A CEO’s Guide to Generative AI

By Mark Larsen

Introduction 

In November 2022, OpenAI unveiled ChatGPT-3.5, an innovative milestone so transformative that it recalibrated our anticipations about AI’s imminent role in daily life. Amassing a staggering 100 million users in just two months, it became the fastest growing app of all time.  

As the tide of generative technologies surges, CEOs are tasked with harnessing its potential, optimising advantages while curtailing risks. 

This guide offers clarity in navigating this complex landscape.   

We’ll Cover 

• The Difference Between Generative AI and Traditional AI 
• Categories and Popular Tools 
• Use Cases and Concerns by Department 
• Business Opportunities 
• How to Embrace Generative AI in Your Organisation.  

The Difference Between Generative AI and Traditional AI 

Generative AI boasts versatility and creativity, allowing it to tackle diverse tasks and content generation effortlessly. In contrast, traditional AI operates within a more confined scope and requires greater configuration and expertise to solve specific problems.  

Differences have amplified as the technologies unfold.  Generative AI’s versatility allows for conversational experience. This brings accessibility.  With ChatGPT, users can quickly produce meaningful outputs, paving the way for broader applications. It’s this expansive potential that has transformed global perspectives. 

Categories and Popular Tools 

Most generative AI tools specialise in one of the following outputs:    

• Text  
• Image 
• Video 
• Audio 
• Code 
• Design 

Examples and Popularity

(Percentages indicate web traffic as of Sept 2023) 

Text (89%) 

• ChatGPT 4.0 (58%) – The head-turning natural language marvel. Trained on data up to Jan 2022.  
• Bard (now Gemini) (7%) – Google’s contender, slightly less captivating, but gaining market share since the release of Gemini. Its standout feature is access to real-time data. 

Image (8%) 

• Photoroom (2%) – Dubbed the ‘magic’ photo studio. It can switch backgrounds, erase undesired objects, and enhance image quality. 
• Midjourney (2%) – Converts natural language descriptions into images. Allows users to iterate outputs to improve results.  

Video (<1%) 

• Veed – Offers automation for video tasks like captioning, trimming, and effects addition. 

• Synthesia – Transforms text into realistic video content, with digital hosts fluent in various languages. 

Audio (<1%) 

• Speechify – Transforms text into audio. 
• VocalRemover – Extracts and isolates vocals from tracks. Useful for remixes or karaoke. 

Code (<1%) 

• Tabnine – Assists with code completion and pattern recognition. 
• OpenAI Codex – Translates natural language into code. 

Design (<1%) 

• Tome – Generates multimedia presentations that are more dynamic than slideshows and simpler than webpages. 

Furthermore, there’s an expanding suite of platforms where users can share and refine AI models: 

Platforms (<2%) 

• Hugging Face (1.4%) – Renowned for its natural language processing models and libraries. Equipped with tools for text-driven tasks such as translation, sentiment analysis, and text generation. 

Tool Selection 

ChatGPT 4.0 is the clear leader in its category but other LLMs (Large Language Models) are gaining traction. For those with above average risk tolerance, integrating an LLM and facilitating its training becomes essential. Standing up a private instance of an LLM can mitigate risks, ensuring both prompts and outputs remain confidential. 

Although text generating tools, primarily ChatGPT, are driving most of the buzz around generative AI, tools tailored to specific domains have significant promise.  

As with conventional tool selection, you’ll want to narrow focus. A lean tech stack is key, so a tool’s value must outweigh the cost of complicating the tech stack. Often, it’s pragmatic to opt for straightforwardness, even if it means forgoing tools with discernible benefits. 

Use Cases 

By department, in Order of Bullishness 

IT & Software Development 

67% of IT leaders surveyed said they have prioritised generative AI for their business within the next 18 months 

Inexpensive SAS tools can dramatically reduce development time. Uses include: 

• Code creation 
• Code completion 
• Improved readability / commenting code  
• Debugging 
• Testing 

Generative AI isn’t a substitute for skilled engineers. Experienced engineers are likely to benefit most, as they can better navigate the pitfalls, while maximising benefits.  

Marketing 

51% of marketers are currently use generative AI to boost marketing efforts 

Seven in 10 marketers (71%) expect generative AI to help eliminate busy work and allow them to focus more on strategic work.  

Marketers expect to utilise most categories of Gen-AI. Activities include:   

• Content creation 
• Writing/improving copy 
• Inspiring creativity 
• Analysing market data 
• Generating royalty-free images 

Sales 

35% of sales professionals are using Gen-AI to accelerate their sales processes  

Uses include: 

• Content creation  
• Analysing market data 
• Helping to generate sales reports 
• Automating personalised sales communications 

Customer Support  

Only 24% of customer support professionals said they were currently using generative AI for work. However, 9 out of 10 from those who do say it helps them serve customers faster. 

The latter suggests the low adoption from this group doesn’t stem from lack of opportunities.  

Uses include: 

• Implementing chat bots to resolve common issues 
• Creating and personalising customer communications  
• Generating case summaries 
• Generating knowledge articles 

Concerns 

From IT Leaders 

IT leaders balance their overall bullish outlook with practical concerns:  

• 71% say Generative AI will introduce new security threats to their data 
• 66% believe their employees lack the skills to use it successfully  
• 60% note Generative AI can’t yet integrate with their business’s tech stack 
• 59% highlight their organisation lacks a unified data strategy 

From other employees 

The main concerns employees themselves raised were: 

• Not knowing how to get the most value from generative AI at work.  
• No knowing how to use it safely or effectively 
• Lack of training 
• Worries they will lose their job if they don’t learn how to use generative AI at work. 

