From CES Showroom to Banking Boardrooms

Dear Readers,

Welcome to this week's edition of FutureOrg AI, your go-to source for the latest in AI developments affecting the business world. Grab your coffee or drink (for happy hour) and let's dive into some exciting news!

It’s 2025! I hope everyone had a wonderful year-end celebration. I did have a good time with my colleagues, family and friends. Looking ahead, I see more exciting news on AI and all the cutting-edge technologies. Just want to say how grateful I am for all of you. Without you, this newsletter without have been possible. I look forward to sharing with you everything I’ve learned about AI business insights with everyone.

Cho-Nan

INSIDE THIS WEEK:

🌏 COVER STORY: CES - AI everywhere 

Top use cases of AI in banking world

Popular AI agent frameworks for enterprise environment

CES - AI everywhere

Having attended CES 2024 last year, I'll unfortunately miss this year's event due to scheduling conflicts. While the convention will undoubtedly showcase numerous AI applications, tools, and generate its share of industry buzz, I feel compelled to offer a word of caution. As both a practitioner and technologist in the AI field, I emphasize the critical importance of prioritizing pragmatism and ROI when developing AI software. We've witnessed considerable hype in this space, along with several notable failures. It's time for our industry to embrace accountability and maintain a clear-eyed perspective on AI's real-world applications.

For those who have never attended CES, it remains an exceptional experience worth considering - starting Monday Jan 5th. The event draws massive crowds and generates palpable excitement, with its standout feature being the impressive lineup of keynote speakers. If I were able to attend this year, I would particularly look forward to presentations from NVIDIA CEO Jensen Huang, Accenture CEO Julie Sweet, and Waymo Co-CEO Tekedra N. Mawakana. Their keynotes alone would justify the trip.

Watch keynotes here.

Top use cases of AI in banking world

After spending the past year conducting in-depth research for financial sector clients, I can identify clear patterns in how banks are adopting AI—findings that align with numerous professional publications and industry surveys. Four main areas of AI implementation have emerged: customer service chatbots, fraud detection systems, personalized financial services, and risk management solutions. While traditional banks have taken a measured approach to adoption, digital-first fintech players have already deployed many of these applications in production environments. It's worth noting that while banks often don't explicitly differentiate between generative AI and traditional AI in their reporting, our research indicates they tend to favor traditional AI applications over generative ones.

Use case types and examples

1. customer service chatbots

   1. Bank of America's virtual assistant Erica has become a cornerstone of their digital strategy, handling an impressive 2 billion interactions since its 2018 launch. With 42 million clients and 2 million daily engagements, it demonstrates the scale at which AI can operate. Similarly, RBC's NOMI platform has driven a 20% increase in mobile app usage, with customers consuming over 100 million insights in just five months.

2. fraud detection

   1. In fraud detection, traditional methods are being augmented by AI's pattern-recognition capabilities. TD Bank's implementation of AI-powered fraud detection has yielded remarkable results, boosting detection rates by 60% while cutting false positives in half. Citibank's partnership with Feedzai for real-time monitoring of corporate payments illustrates the banking industry's commitment to AI-driven security solutions.

3. personalized financial services

   1. Personalization has emerged as another crucial application. Truist Financial's AI assistant has improved customer satisfaction by 15%, while Scotiabank's C.MEE platform has driven a 36% increase in customer consultations. These improvements suggest AI's ability to deliver more relevant, timely financial guidance.

4. risk management

   1. In risk management, early results are promising. Ally Financial's AI platform, Ally.ai, has achieved 82% accuracy in generating summaries without human intervention. Meanwhile, BMO has integrated AI into its anti-money laundering and know-your-customer processes, strengthening regulatory compliance while improving operational efficiency.

Popular AI agent framework for enterprise environment

As artificial intelligence becomes increasingly central to business operations, companies are seeking efficient ways to develop AI agents. While building these agents from scratch using LLM APIs is possible, it's often an exhaustive process that demands significant resources. The challenge extends beyond basic functionality – production-ready AI agents require sophisticated features like memory management, workflow orchestration, loop handling, and comprehensive monitoring systems.

For enterprise organizations venturing into AI agent development, established frameworks offer a compelling starting point. While these frameworks do add layers of abstraction, they significantly accelerate development timelines and incorporate industry best practices refined through extensive real-world implementation.

Here are three frameworks worth considering for organizations beginning their AI agent development journey:

AutoGen

AutoGen is an open-source programming framework developed by Microsoft for building AI agents and facilitating cooperation among multiple agents to solve tasks. It aims to simplify the orchestration, automation, and optimization of complex LLM workflows.

Key features of AutoGen include:

• Support for multi-agent conversations with minimal effort

• Customizable and conversable agents

• A collection of working systems demonstrating diverse conversation patterns

• Autonomous and human-in-the-loop workflows

AutoGen is particularly useful for creating next-generation LLM applications based on multi-agent conversations, maximizing the performance of LLM models while overcoming their weaknesses.

LangChain

LangChain is a framework for developing applications powered by large language models (LLMs). It simplifies every stage of the LLM application lifecycle, from development to productionization and deployment.

LangChain offers:

• Open-source building blocks, components, and third-party integrations

• LangGraph.js for building stateful agents with streaming and human-in-the-loop support

• LangSmith for inspecting, monitoring, and evaluating chains

• Tools for turning LangGraph applications into production-ready APIs and Assistants

LangChain consists of multiple libraries, including @langchain/core for base abstractions, @langchain/community for third-party integrations, and partner packages for specific integrations.

LlamaIndex

LlamaIndex is an advanced orchestration framework designed to amplify the capabilities of LLMs by bridging the gap between these models and private or domain-specific data. It offers a structured way to ingest, organize, and harness various data sources, including APIs, databases, and PDFs.

Key features of LlamaIndex include:

• Diverse data source compatibility

• Array of connectors for data ingestion

• Efficient data retrieval through an advanced query interface

• Customizable indexing options

LlamaIndex excels at indexing and structuring data, making it easily accessible and searchable by LLMs. It is particularly useful for search and retrieval applications, optimizing for quick data access and concise outputs.

These frameworks can be used individually or in combination to create powerful LLM-based applications that leverage external data sources, multi-agent conversations, and efficient data retrieval capabilities.

That's all for this week's FutureOrg AI newsletter. Remember, staying informed about AI developments isn't just about keeping up—it's about staying ahead. Have a great week, and we'll see you next week!

ps. If you like this newsletter, please forward along to a friend!

pps. If you want to subscribe, please sign up here.

pps. If this newsletter is not for you, feel free to unsubscribe. I won’t be offended and I’ll thank you for it.