HomeBlock ChainLeveraging Conversational Search for Generative AI Innovation: Introducing IBM Watsonx Assistant

Leveraging Conversational Search for Generative AI Innovation: Introducing IBM Watsonx Assistant

Published on

Generative AI has become extremely popular in the business world, with organizations worldwide trying to utilize these advancements while managing the risks associated with implementing these models on a large scale. Despite concerns such as hallucinations, traceability, data training, IP rights, skills, and costs, enterprises are still driven by the potential to revolutionize customer and employee experiences through AI. Large language models (LLMs) have been at the forefront of generative AI, transforming the way we access and interact with knowledge. Previously, enterprises relied on keyword-based search engines to support customers and employees, but now, with the introduction of conversational search in watsonx Assistant, AI Assistants can provide faster and more accurate answers by leveraging IBM Granite large language model and Watson Discovery. Conversational Search is seamlessly integrated into the augmented conversation builder, empowering customers and employees to automate answers and actions. IBM’s Granite model, which was recently made generally available, allows users to utilize a pre-trained Granite LLM model specialized for enterprises and apply it to watsonx Assistant for quick and comprehensive question-answering capabilities. Conversational Search expands the range of queries that can be handled by AI Assistants, reducing training time and increasing knowledge delivery. Users of watsonx Assistant’s Plus or Enterprise plans can now request early access to Conversational Search and can contact their IBM Representative to schedule a demo or gain exclusive access to the beta version. The behind-the-scenes working of Conversational Search involves determining how to assist a user by triggering prebuilt conversations, conversational search, or escalation to a human agent. Retrieval and generation are the two key steps for successful conversational search. IBM watsonx Assistant utilizes the Retrieval Augmented Generation framework to reduce the need for continuous training of the LLM model. For retrieval, watsonx Assistant leverages search capabilities from Watson Discovery to retrieve relevant information from business documents, enabling a better understanding of context and meaning. Once retrieved, the information is passed to an LLM (Granite model) to generate contextual responses based on the content, ensuring traceability of the answers. IBM emphasizes responsible AI usage and allows organizations to enable conversational search only for certain topics or as a fallback option for long-tail questions. The functionality can be adjusted based on corporate policies. A real-life scenario demonstrating the functionality of Conversational Search involves a customer asking about welcome offers for a credit card. The assistant routes the user’s message to Conversational Search to retrieve relevant information from the bank’s knowledge documents. If the user asks a sensitive question, the assistant can escalate to a human agent for resolution. IBM aims to drive open innovation by offering flexible deployment options for watsonx Assistant Conversational Search, including IBM Cloud, Cloud Pak for Data, and future software and SaaS deployments. Organizations can also bring their proprietary data to IBM LLM models or leverage third-party models for customization. IBM Consulting offers client briefing sessions to assist with generative AI journey for customer service.

Latest articles

Analyst Suggests Ethereum Price Could Surge to $3,100 with Bullish Momentum

Ethereum (ETH), the second-largest cryptocurrency, has seen a significant price increase over the past...

Cboe Digital Announces Plan to Introduce Margin Futures Trading for Bitcoin and Ethereum by 2024

On Nov. 13, Cboe Digital announced that it will soon launch trading and clearing...

Top Trader Predicts Significant Price Increases for Sushi and Apecoin – Check Out His Projections

A crypto strategist says more rallies are up ahead for prominent decentralized exchange (DEX)...

What is the difference between NVMe and M.2?

`` Over the last decade, solid-state drives (SSDs) have become the top choice for many...

More like this

Analyst Suggests Ethereum Price Could Surge to $3,100 with Bullish Momentum

Ethereum (ETH), the second-largest cryptocurrency, has seen a significant price increase over the past...

Cboe Digital Announces Plan to Introduce Margin Futures Trading for Bitcoin and Ethereum by 2024

On Nov. 13, Cboe Digital announced that it will soon launch trading and clearing...

Top Trader Predicts Significant Price Increases for Sushi and Apecoin – Check Out His Projections

A crypto strategist says more rallies are up ahead for prominent decentralized exchange (DEX)...
bitcoin
Bitcoin (BTC) $ 67,721.14 4.00%
ethereum
Ethereum (ETH) $ 3,775.59 8.12%
tether
Tether (USDT) $ 0.999949 0.10%
bnb
BNB (BNB) $ 420.39 1.04%
solana
Solana (SOL) $ 132.68 1.26%
staked-ether
Lido Staked Ether (STETH) $ 3,760.64 7.75%
xrp
XRP (XRP) $ 0.651759 0.00%
usd-coin
USDC (USDC) $ 1.00 0.33%
cardano
Cardano (ADA) $ 0.775808 0.35%
dogecoin
Dogecoin (DOGE) $ 0.190053 17.88%