Salesforce bets on generative AI agents as the future of customer service

Sometimes all a customer needs is an article that tells them how to do something step by step. If this is a scenario your company is familiar with, Gen AI can help you generate automatic recommendations based on keywords, history of interactions, and similar requests from other users. The GPT in ChatGPT stands for Generative Pre-trained Transformer architecture, which is a language model capable of understanding natural language and performing related tasks. These tasks include creating text based on a prompt and engaging in a conversation with users. This need culminated in the emergence of Restricted Boltzmann Machines (Late 1990s), a genre of generative models founded on probabilistic modeling and unsupervised learning.

Sprinklr’s “call note automation” solution aims to overcome this issue by jotting down crucial information as the customer talks. Again, the contact center must plug the solution into various knowledge sources for this to happen – as is the case across many other use cases – and an agent stays in the loop. For example, in healthcare, digital assistants streamline appointments and inquiries, as seen in Memorial Healthcare Systems’ reduced call volumes. Similarly, Carbon Health reduced patient wait times and clinic answer rates by 40%. We broke down barriers with Industry Experience Clouds—an innovation that pre-designed and integrated AI specifically tailored for various verticals. Because adoption and evolution of the technology will take place almost simultaneously, generative AI will be continually disruptive.

According to

Accenture’s 2024 Technology Vision report, 95 percent of

executives believe generative AI will compel their organization to modernize their technology architecture.​ Many are turning to trusted platforms. Drift, now owned by Salesloft, is known for its ability to upgrade buyer experience and encourage prospects to make a purchasing decision faster. To proactively engage with buyers and help them make a purchase, you only have to set the high-intent buying signals in the platform. Based on previous data and new data input, Drift can also identify leads that are likely to convert with a little push.

Learn all you need to know about predictive marketing and how generative AI and a customer data platform play a role in enabling businesses to succeed. In the blink of an eye we could start to see the capabilities of AI assistants powered by GenAI change from FAQ and query support, to perhaps one day assisting in more complex query resolution. You can train your AI chatbot to understand the intent behind a question, so they can better address and answer the query. Launch regular customer satisfaction surveys with an AI chatbot that can collect responses and feedback directly in chat.

To achieve the promise of AI-enabled customer service, companies can match the reimagined vision for engagement across all customer touchpoints to the appropriate AI-powered tools, core technology, and data. Exhibit 1 captures the new model for customer service—from communicating with customers before they even reach out with a specific need, through to providing AI-supported solutions and evaluating performance after the fact. One of the major reasons why AI is being used for customer service is to improve agent experience. Call centers are known for being over-loaded with mundane and repetitive questions that can often be resolved with a chatbot. Offloading these queries to an AI chatbot or AI assistant can help improve agent experience by allowing them to focus on more complex queries and lighten their workload, which gives them more time to offer personalized experiences to users.

For the purposes of this report, we define generative AI as applications typically built using foundation models. These models contain expansive artificial neural networks inspired by the billions of neurons connected in the human brain. Foundation models are part of what is called deep learning, a term that alludes to the many deep layers within neural networks. Deep learning has powered many of the recent advances in AI, but the foundation models powering generative AI applications are a step-change evolution within deep learning. Unlike previous deep learning models, they can process extremely large and varied sets of unstructured data and perform more than one task.

Implementing Generative AI, and doing it well, will take serious commitment and prioritization. Support & Customer Success leaders will need to make strong business cases at a time when many are looking to trim costs. However, it’s important to note that this approach may be limited by the expertise of the internal team. Generative AI and ChatGPT require specialized skills and knowledge that may be limited within the organization.

Einstein Copilot can assist with tasks like answering questions using your knowledge base. Einstein Copilot uses advanced language models and the Einstein Trust Layer to provide accurate and understandable responses based on your CRM and external data. Tools like AI-powered virtual assistants are paving the way for a new era of customer and agent experiences. Generative AI-powered capabilities like case summarization save agents time while

improving the quality of case reports for the most critical hand-offs.

With generative AI’s enhanced natural-language capabilities, more of these activities could be done by machines, perhaps initially to create a first draft that is edited by teachers but perhaps eventually with far less human editing required. This could free up time for these teachers to spend more time on other work activities, such as guiding class discussions or tutoring students who need extra assistance. Researchers start by mapping the patient cohort’s clinical events and medical histories—including potential diagnoses, prescribed medications, and performed procedures—from real-world data. Using foundation models, researchers can quantify clinical events, establish relationships, and measure the similarity between the patient cohort and evidence-backed indications. The result is a short list of indications that have a better probability of success in clinical trials because they can be more accurately matched to appropriate patient groups.

Notably, these machines powered collaborative filtering, a technique that leveraged past interactions to tailor solutions for contemporary users. But one thing is for sure, generative AI helps speed up customer service and improves customer satisfaction with brands. Exploring how to implement, train, and launch an AI assistant is beneficial for any brand that is overloaded with simple queries and low CSAT scores. Generative AI carries a lot of potential when it comes to providing information fast and accurately.

Since customers can quickly access answers to their queries, and the wait times for call centers are generally reduced, time to resolution drops, making customer support a much more pleasant experience. We covered how GenAI can lower the number of mundane queries to agents and enable self-service query resolution which improves overall customer support. These are intent based chatbots that use natural language processing to interact with users. They recognize keywords and use machine learning to recognize why the end user is starting a conversation and understand patterns of behavior.

This is largely explained by the nature of generative AI use cases, which exclude most of the numerical and optimization applications that were the main value drivers for previous applications of AI. Executives estimate that 40 percent of their employees

will need new skills in the next three years due to GenAI implementation. Critical to GenAI implementation is upskilling and reskilling agents for the inevitable changes in their roles. Once you’re up and running with your monitoring and alerting, the Observability AI Assistant can help to answer any questions you have about the data you collect. Monitoring and alertingThe Support Assistant can help with providing steps for setting up monitoring for your deployment. Whether you need to configure Kibana dashboards or set up alerting for specific events, the Assistant can walk you through the necessary steps, ensuring your deployment remains healthy and issues are flagged promptly.

The launch of ChatGPT will be remembered in business history as a milestone in which artificial intelligence moved from many narrow applications to a more universal tool that can be applied in very different ways. While the technology still has many shortcomings (e.g., hallucinations, biases, and non-transparency), it’s improving rapidly and is showing great promise. It’s therefore a good time to start thinking about the competitive implications that will inevitably arise from this new technology. Many executives are wrestling with the question of how to take advantage of this new technology and reimagine the digital customer experience?

Automating Post-Call Processing

In fact, many companies are already taking concrete steps to reduce the burden on their employees. According to our Customer Service Trends Report 2023, 71% of support leaders plan to invest more in automation to increase the efficiency of their support team. Support teams facing both high-stress situations and an endless procession of repetitive tasks are often left with burnout.

Retailers can create applications that give shoppers a next-generation experience, creating a significant competitive advantage in an era when customers expect to have a single natural-language interface help them select products. For example, generative AI can improve the process of choosing and ordering ingredients for a meal or preparing food—imagine a chatbot that could pull up the most popular tips from the comments attached to a recipe. There is also a big opportunity to enhance customer value management by delivering personalized marketing campaigns through a chatbot. Such applications can have human-like conversations about products in ways that can increase customer satisfaction, traffic, and brand loyalty. Generative AI offers retailers and CPG companies many opportunities to cross-sell and upsell, collect insights to improve product offerings, and increase their customer base, revenue opportunities, and overall marketing ROI.

• All of us are at the beginning of a journey to understand this technology’s power, reach, and capabilities.
• There is no guarantee that information you provide will be kept secure or confidential.
• Many executives are wrestling with the question of how to take advantage of this new technology and reimagine the digital customer experience?
• That final part is crucial, keeping a human in the loop to lower the risk of responding with incorrect information and protecting service teams from GenAI hallucinations.

As the company behind Elasticsearch, we bring our features and support to your Elastic clusters in the cloud. Venturing into the 1990s, Recurrent Neural Networks (RNNs) surfaced as a milestone, imbuing networks with memory and temporal continuity. RNNs enabled sequential data utilization, propelling applications such as language translation, Siri’s functionality, and automated YouTube captions. The Backpropagation Algorithm (1986) emerged as a transformative breakthrough, resuscitating neural networks as multi-layered entities with efficient training mechanisms. This ingenious approach entailed networks learning from their own errors and self-correcting – a paradigm shift that significantly enhanced network capabilities.

