Search and Conversation with Vertex AI are now accessible

For generative AI to reach its full potential, it must be widely available and easy to integrate into various systems. Our customers may use Vertex AI Search and Conversation, available today, to simplify generative search and chat apps. These products let developers without machine learning experience build and deploy intelligent apps in hours.

Vertex AI Search and Conversation, launched earlier this year in preview as Enterprise Search and Conversational AI on Generative AI App Builder, offers a simple orchestration layer to combine enterprise data with generative foundation models, conversational AI, and information retrieval technologies.

Enterprise developers can quickly ingest data, customize, and build a search engine or chatbot that interacts with customers and answers questions based on their enterprise website and specified structured and unstructured data sources in a few clicks instead of months. This capacity to swiftly develop generative apps allows corporations employ them for food ordering, banking, and customer support.

Today, we’re launching new tools to help developers build more appealing apps that allow users find critical information using natural language and take actions on their behalf. New features include:

• Multi-turn search allows for follow-up questions without restarting the interaction; conversation and search summarization provide concise summaries; and tools allow developers to pre-program prompts and responses for specific queries in natural language.

• Vertex AI extensions and data connectors enable real-time data retrieval and action across Google, Datastax, MongoDB, and Redis, connecting generative applications to enterprise systems like Salesforce, Confluence, and JIRA.

By grounding generative outputs in enterprise data, generative AI search and conversational applications can gain confidence. This data can be supplemented with foundation model training data at organizations’ discretion. Citations can also improve user confidence in the results.

Let’s examine Vertex AI Search and Conversation’s capabilities

Personalizing, appealing generative apps with Vertex AI

Vertex AI Search allows enterprises create Google Search-quality, multimodal, multi-turn search apps driven by foundation models, including the flexibility to ground outputs in enterprise data alone or to complement the foundation model’s original training. It will soon offer enterprise access controls to limit information to appropriate individuals and citations, relevance scores, and summaries to boost confidence and usefulness.

LLM embeddings and vector search can power semantic search, personalized recommendations, chat, multi-modal search, and other generative AI apps in complex use cases. Vector search provides enterprises with an easy-to-use vector similarity search solution, the same technology used to power Google Search and YouTube at scale.

Foundation models with audio and text capabilities enable Vertex AI Conversation to create natural-sounding, human-like chatbots and voicebots. With a few clicks, developers may design a chatbot based on a website or documents. Vertex AI lets developers combine deterministic workflows with generative outputs, rules-based processes with dynamic AI to create engaging, reliable apps with transaction capabilities so users can ask AI agents to book appointments or make purchases. Organizations can tailor chats using data from websites, documents, FAQs, emails, and agent conversation histories and generate interaction summaries, citations, and other data to support AI app-human agent handoffs.

• Create tailored, immersive experiences for customers and employees, transforming lengthy processes into rapid searches or conversations. Apps respond naturally to user inputs and give companies control over tone, conversation flows, data access, and more by exposing the right information in the right context, executing activities, generating graphics, recording citations, and making recommendations.

• Eliminate data chunking, embeddings, indices, and conversational decision trees. Instead, developers may use a simple orchestration interface to construct apps without coding or machine learning knowledge.

• Maintain grounding, application actions, and data control. Developers can eliminate hallucinations and increase output relevance by using enterprise data. They also manage what data apps can access via connections and what actions apps may take for users. Access controls and HIPAA, DRZ, and CMEK compliance standards assist organizations secure data. Data is held in an enterprise’s Google Cloud instance, and Google does not access or train our algorithms.

Vertex AI Search and Conversation enables faster time-to-deployment and time-to-value for search and conversational use cases, one of the most widely-applicable new AI initiatives we see enterprises exploring. The early adopters are succeeding:

• Six Flags’ Chief Digital Officer, Omar Omran, highlighted the transformative potential of Google Cloud and Vertex AI Conversation collaboration. This partnership marks a major strategic shift for Six Flags, empowering it with technology. We use Google Cloud technologies to improve park operations and provide unforgettable, personalized guest experiences. Our operations will be more agile and responsive, reinventing how we serve customers and setting new standards in the amusement park business.”

According to Harsh Kumar Sarohi, Senior Vice President, Technology, TradeIndia.com, a leading B2B portal in India, Vertex AI Search from Google Cloud has reduced search drop-off by 50% and increased user engagement by 6%. We’re also using Vertex AI Search’s analytics to identify portfolio gaps, such as top search results with no inquiries.”

• C1, a top contact center provider, utilizes Vertex AI Conversation for customer and employee support. C1 Auto Pilot, their agent assist bot, provides real-time customer contact suggestions, cross-sell/upsell opportunities, sentiment analysis, and automatic call insights summarizing to improve customer experience. According to CTO Mark Langanki, C1 The C1 Auto Pilot solution helps agents improve customer experience by reducing time spent on administrative tasks like note-taking and research and providing them with the information they need to have a more productive and engaging conversation with customers. Providing agents with customer-specific value add information has reduced call handling times.  The agent experience and consumer satisfaction should continue to improve.”

• Dan Burgin, Sr Director AI Automation, C1, highlights the quick and simple build process with Vertex AI Conversation. We were impressed with how rapidly Vertex AI Conversation lets us design and deploy chatbots while protecting client and end-customer data. We’re also finishing testing a generative AI-powered chatbot for our staff that can answer frequently requested HR inquiries on our Intranet site quickly and securely.

Start building

Search and conversation use cases allow enterprises to quickly learn and profit from generative AI. We look forward to seeing more of our customers use Vertex AI Search and Conversation to delight their customers and workers. For more information, see our product website and attend our Next ’23 presentations.

We aim to serve organizations across the spectrum of AI needs and expertise levels, so if you’re a machine learning engineer or data scientist looking to build customized applications, check out Vertex AI updates with Model Garden, foundation models, and tuning options and Colab Enterprise news.

News source: