A lot of SaaS companies have been experimenting with generative AI and LLMs over the last 12-18 months. And increasingly they’re asking, how are companies doing this in production at scale? Are they seeing the kind of ROI that we all feel could be there? As the leader of the Applied AI team at Google Cloud, Duncan Lennox sets out to answer these questions and shares six ways businesses are currently transforming their businesses at scale and with real ROI.
Duncan’s team at Google Cloud recently ran a survey with around ~2.5k businesses around the world, asking how they are using Generative AI today. The data is clear.
- 61% of the organizations are already using GenAI in production.
- Of those doing that, 86% already see an annual revenue boost of 6% or more.
- 74% are seeing ROI within the first year, which is astoundingly fast for adopting a new technology.
Let’s look at each of the areas where customers are using GenAI and LLMs in production at scale and where they’re getting ROI so you can take those ideas and make them work for you.
GenAI Opportunity #1: Customer Experience
So this is probably one of the first areas that you think of when you think about applying generative AI to businesses. This is everything from chatbots that are able to engage directly with customers and answer their questions, to fully automated support tickets. What’s interesting is, that Google Cloud is seeing across its customer base not only increased customer satisfaction but also increased customer loyalty, sales, and revenue.
Why? Because one of the things that’s happening with the use of AI here is the ability to not only solve problems for your customers faster but also actually offer highly relevant upsell opportunities that get much higher acceptance rates than traditional upsell opportunities. So this is doing everything from personalizing the product and supporting you to creating recommendations. It’s about being able to meet your customers where and when and how they want.
AI improves the customer experience through:
- Personalized product recommendations and omni-channel marketing, meeting the customer where, when, and how they want.
- Delivering nuanced, precise responses to customer queries.
- Responding to many different languages in real-time.
- Sentiment analysis and addressing the biggest issues customers are having right now.
GenAI Opportunity #2: Employee Productivity
How can we take some monotony out of repetitive tasks that folks do every day that were traditionally difficult to automate and increase productivity overall? AI is currently doing that for many businesses, becoming a single source of knowledge sharing and collaboration in the form of an assistant who knows all about your company, policies, and specific work documents.
It can take some of those repetitive tasks away, such as summarizing documents and meeting notes, but it also has advanced capabilities now, like advanced reasoning and analysis, to analyze financial data and provide insights from that. The most interesting area here is where an LLM can tailor training to a person, their needs, and areas of weakness.
Generative AI is helping employee productivity in areas such as:
- Automating repetitive tasks (data entry, scheduling, reporting)
- Summarizing documents and meeting notes
- Offering assistance with regulatory compliance
- Providing on-demand access to company policies and procedures
- Analyzing financial data and provide insights
- Creating personalized training materials
GenAI Opportunity #3: Marketing and Creative Content
Marketing and content creation is an area where we really see AI augmenting human creativity, even though people fear it replacing marketing outright. It’s not designed to replace humans but to be your digital muse, your 24/7 brainstorming partner to work through and come up with creative ideas you might not have thought of on your own. Of course, that can include generating marketing copy, creating digital assets like video and images, and it lets you get into creating personalized experiences for your audience because it learns your brand, its voice and customers.
Generative AI is helping marketing and content creation in the following areas:
- Generating marketing content (images, video, copy, audio, music)
- Automating tasks like image editing and video production
- Designing product packaging and marketing materials
- Personalizing content for specific audiences
- Creating personalized customer experiences
- Generating unique product concepts and ideas
GenAI Opportunity #4: Accelerate Data Analysis and Decision-Making
A meta-theme across these examples is about advanced reasoning and insight capabilities of generative AI and LLMs. Data analysis is a perfect area to get insights that make a difference. The volume of data everyone deals with daily is skyrocketing, and your data is trying to tell you a story. The question is, are you listening?
It’s very difficult to get those insights due to the complexity and volume of data, so it’s an area seeing substantial impacts from using GenAI. It starts with ingesting the data, cleaning and organizing it so it can be used, and moving on to initial, more straightforward but meaningful tasks like pricing optimization and inventory strategies. These tools can also get on more interesting tasks like advanced pattern recognition with things like anomaly stack, predictive modeling, equipment failure, and so on.