Software engineers can use generative AI in pair programming and to do augmented coding and train LLMs to develop applications that generate code when given a natural-language prompt describing what that code should do. We estimate that generative AI could increase the productivity of the marketing function with a value between 5 and 15 percent of total marketing spending. Following are four examples of how generative AI could produce operational benefits in a handful of use cases across the business functions that could deliver a majority of the potential value we identified in our analysis of 63 generative AI use cases.

AI trained on these models can perform several functions; it can classify, edit, summarize, answer questions, and draft new content, among other tasks. The Support Assistant is the latest enhancement to the Elastic Support Hub, reflecting our ongoing commitment to empowering our customers through self-service knowledge discovery and agent-driven support cases. Accuracy has always been a priority for us, beginning nearly a year ago with our transition to semantic search, and the addition of the Support Assistant is no exception. To fully harness the power of search and drive GenAI innovation across your enterprise, we highly recommend partnering with Elastic Consulting. Whether you’re developing highly personalized ecommerce experiences or implementing interactive chatbots, our consultants have the technical expertise to design and deploy GenAI solutions tailored to your unique business needs.

24/7, real-time support

Generative AI has the potential to revolutionize the entire customer operations function, improving the customer experience and agent productivity through digital self-service and enhancing and augmenting agent skills. The technology has already gained traction in customer service because of its ability to automate interactions with customers using natural language. Crucially, productivity and quality of service improved most among less-experienced agents, while the AI assistant did not increase—and sometimes decreased—the productivity and quality metrics of more highly skilled agents. This is because AI assistance helped less-experienced agents communicate using techniques similar to those of their higher-skilled counterparts.

Foundation models and generative AI can enable organizations to complete this step in a matter of weeks. Our analysis captures only the direct impact generative AI might have on the productivity of customer operations. Like other AI tools for customer service, Ada also uses resources like repositories and guidelines to answer customer queries instantly. It is even known for engaging with customers at human-level reasoning and ensuring they don’t leave without a solution.

However, they can be difficult to find, and customers often don’t have the time or patience to search for them. Unsurprisingly, most customers end up being routed to a human agent, even for relatively simple queries; it’s often too complex to program traditional chat or voice bots to provide the right answer or think of all potential questions someone might ask. Increase customer satisfaction and reduce agent handle time with AI-generated replies on SMS, Whatsapp, and more. Use Einstein Service Replies on any channel to analyze content from customer conversations in real time.

Automated customer service interactions sometimes break down when customers change their intent halfway through a conversation – confusing the virtual agent. To automate customer queries, GenAI-based solutions drink from various knowledge sources. Technically, this works, and agents and customers can engage in phone conversations while speaking different languages. This enables the service team to prioritize actions to improve contact center journeys.

Entos, a biotech pharmaceutical company, has paired generative AI with automated synthetic development tools to design small-molecule therapeutics. But the same principles can be applied to the design of many other products, including larger-scale physical products and electrical circuits, among others. Generative AI’s potential in R&D is perhaps less well recognized than its potential in other business functions. Still, our research indicates the technology could deliver productivity with a value ranging from 10 to 15 percent of overall R&D costs. In 2012, the McKinsey Global Institute (MGI) estimated that knowledge workers spent about a fifth of their time, or one day each work week, searching for and gathering information.

Labor economists have often noted that the deployment of automation technologies tends to have the most impact on workers with the lowest skill levels, as measured by educational attainment, or what is called skill biased. We find that generative AI has the opposite pattern—it is likely to have the most incremental impact through automating some of the activities of more-educated workers generative ai for customer support (Exhibit 12). These examples illustrate how technology can augment work through the automation of individual activities that workers would have otherwise had to do themselves. Over the years, machines have given human workers various “superpowers”; for instance, industrial-age machines enabled workers to accomplish physical tasks beyond the capabilities of their own bodies.

These startups are known for their agility and innovative approach to problem-solving, making them well-suited to deliver unique solutions tailored to specific needs. They work closely with organizations to refine and enhance their technology, providing flexibility in the implementation process. Despite the large corpus of facts and answers it can generate from its training data, LLMs like GPT-4 can’t empathize with customers. This improves the efficiency of support-related processes and activities, accelerates resolution, and enables SMB to enterprise support teams to manage support ticket queues more effectively. You can foun additiona information about ai customer service and artificial intelligence and NLP. One of the biggest challenges we hear from customer service leaders is around limitations imposed by their current infrastructure.

