How to create value with generative AI beyond chatbots
After a year of excitement and experimentation, I wanted to dig deeper into some real ways your company can use generative AI to create value.
As organizations are building gen AI-powered products and services, chatbots are the obvious thing most gravitate towards. Let’s broaden that perspective and examine how gen AI can be helpful to your organization in multiple ways.
Here are five categories of how gen AI can help your business:
Extreme personalization
Amplified customer service
Operational efficiency
AI-powered creation
Supporting decisions
I’ll explain each category’s meaning, real examples, cases, and key considerations.
1. Extreme personalization
What it means
Generative AI is enabling each customer to have their own personalized experience.
Instead of a one-size-fits-all website, app, or email, each customer can have their own generative AI-enabled experience.
Duolingo brings a personalized language learning experience for students with AI
Example cases
A global retailer is building personal journeys for each customer. Based on your browsing and purchase history on their site, they will generate personalized deals via SMS that are completely custom-made with a Large Language Model.
Language learners can now use the Duolingo app to practice Spanish or French with the help of real-time roleplay. The GPT-4-powered speaking partner will converse with users in everyday situations. AI also provides detailed feedback on users' language mistakes
Questions to consider
Consider your current customer journey; what would it look like if all customers had a unique experience?
What are the unique data assets your company has? How could you combine your product and customer data to create a differentiated experience?
2. Amplified customer service
What it means
Amplified customer service means combining human skills with AI to enable the best possible experience at a reasonable cost.
According to recent research, AI is automating some customer service tasks, not affecting some tasks and augmenting others. It also frees up customer service representatives to focus on completely new, higher-value tasks.
Companies are now experimenting with arming their front-line employees with powerful AI tools to help serve customers better (see examples).
I worked with a customer service AI chatbot start-up last year. One of the key pieces of designing the experience was designing a smooth handover between the AI and humans. How can we enable frictionless and transparent collaboration for the user and the customer service agents?
Stripe uses AI to create a summary answer based on several documents for users of its help center. (source: Stripe & OpenAI)
Example cases
Customer service representatives in a large enterprise software company could better respond to customer questions with the help of an AI system that read their chats and suggested responses based on what had produced good outcomes in the past.
According to the study detailed in HBR, during the 7-week pilot, agents could handle 15% more chats, resolve them 10% faster while increasing customer satisfaction.Customers can troubleshoot their issues faster with gen AI-powered help centers. The payments company Stripe is using GPT-4 to quickly generate a summary answer to a question from multiple technical documents. Instead of clicking around to piece together an answer, customers receive a full answer immediately.
Questions to consider
What is the right balance for your company between AI-powered human service agents and automated AI-bots? A white-clove human service amplified by AI makes more sense than complete automation for companies with high average customer lifetime value.
3. Operational efficiency
What it means
Companies are taking repetitive and error-prone human tasks and finding ways to increase their speed and efficiency with AI.
Employee onboarding, quality control, answers to questions found in knowledge bases, and manual data entry are prime subjects for introducing generative AI.
Ironclad helps legal teams draft legal documents faster with AI review and suggestions (source: Contractbook)
Example case
Ironclad is a contract lifecycle management software. It now helps legal teams create high-quality contracts faster and with AI. The AI assist feature suggests corrections or additions to contracts that the human professionals might have missed. Humans always have the last word and can accept or override AI suggestions.
Questions to consider
What are the most time-consuming manual processes in your company?
How could generative AI help in information retrieval, quality control, summarization or content creation to improve its efficiency?
4. AI-powered creation
What it means
Generative AI can help your company draft content from copywriting to reports and designs.
AI-powered content creation does not mean outsourcing to AI but using generative AI as an amplifying assistant in all phases of producing content.
Jasper helps marketing teams generate campaign content with AI based on input text, asset types, and tone.
Example cases
A Finnish staffing and recruiting agency use GPT-4 to help them generate job postings based on rough role descriptions. Human operators guide the AI and iterate on its work, but the AI does most of the heavy lifting of content creation.
Several consultants I’ve spoken with are experimenting with using Large Language Models to create reports and proposals for their clients. Models like GPT-4 have been trained on their tone of voice and are fed rough notes with no identifiable or private details to generate fluid texts as first drafts.
I use GPT-4 to brainstorm ideas for this newsletter. I explore different angles I could take on a topic but never rely on AI to write my content as I want it to retain my voice and improve my own thinking.
Questions to consider
What are the most resource-intensive content creation processes in your company? How could AI help you create content more efficiently
How could you use AI in different phases of content creation? How could you use it for research, ideation, creation, and content refinement?
What are the parts of content creation that you shouldn’t rely on an AI for to retain your voice and level of quality and improve your thinking?
5. Supporting decisions
What it means
Combined with luck, the quality of decisions ultimately decides a company's success.
How can we tap into generative AI to improve our decision-making? One superpower of gen AI is taking massive amounts of unstructured data and finding patterns we miss.
In addition to finding insights, we can use generative AI to sharpen our own thinking with collaborative strategy and sparring sessions. I’ve used it several times as a mentor to ask the right questions to form the strategy of my business.
Example cases
A startup called Viable helps companies find insights from unstructured feedback data. Combining insights from customer service, social media, and reviews, decision-makers can ask questions like “What are the biggest frustrations our customers have?”
Leaders can use a tool like GPT-4 as a devil's advocate on their strategic thinking with a prompt like “You’re a world-class business consultant. Help us examine the potential risks of this strategy. We’re planning to (include 5 bullet points of your strategy)”. Use secure solutions like ChatGPT Enterprise or Azure to protect confidentiality if needed.
Questions to consider
What are the most critical decisions our business makes? What data sources could help us make better decisions? How could we leverage generative AI to analyze that data and uncover new insights?
How could we use a tool like GPT-4 for our strategic planning? How could it give us ideas and help us refine our thinking?
Going beyond chatbots in generative AI
As we look at the landscape of how generative AI is used, it’s clear that chatbots only scratch the surface.
In virtually all areas of business where people use information to make decisions or create content, generative AI can add efficiency and improve quality.
This year, it’s crucial to broadly explore how generative AI can help our organizations in real and tangible ways. Doing so is vital within robust ethical, safety, and legal guardrails.
If your company needs help with any of this, I just launched a new workshop where we ideate to uncover the most promising use cases for generative AI
Want to read more?
This article was first published on my weekly newsletter Generative AI for Design and Innovation.