Generative AI is helping data analysis and decision-making in the following areas:
- Clean, organize, and analyze complex data sets
- Optimize pricing and inventory management
- Develop personalized recommendations for customers
- Generate personalized data visualizations
- Identify patterns, predict customer behavior, segment customers
- Identify data anomalies and potential risks
- Predictive modeling: forecast sales, customer churn, equipment failures, market trends
Data Analysis Example: neo4j
Neo4j is a leading graph database company. They’ve integrated GenAI into their product suite so customers can query the knowledge graph using natural language. This allows you to take that knowledge graph and use it as another part of your GenAI solutions, and it becomes a source of truth to avoid hallucinations.
GenAI Opportunity #5: Assist Product Development
Software engineering and how to use generative AI around developer productivity is another area that had a lot of early-on excitement. Coding itself is a complex task and involves understanding the domain space you’re working in, the technical complexities of your architecture, the language you’re using, and so forth.
Coding assistance can take out some of the manual steps for software engineers so they can focus on higher-value problems. It can help generate code from natural language and assist in code completion, and this can be tailored to your own private code repositories. So, it’s not just the broad coding knowledge the LLM has but also enhancing and fine-tuning it with your own code repositories.
Generative AI is helping accelerate product development in the following areas:
- Generate code, or complete code as you write
- Identify and fixing bugs in code
- Create automated tests for code
- Explain complex code sections
- Convert natural language to code
- Convert code between different languages
- Generate technical documentation
Identifying and fixing bugs is another great area where we can identify problems more easily. The earlier you can catch and fix a bug, the substantially cheaper it is to deal with.In addition to bug fixes, compilers have advanced in many ways and provide syntactic and semantic error checking. But compilers don’t always understand what your code is trying to do or what business problem it’s trying to solve. By leveraging LLMs, you can get insights into what the code is trying to do and recommend how to fix it.
Generative AI also helps in the area of explainability. One of the biggest productivity syncs is when a new engineer joins a team and a product. They have to get up to speed on the code that exists for that product. Even if you’re the most well-documented code base in the world, it’s still a complex journey to wrap your head around that. LLMs help explain what a chunk of code is doing and get engineers up to speed faster.
Developer Productivity Example: Turing
This is an AI-enabled tech services company leveraging this technology for developer productivity. They have been able to take their own private repositories and code bases to augment the LLM. With this code assistance, they’re seeing 30% productivity improvements from their engineers.
GenAI Opportunity #6: Increase Security, Risk Management, and Compliance
Security is very much needle-in-a-haystack kind of work. How do we spot an anomaly or something that’s going on that could result in fraud, cyberattacks, or noncompliance? Needle-in-a-haystack work is perfect for generative AI and large language models.
Companies are doing investigation assistance and event summarization. When there’s a security event, a lot is happening in real-time, so the ability to have accurate, up-to-date summarization of where the event is right now and the next steps is incredibly valuable. Security engineers can stay focused on managing and mitigating the incident.
Generative AI is helping to increase security, decrease risk management and speed up compliance in the following ways:
- Identify and address regulatory compliance issues
- Investigation assistance with security event summarization and next-step guidance
- Safeguard client data and identify suspicious activity
- Proactively detect and prevent fraudulent transactions
- Protect infrastructure from cyber attacks with continuous network monitoring
- Create personalized training materials
Anomaly detection is another great problem for GenAI and LLMs to solve. It does everything from safeguarding client data to looking at fraudulent transactions across a number of parts. LLMs allow teams to respond to events much faster.
Risk Management Example: PaloAlto Networks
They have integrated advanced AI search capabilities and their own corpus of data into the responses that they’re able to provide to customers, making sure they’re always providing the most up-to-date and accurate information.
Key Takeaways
Customers are using generative AI and LLMs in production at scale today. The key takeaways from these examples are:
- AI is transforming business areas from customer experience to compliance.
- Companies are achieving real-world ROI from their investments.
- Explore AI solutions for your priority use cases to deliver the fastest ROI.
