Today, Google revealed that it is integrating its conversational,Bard chatbot into numerous of its most well-known applications. This will enable users to search Google Docs and Youtube, receive human-like answers to their queries, and summarize Gmail messages.
Bard faces fierce competition from other leading generative AI solutions, such as Microsoft’s Bing conversational AI tool and OpenAI’s ChatGPT. Through its own chatbot assistant, Copilot, Microsoft has integrated ChatGPT capability to its range of Microsoft 365 business productivity and collaboration products.
Forrester Research principal analyst Max Ball says that although Google and Microsoft are pursuing internal use cases for chatbots to boost staff productivity, Salesforce is now vying for market share in this area.
“Talking with a knowledgeable assistant who can answer all of your questions and compile the information for you in an acceptable way is beneficial. However, the value increases significantly when systems are able to execute real activities and retrieve particular data that is typically hidden in back-end systems, according to Ball.
Salesforce introduced the most recent iteration of Einstein Copilot, a conversational assistant integrated into each Salesforce application, earlier this month. Through the app, customers may ask questions in natural language and receive responses based on company data that is private to Salesforce Data Cloud.
Google claimed that its new Bard Extensions function as a natural language assistant that can search for information across Google products, similar to Copilot, which can automate activities and serve as an assistant in Word, Outlook, Teams, Excel, PowerPoint, and other applications. Not only can users search Google Docs, but they can also summarise Gmail messages, assist with trip planning, and use Google Flights to check airline fares.
Using Google Maps to find directions to the airport, searching for real-time flight and hotel information, watching YouTube videos of things to do at the destination, and more, a user can now ask Bard to do various tasks within a single conversation when planning a trip.
Google stated in a blog post that users may ask Bard to create a trip planning document for them and their friends, write up an online marketplace ad, or assist in explaining a scientific concept to their children.
Google has previously experimented with genAI assistants. The business declared last month that Duet AI for Workspace productivity apps was now widely accessible. The Duet AI features will set users back $30 a month, matching the cost of rival Microsoft’s planned Copilot program.
The Bard chatbot is available in forty languages, but at first, Extensions will only support English-language inquiries.
Users can choose Bard’s “Google it” button to verify the search engine’s preliminary results when they receive English-language responses. Google stated, “When you click on the “G” icon, Bard will read the response and determine whether there is content available online to support it.” “You can click the highlighted phrases to learn more about the information found by Search that supports or contradicts a statement when it can be evaluated.”
A user can ask Bard more questions about a subject, carry on the conversation, or utilize the exchange as a springboard for their own thoughts when they share a Bard chat with someone via a public link.
Google updated the PaLM 2 large language model, the chatbot’s platform, which gives Bard more accuracy in its responses. This version includes the new functionality with Bard Extensions.
According to Forrester’s Ball, chatbot technology is still a long way from reaching maturity and being able to do more complex tasks like ordering products, transferring money across accounts, and scheduling appointments and travel.
Ball stated, “Shopping cart management and CRM system updates will require additional functionality, even though generative AI is fantastic at talking.”
The expected trajectory for chatbot maturation is as follows:
Building a generative artificial intelligence system that can communicate with customers to learn what they need to know or desire to accomplish.
Establishing a connection with a CRM or other back-end systems in order to collect data or carry out transactions.
Determining whether a human agent is required and transferring all necessary data to that agent.
Proactively notifying a consumer that something has changed and asking them how they would want to react. (Since you will be visiting the lake house over the weekend, should we forward your package, which is scheduled to arrive two days late?)
“When you start considering the possibility of managing the LLMs or generative AI systems when different ones are better at different tasks, it gets bigger and more complex,” Ball stated.
