By: Sheba Fernando, AVP & Head of Enterprise AI Solutions, LTIMindtree

Generative AI or Gen AI has been the buzzword around most boardrooms over the last 2-3 years and the hype around it refuses to die. But is it just that, a hype or a game-changing opportunity that business leaders must capitalize on? The answer is a bit of both. Let’s find out why and look at some of our top use cases that can solve real-world industry challenges.

Traversing the generative AI roadmap

Reuters[i] reported that the public version of ChatGPT reached 100 million subscribers in just two months, becoming the fastest-growing consumer application in history. By comparison, Instagram took 2.5 years to reach this milestone. This democratized the technology and allowed users to carry out a multitude of tasks. However, this versatility and the lack of reliable training data have reduced the accuracy of the results and led to global calls for creating AI guidelines on ethics, regulatory compliance, transparency and risk management. Leaders also realized that enterprise needs were far more sophisticated than what the plug-and-play Gen AI models could deliver.

This has led to the adoption of customized solutions with proper guardrails for security and compliance. The results are already bearing fruit with the customized Gen AI solutions creating new and exciting use cases for businesses and improving, scaling, or speeding up existing ones. Our recent survey[ii] showed that early or extensive adopters use Gen AI to improve customer experience, process efficiency, productivity, product design/innovationand analytics. This is also consistent with what we are seeing with our clients.

Top 5 Gen AI use cases

At the start of 2023, Generative AI was seen as a chatbot that could only generate content in the form of text, imagesand code based on simple prompts. Since then, it has evolved to drive value and eliminate real-world challenges, including the following five use cases that we have implemented and industrialized for our clients.

Use case 1:

Facilitating real-time language translation for an American life insurance major

Value captured:

Customer experience, productivity and process efficiency

Challenges:

Global businesses work with customers across geographies, which means overcoming language barriers, among other things. Previously, this required using a human interpreter during all meetings, which increased translation time and complexity, compromised end-user experiences and resulted in information breaches.

Overcoming with Gen AI:

Using generative AI models with Azure AI Speech services and Azure OpenAI, the solution helps translate spoken words in real time from the source to the destination language.

Benefits:

• Risk management: Reduced the risk of information breach and loss of meeting data by storing it for future reference.
• Process efficiency: Saved manual cost, time and resources by up to 95% and increased efficiency by 10X.
• Streamlined conversion processes: The solution allowed insurance agents to collaborate and create tailored customer contracts by overcoming the language barrier.

Tools used:

Azure OpenAI–(Model: gpt-35-turbo) and Azure Cognitive Services

Use case 2:

Simplifying product research and design for an FMCG major

Value captured:

Product design/innovation, process efficiency and data analytics

Challenges:

Enterprises spend considerably on their research and development in marketing design initiatives and want to secure their proprietary research and design at all costs. This also means that this data is stored in silos in the form of PDFs and images. This makes it difficult for future researchers to find related reference data for further innovation.

Overcoming with Gen AI:

Customized Gen AI platforms generate innovative marketing designs, both text and images, securely from your proprietary data with access control and pull out old designs and research based on the keywords/context. Using generative AI and Stable Diffusion models, the solution helps translate NLP words to different banner and brochure design ideas, for the marketing team to enable product marketing design for campaigns.

Benefits:

• Process efficiency: Improved the efficiency of research and design teams by up to 30% with Gen AI.
• Go-to-market: Improved the ability to innovate and current deliverables faster, accelerated product launches and increased return on investment.
• Risk management: Minimized data breaches using access-controlled solutions with user activity data.

Tools used:

Azure Open AI GPT 4.0, Stable Diffusion models

Use case 3:

Optimizing website content and design for a global CPG Organization

Value captured:

User experience, process efficiency and data analytics

Challenges:

Capturing customer attention is hard, especially in the B2C sector. It requires generating content that is relevant, informative, user-friendly and optimized for search engines and devices. However, large global Consumer Packaged Goods (CPG) companies have been creating content for decades and much of it is dated and irrelevant. Updating it requires massive effort, technical capabilitie sand compliance with brand guidelines and website best practices.

Overcoming with Gen AI:

Creating/rewriting content using generative AI is a multi-step process. It starts with analyzing the client’s existing content to understand the tone of voice and writing style. Next, the tool analyzes competitor content from WebAnalytics 3^rd party data to identify the trending keywords and topics, which are used to create fresh Search Engine Optimization (SEO) and brand guidelines-compliant content. Gen AI is then used to generate meta descriptions, blog content and URL suggestions.