Perhaps one of the most obvious applications – and certainly one we’re seeing enthusiastic adoption of – is chatbots. In the past, most of us will probably have experienced the frustration of dealing with slow, clumsy and far-from-intelligent voice recognition and automated customer support technology. Today, thanks to the application of chatbots built on LLMs, bots can have conversations that are close to being as dynamic and flexible as those of humans. Holistically transforming customer service into engagement through re-imagined, AI-led capabilities can improve customer experience, reduce costs, and increase sales, helping businesses maximize value over the customer lifetime. Humans still and will always likely play a major role in training, assisting customers, and ensuring that AI responses are accurate, relevant, and reliable for customer service. Generative AI can also help streamline business processes to make customer support agents more efficient at their job.

The Search AI Company

Empower agents to review, edit, and save these summaries to feed your knowledge base. The analyses in this paper incorporate the potential impact of generative AI on today’s work activities. They could also have an impact on knowledge workers whose activities were not expected to shift as a result of these technologies until later in the future (see sidebar “About the research”).

Instead of hard-coding information, you only need to point the agent at the relevant information source. You can start with a domain name, a storage location, or upload documents — and we take care of the rest. Behind the scenes, we parse this information and create a gen AI agent capable of having a natural conversation about that content with customers. It’s more than “just” a large language model; it’s a robust search stack that is factual and continually refreshed, so you don’t need to worry about issues, such as hallucination or freshness, that might occur in pure LLM bots. While traditional AI approaches provide customers with quick service, they have their limitations. Currently chat bots are relying on rule-based systems or traditional machine learning algorithms (or models) to automate tasks and provide predefined responses to customer inquiries.

• While traditional AI approaches provide customers with quick service, they have their limitations.
• While the technology still has many shortcomings (e.g., hallucinations, biases, and non-transparency), it’s improving rapidly and is showing great promise.
• Vertex AI data connectors help your applications maintain freshness and extend knowledge discovery with read-only access to enterprise data sources and third-party applications like Salesforce, JRA or Confluence.
• A generative AI bot trained on proprietary knowledge such as policies, research, and customer interaction could provide always-on, deep technical support.
• Thus, significant human oversight is required for conceptual and strategic thinking specific to each company’s needs.

The Support Assistant is designed to enhance our customers’ Elastic technical product knowledge, and its accuracy is continually being refined. However, as with all AI tools, users should exercise caution, as responses may vary. It is recommended to verify the information provided with source documentation to ensure accuracy. In a support context, this means it can quickly analyze large volumes of tickets or inquiries, categorizing them according to the sentiment of the customer. This could even take place in real-time, for example, by guiding human agents on how to respond during person-to-person interactions. They can be continuously kept up-to-date with the latest developments in best practices so that human agents will always have access to the most current information and insights.

Frank Rosenblatt’s creation of the Perceptron (1958) introduced a single-layer neural network with the ability to learn and make decisions based on input patterns. This innovation hinted at the expansive array of potential applications, including image recognition, but it wasn’t without limitations. To leapfrog competitors in using customer service to foster engagement, financial institutions can start by focusing on a few imperatives. Brands that need a chatbot to handle FAQ use cases on a large scale and offer human-like responses. Conversational experiences and generative AI are all the rave these days, and they have proven to be a game-changer for many businesses. Indeed, this list of generative AI use cases for customer service originally included 20 examples.

By comparison, the bulk of potential value in high tech comes from generative AI’s ability to increase the speed and efficiency of software development (Exhibit 5). Generative AI has taken hold rapidly in marketing and sales functions, in which text-based communications and personalization at scale are driving forces. The technology can create personalized messages Chat GPT tailored to individual customer interests, preferences, and behaviors, as well as do tasks such as producing first drafts of brand advertising, headlines, slogans, social media posts, and product descriptions. In other cases, generative AI can drive value by working in partnership with workers, augmenting their work in ways that accelerate their productivity.

You always need to vet answers, except for basic queries that require linear, straightforward replies. This way, you can educate customers and provide proactive customer support to preempt known issues before they raise them. Ultimately, average handle time is something of a paradox—the more calls your agents can cram into a day, the better. So, balancing speed and quality conversations is basically impossible without hiring more agents. These digital assistants enable end-users and provide customer self-support that provides a better overall customer experience, reduces time-to-resolution, and deflects support tickets. When it comes to making communication easier during complex calls, generative AI truly shines.

