicine used AI to discover a potential fibrosis treatment in 46 days.
• AI-generated voice clones help individuals with motor neuron disease communicate more effectively.

4. Financial Services

Gen AI enhances fraud detection, regulatory compliance, and sustainable finance. It analyses unstructured data for ESG assessments and flags financial irregularities.

• ING Bank uses a chatbot to enhance internal service quality.
• Mastercard applies AI to identify compromised cards and prevent fraud.

5. Public Sector and Nonprofits

Government bodies use gen AI for policy reporting, data analysis, and public engagement. Nonprofits benefit from AI-driven grant writing, fundraising, and impact reporting.

• Opportunity International supports Malawian farmers with a Chichewa-speaking AI chatbot.
• Bower gamifies recycling through an AI-enhanced app in the UK and Nordics.

6. Manufacturing and Electronics

Manufacturers deploy AI to design products, manage supply chains, and improve quality control. Consumer technology companies embed AI into devices for enhanced functionality.

Example: Oppo and OnePlus integrate Google's Gemini models for features like AI summaries and real-time transcription.

7. Media, Marketing, and Gaming

AI is widely used in video editing, automated journalism, and dynamic content creation. It supports tailored ad generation, real-time campaign adjustments, and game development.

Example: Copy.ai assists marketers in creating consistent content aligned with brand voice.

8. Hospitality and Travel

AI enhances booking systems, customer service, and itinerary planning. It offers real-time updates, personalised recommendations, and dynamic pricing.

Example: HomeToGo’s "AI Sunny" supports users throughout the travel booking process and is set to become a comprehensive AI travel assistant.

9. Technology Sector

AI streamlines digital interactions through predictive analytics, chatbots, and voice assistants. It automates backend operations and personalises user experiences.

Example: Gojek’s voice-enabled assistant, Dira, facilitates easy access to financial services through conversational interaction.

Transforming workflows and workforce efficiency

Generative AI is revolutionising business operations by enhancing efficiency across key functions. According to a Gartner survey, executives are prioritising AI adoption for several strategic objectives, with 38% focusing on customer experience and retention, 26% on revenue growth, 17% on cost optimisation, and 7% on business continuity. These priorities reflect the technology’s potential to drive both competitive advantage and operational resilience.

Revenue Opportunities: One of the most significant impacts of generative AI lies in accelerating product development cycles, enabling businesses to bring innovations to market faster. Additionally, it facilitates the creation of new digital services and products, opening previously untapped revenue streams. Companies leveraging AI-driven design and prototyping, for instance, can reduce time-to-market while maintaining high-quality outputs.

Productivity Gains: Generative AI enhances workforce efficiency by automating repetitive tasks, such as content creation, document review, and even coding. Rather than replacing human workers, AI acts as a collaborative tool—handling routine processes while employees focus on strategic decision-making and creativity. For example, AI-generated drafts can be refined by human editors, streamlining workflows and reducing manual effort.

Risk Management: Beyond efficiency, generative AI strengthens risk mitigation by improving compliance monitoring and predictive analysis. AI systems can scan regulatory updates, flag potential violations, and forecast operational risks based on historical data. This proactive approach allows businesses to address vulnerabilities before they escalate, ensuring smoother operations.

Risks and challenges of generative AI

Despite its advantages, gen AI presents notable risks:

• Bias and Discrimination: AI can replicate and amplify societal biases. For example, Fable's AI-generated book summaries were criticised for making discriminatory remarks.
• Misinformation and Deepfakes: AI-generated content can spread disinformation. A notable case involved a deepfake video of Ukraine’s President announcing a false surrender.
• Hallucinations: AI outputs can be inaccurate or misleading. Examples include Google's Bard providing false astronomical facts and Meta’s Galactica model producing biased text.
• Security and IP Issues: AI systems may be vulnerable to manipulation and raise legal questions about content ownership. Artists have sued AI firms for allegedly copying their work.
• Lack of Accountability: Determining who is responsible for AI-generated harm remains complex.
• Mental Health Concerns: Prolonged interaction with AI chatbots may increase social isolation and emotional dependency, as highlighted in studies by OpenAI and MIT.

The future of generative AI in business

Generative AI is evolving from a support tool into a central element of business operations. Over the next decade, generative AI is expected to:

• Become integrated into unified interfaces for business operations
• Enable real-time, collaborative work between humans and multiple AI agents
• Redefine roles such as marketing, design, coding, and content creation

By 2026, more than 100 million people are expected to collaborate with AI agents at work. The shift will require new competencies, prompt engineering, data curation, and ethical oversight.

According to McKinsey, industries like technology, banking, education, and professional services will experience profound changes. Marketing and sales functions are set to see the most widespread gains.