Apart from the abovementioned process, we plan to use gen AI to analyze existing content for gaps in SEO and relevancy and update it using keywords from Google AdWords and SEMrush. We’ll also use it to recommend topics based on existing content and SEMrush keyword trends.

Benefits:

• Improved website conversions:Enhanced branding and improved traffic to website from both paid and unpaid sites
• Content best practices:Optimized content for SEO, search engines and client brand guidelines and eliminated plagiarism.
• User experience:Increased user awareness and helped them move to the consideration stage of the buyer’s journey. Created backlinks to internal product pages for a seamless user journey.

Tools used:

• Azure Open AI GPT 3.5
• Azure Open AI GPT 4
• Prompt Engineering
• LangChain Framework

Use case 4:

Automating financial report generation for financial services organization

Value captured:

Data analytics and process efficiency

Challenges:

Financial research reports provide a detailed analysis of investment opportunities and analysts and experts spend between 40 and 80 hours per report. The reports are consumed by enterprises looking to maximize their investments and are prepared manually by researchers. The manual processes take over two weeks to create such reports, which might be too long for enterprises.

Overcoming with Gen AI:

Using a financial research co-pilot that uses generative AI’s NLP capabilities boosts the productivity of the researchers by automating the drafting of financial reports. Pre-training the model on previous financial reports helps it replicate human writing styles instead of machine-generated content. It also summarizes large PDFs and generates relevant research report sections.

Benefits:

• Increased productivity: The co-pilot saved the time taken to generate reports by 70% and increased productivity by 5x.
• Reduced errors: The generative AI platform eliminated repetition and reduced manual fatigue while creating lengthy reports.
• Faster go-to-market: The timeline for publishing reports, including generation and review, reduced from 5-10 days to within a day.

Tools used:

Open AI GPT 3.0 – Davinci Model

Use case 5:

Integrating ChatGPT with enterprise policy data for a leading non-profit

Value captured:

Process efficiency and customer experience

Challenges:

Multinational non-profit organizations use different policy documents to comply with regional regulations. Storing these documents in a common knowledge repository ensures easy access for employees but increases confusion. Their miscellaneous queries about company policies, such as reimbursement or travel allowance across countries or grades, are not easily resolved and require an analyst to be always available. This complicates resource planning and increases operational expenses.

Overcoming with Gen AI:

Using a chatbot interface with NLP capabilities to parse through company policy documents facilitates the necessary information through an integrated search operation. This includes enabling a ChatGPT-like experience, with Azure Open AI and Azure Cognitive Service for capabilities like querying, searching, summarizing and comparing different documents. Pre-training the model on previous searches enables replicating responses through machine-generated content. This greatly reduces the dependency on human intervention as accurate policy documents are highlighted based on the requested search with minimal human intervention.

Benefits:

• Increased productivity: Minimized the time taken to respond to queries by 50% and increased employee productivity by 2x.
• Increased automation: Reduced manual intervention needed to search for specific content in the document freed the team to take up critical work.
• Reduced errors: Eliminated repetition and reduced manual fatigue while responding to queries.
• Enhanced customer experience: Responded to employee questions in real-time based on semantic search and summarization of various policy documents.

Tools used:

• Azure Cognitive Search
• Azure Open AI (GPT3.5)
• Prompt Engineering

In Conclusion

Our study[iii] shows that 33% of leaders see Gen AI adoption increasing their revenue by 20% or more. McKinsey research[iv] also suggests that it can add an equivalent of up to $4.4 trillion of value annually to the global economy. So, does this mean that gen AI has moved beyond hype and leaders must forge full speed ahead to take advantage? Yes, but with a caveat.

While the possibilities around Gen AI are limitless, the rhetoric surrounding the technology is inflating user expectations and making them underestimate the challenges, including counterfeit, fraud and reputational risks.

However, implementing the right use case with proper guardrails and security and risk management can help businesses overcome the challenges and unleash the full potential of generative AI.

To know more, please download our report here.

[i] Reuters, ChatGPT sets record for fastest-growing user base – analyst note, January 2, 2023: ChatGPT sets record for fastest-growing user base – analyst note | Reuters

[ii] LTIMindtree, State of Generative AI Adoption: https://www.ltimindtree.info/gen-ai/

[iii] LTIMindtree, State of Generative AI Adoption: https://www.ltimindtree.info/gen-ai/

[iv] Mckinsey, The economic potential of generative AI: The next productivity frontier, June 14, 2023: https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/the-economic-potential-of-generative-ai-the-next-productivity-frontier#business-value