Thanks to multi-modal foundation models, your virtual agents or chatbots can have conversations that include voice, text, images and transactions. With the call companion feature in Dialogflow CX (in preview), you can offer an interactive visual interface on a user’s phone during a voicebot call. Users can see options https://chat.openai.com/ on their phone while an agent is talking and share input via text and images, such as names, addresses, email addresses, and more. They can also respond to visual elements, such as clickable menu options, during the conversation. Improved customer experience and more time for human agents to handle complex calls.

If you’re going with a pre-integrated generative AI assistant (from Zendesk, Intercom, HubSpot, etc.), you may be able to skip this step since your customer conversations and help library live on the same platform, which your AI assistant has easy access to. If you grant it access to your customer database, an LLM can use customer data, such as purchase history and demographics, to customize help experiences, offers, and follow-ups better than a human agent can. With a sufficiently large trough of data, generative AI-powered support engines can suggest complementary purchases, seasonal gifts, discounts, etc., customized to individual customers.

This makes it an invaluable tool in various applications, including image and video generation, natural language processing (NLP), and music composition. Put together, next-generation customer service aligns AI, technology, and data to reimagine customer service (Exhibit 2). That was the approach a fast-growing bank in Asia took when it found itself facing increasing complaints, slow resolution times, rising cost-to-serve, and low uptake of self-service channels. The most mature companies tend to operate in digital-native sectors like ecommerce, taxi aggregation, and over-the-top (OTT) media services. In more traditional B2C sectors, such as banking, telecommunications, and insurance, some organizations have reached levels three and four of the maturity scale, with the most advanced players beginning to push towards level five. These businesses are using AI and technology to support proactive and personalized customer engagement through self-serve tools, revamped apps, new interfaces, dynamic interactive voice response (IVR), and chat.

The pace of workforce transformation is likely to accelerate, given increases in the potential for technical automation. All of us are at the beginning of a journey to understand generative AI’s power, reach, and capabilities. This research is the latest in our efforts to assess the impact of this new era of AI. It suggests that generative AI is poised to transform roles and boost performance across functions such as sales and marketing, customer operations, and software development.

Ways To Prep Your Knowledge Base for AI Grounding

At Next ’23, we also launched a CCAI-P “Intelligent Virtual Agent only” option, which gives you a way to access all of our gen AI services with a light touch pipeline from your existing contact center to Google Cloud. This feature allows you to work with whatever infrastructure you have, whether you are on-premises or using a CCaaS platform outside of the Google Cloud partner program. Additionally, we offer ongoing support and optimize operational and outcome metrics to measure Generative AI ROI accurately for strategic decisions. This article will discuss the key metrics, KPIs establishing guides, and strategies to maximize the return on investment when implementing Generative AI for businesses. Find out how Service Cloud helps you deflect 30% of cases and deliver value across your customer journey with CRM + AI + Data + Trust.

This beats the typical chatbot workflow that requires customers to go through an elimination process to narrow down their question. Large language models can be trained on all your support tickets to date to ‘learn’ where to classify specific queries based on the words referenced against previous tickets. It can also create an algorithm to automatically segment your support tickets into priority levels. Connecting to these enterprise systems is now as easy as pointing to your applications with Vertex AI Extensions and connectors. Agent Assist is easy to deploy, requires almost no customization work, and operates in a Duet mode with a human agent in the middle — so it’s completely safe. It delivers measurable value across KPIs like agent handling time, CSAT (customer satisfaction score), and NPS (net promoter score).

Unfortunately, there are seemingly no purpose-built solutions for contact centers quite yet. Still, Google has pledged to make such a feature available on its Google Contact Center AI Platform soon. Because they leverage speech-to-text to create a transcript from the customer’s audio. It then passes through a translation engine to pass a written text translation through to the agent desktop. That makes it easier for future agents – handling follow-ups – to get to grips with what happened on the previous call.