This matches well with what they need to feel enabled: 

• 70% of non-users would use generative AI more if they knew more about the technology.  
• 64% would use generative AI more if it was more safe/secure.  
• 45% would use generative AI more if it was integrated into the technology they already use. 

Business Opportunities 

Ranked from Most to Least Accessible 

Internal Adoption 

Integrating GenAI into your organisation is the quickest avenue to experience its advantages. Enabling your teams to boost their productivity has immediate benefits and prepares them well for the changing climate. The role of the CEO is instrumental in shaping the organisation’s AI-centric strategies, culture, and frameworks. 

We explore this deeper in the next section: How to Embrace Generative AI.   

Knowledge Management 

This is the main horse Data Cubed is backing. Connecting ChatGPT, or similar, to the mountain of files and reports within your organisation allows users to get quick, succinct answers with ease.  

LLMs can be connected to both unstructured (pdfs, ppts, word docs etc) and structured data (tables, databases etc) allowing insights to be served effortlessly.  

Our Data Interrogator allows you to explore your knowledge base using a private instance of ChatGPT, so prompts, outputs and sources are never shared.  

Templating 

A huge amount of manual resource is spent pulling information from multiple sources and shaping them for a company template. As well as succinctly organising information, ChatGPT can also follow specific formatting guidelines. The shaping part is now easy.  

Tools like Zapier have integrated ChatGPT and can connect to thousands of other applications. This allows you to automate many templating processes natively. Alternatively, 3^rd party services (like ours) can help you automate more custom templating requirements.  

Applications 

Products and services that leverage existing foundational models often present the most straightforward opportunities for revenue and lead generation.  

Consider the following: 

• WonderPlan automates holiday itineraries.  
• Recently trending, BarbieMe turns users into a doll.  

Author as a doll

These applications cleverly harness existing models, tailoring them in ways that resonate with their intended audience. This is low effort, high reward when done well.  

Generative AI predominantly emerges in an open, collaborative environment, with APIs being widely available. This means barriers to entry are surprisingly low.  

Transitioning from an initial idea to a Proof-of-Concept and ultimately to a full-fledged product can be seamless. Tools like Power Apps, which emphasise low-code development, further simplify this path. With a little know-how, you could build a ChatGPT component in Power Apps in less than an hour.   

If you have an idea but you’re unsure how to get started, reach out. Our Data Cubed experts will know the simplest way to bring your ideas to fruition.  

Services 

Supporting organisations to effectively leverage generative AI is clearly a much-needed service.  

Opportunities include:  

• Setting up private infrastructure 
• Templating 
• Bespoke training 
• Courses/educational content 
• Consultancy 
• Workshops 

• SaaS tools to facilitate adoption (benchmarking, roadmaps etc)   

Platforms  

Developing platforms and tools that empower users to curate, refine, showcase, and collaborate on either the foundational models themselves or content derived from existing models is a promising avenue. 

For example, CivitAI offers a space for creators to share and evolve AI generated art.  

 
Niche Opportunities 

Delving into more specific areas of expertise, opportunities arise from creating in-house foundational models to offering hardware and cloud solutions to meet rising computational needs.  

Reflecting on generative AI’s influence on the prevailing norms and its potential effects on relevant sectors can uncover fresh avenues of opportunity. 

How to Embrace Generative AI 

The fastest way to benefit from generative AI is to embrace the technology and drive adoption throughout your organisation. Generative AI is a rapidly evolving field, and as leaders, it’s crucial to foster an environment that encourages experimentation while addressing concerns and minimising risks. 

Identify and Prioritise Use-Cases and Focus on Successes  

• Create opportunities for team members to share and discuss use cases, stories, successes and failures. 
• Welcome and actively seek out ideas for potential use cases and facilitate the development of proof of concepts (POCs). 
• Ensure team members know which tools they can access and processes for subscribing to new tools.  
• Use POCs to evaluate potential quickly.  
• Concentrate on early successes and promote a culture of scaling and evolving what works. 
• Ensure successes are well-coordinated across departments.  

Understand the Risks  

• Leaders should have a deep understanding of risks for each potential use case. 
• Risk assessments: Establish structured processes for ongoing check-ins on AI use. Identify and flag potential risks in a timely manner. 
• Encourage open communication and a culture of support and facilitation.  
• Address challenges proactively to ensure risks and tools can be managed effectively.  
• Emphasise practices that align with the organisation’s risk tolerance and strategic goals. A law firm may opt to restrict use to minimise risks, while a tech startup may accept higher levels of risk to accelerate adoption.  

User Education and Guidelines:  

• Define and communicate a generative AI strategy. Align all stakeholders with the organisation’s objectives and expectations. 
• Offer tool specific training, risk-assessment and guidance.  
• Ensure that guidance is both succinct and pertinent. For example: “Never prompt ChatGPT with content that violates non-disclosure agreements; only use publicly available information in prompts.” 
• When uncertainty arises, prioritise decisiveness.  If team members are unclear on something, make a decision and communicate a policy. 

Conclusion 

As we demystify generative AI, challenges start to look more familiar. Business leaders can only focus on a fraction of the opportunities available. The challenge is ranking opportunities, choosing priorities and exploring the most promising in strategic and creative ways.   

Tough decisions lie ahead, and decisive leadership is essential to articulate the strategic direction, coordinate goals, and ensure everyone understands the rules of engagement. 

Generative AI promises to fuel innovation, improve customer interactions, amplify productivity and boost competitive edges. It’s not just a valuable tool; it’s a strategic imperative for those aiming to thrive and lead in the ever-evolving marketplace. 

Get in touch at hello@data-cubed.co.uk to find out more