Since it is powered by generative AI, it can create and customize responses based on a ticket’s content. Zia is also known for its sentiment analysis capabilities, where it dives into the feelings of every ticket and accordingly creates empathetic responses for customers. After all, chatbots are a flagship use case for generative AI, and the process of transitioning from human agents to automated systems began long before the emergence of language models (LLMs). This AI-driven system provides smart responses akin to human intelligence, enabling businesses to engage in dynamic and personalized conversations with their customers.

In the life sciences industry, generative AI is poised to make significant contributions to drug discovery and development. This analysis may not fully account for additional revenue that generative AI could bring to sales functions. For instance, generative AI’s ability to identify leads and follow-up capabilities could uncover new leads and facilitate more effective outreach that would bring in additional revenue. Also, the time saved by sales representatives due to generative AI’s capabilities could be invested in higher-quality customer interactions, resulting in increased sales success. For one thing, mathematical models trained on publicly available data without sufficient safeguards against plagiarism, copyright violations, and branding recognition risks infringing on intellectual property rights. A virtual try-on application may produce biased representations of certain demographics because of limited or biased training data.

Improve search efficiency for agents and customers with AI-powered Search Answers. Quickly generate answers from your trusted knowledge base and display them directly in the search page or agent console. Agents can find results faster with better filtering and support for multiple languages. Customize Einstein Search to match your specific knowledge parameters for optimal results. Generative AI could have an impact on most business functions; however, a few stand out when measured by the technology’s impact as a share of functional cost (Exhibit 3).

Automation potential has accelerated, but adoption to lag

Textbook publisher Wiley implemented Agentforce in time for the back-to-school season, when customer service volumes reach their peak. The company reported a double digit percentage increase in customer satisfaction and deflection rates compared to older technology, alongside a 50% increase in case resolution, due to the help of AI agents, according to Benioff. The McKinsey Global Institute began analyzing the impact of technological automation of work activities and modeling scenarios of adoption in 2017.

But this doesn’t stop the life insurance company from embracing the latest technology. As your generative AI model goes into general availability, you’ll uncover more bugs, errors, and exceptions in the wild. But, you can think of the post-deployment stage as more of an iterative learning process where you observe, refine, and update your generative AI capabilities to fit your agents’ workflows and answer customer queries more accurately.

This can be particularly helpful when you aren’t sure where to find a specific error. Instead of searching the Kibana docs for an error that is actually for Elasticsearch, the Assistant can save time by figuring out the appropriate context for you. The Elastic Support Assistant is now available in the Support Hub for all Elastic customers with either a trial or an active subscription. Unlock the power of real-time insights with Elastic on your preferred cloud provider.

We also offer extensions for first-party applications like Gmail, Drive, BigQuery, Docs and partners like American Express, GitLab, and Workday. Vertex AI data connectors help your applications maintain freshness and extend knowledge discovery with read-only access to enterprise data sources and third-party applications like Salesforce, JRA or Confluence. These connectors index your application data so you’re always surfacing the latest information to your users.

However, workers will need support in learning new skills, and some will change occupations. If worker transitions and other risks can be managed, generative AI could contribute substantively to economic growth and support a more sustainable, inclusive world. Conversica is a conversational AI that intercepts any stage of the sales funnel and provides support that encourages people to make purchase decisions faster. This revenue digital assistant never leaves your leads behind, allowing you to explore untapped potential sales opportunities hassle-free. It not only engages with leads but also helps you verify if they can be converted into customers or not. This is the perfect tool to bring support and sales teams together and deliver the best SQLs to the team.

By pairing this with the Cognigy Playbooks reporting platform, service teams can verify bot flows, validate outputs, and add assertions. Like Nuance and Google, Cognigy has pushed the boundaries of generative AI innovation in customer service, as its “Conversation Simulation” tool exemplifies. As a result, the GenAI application has something to work from – as do live agents during voice interactions –enhancing the contact center’s knowledge management strategy. Elsewhere, a Japanese telecoms provider is trialing a similar software that modifies the tone of irate customers.

After an agent closes a case, she may enter case notes, but these notes can get lost in the ether and other agents may end up problem-solving similar issues from scratch, not knowing their colleague had already solved it. With nearly half of customers citing poor service experiences as the main reason they switched brands last year, the pressure is on for companies to find a better way forward. This can cause latency issues, where the model takes longer to process information and delays response times. With 90% of customers stating instant responses as essential, the response speed can make or break the customer experience. Don’t have the time to work out every single way a customer might ask for a return?