PwC and Salesforce

From scripted chatbots to intelligent agents

hero image
  • Insight
  • March 25, 2025

How the multifaceted AI landscape provides opportunities to advance efficiency and customer satisfaction

AI-powered customer engagement is undergoing a significant transformation. First, chatbots changed the game and now AI-powered agents have introduced a whole new rulebook. Even organizations that haven't fully embraced AI will likely be affected by this shift to cloud-powered, AI-first business, and should explore how AI impacts their business models, service provider relationships, stakeholder expectations, reporting requirements, and more.

All this may have your organization wondering — which technology fits your needs right now and how can you better prepare to scale when the time is right?


AI agent opportunities

  • Hospitality: Recognizing guests, anticipating customer intent, and personalizing interactions
  • Finance: Tailoring relevant recommendations based on goals and needs
  • Manufacturing: Predictive maintenance, process optimization, sales support
  • Consumer goods: Managing marketing campaigns, generating promotional content
  • Healthcare: Patient questions, treatment plans, and scheduling; gleaning provider network insights

The evolution from chatbots to AI agents

AI-enabled chatbots excel at handling defined workflows and simple customer service tasks. They operate within carefully structured parameters, using natural language processing (NLP) to match customer inquiries with predetermined yet natural responses.

This approach has proven valuable for organizations seeking to reduce the burden on customer service reps by handling routine customer interactions such as common questions, returns and refunds, and product recommendations. However, it does require significant effort to train and maintain the bot in matching utterances to intents.

Autonomous and assistive agents introduce a more flexible, sophisticated solution that can respond to complex, unanticipated customer requests without human intervention. In contrast to chatbots, agents rely on pretrained Large Language Models (LLMs) that better match customer language to a topic. This allows greater reasoning and a less scripted dialog flow.

AI agents used for customer service generally fall into two categories:

  • Autonomous agents interact directly with external audiences, leveraging large language models (LLMs) to comprehend and respond to nuanced customer inquiries. They can process natural descriptions of detailed topics, execute multistep workflows, adapt responses based on Customer Relationship Management (CRM) data and real-time needs, and escalate queries to employees as needed.
  • Assistive agents provide internal support to employees by rapidly retrieving information across systems, automating routine tasks, providing decision support, and assuring consistency and awareness across customer interactions. This not only reduces call handling time but improves the customer experience by informing the agent's view.

Strategic considerations for AI-enabled customer engagement

As many organizations weigh the benefits of these technologies, there are a few key differentiators to help identify the ideal solution for a particular challenge or for where you sit on the AI-adoption continuum.

You may find that some interactions can be well managed by chatbots, whereas others may be better supported by autonomous or assistive agents. In fact, the two can work together to complement one another and support a variety of use cases.

How are companies currently leveraging AI innovations?

  Chatbots AI Agents
Use case complexity Standardized processes; Well-defined customer journeys and responses Complex, variable workflows requiring adaptive reasoning
Data compatibility Structured CRM data and explicit retrieval processes Flexible retrieval of internal and third-party data across platforms; Process structured and unstructured data through retrieval augmented generation (RAG)
Implementation Efficient for specific, well-defined use cases and repetitive, narrow tasks Preconfigured agents for common sales, service, marketing, and commerce use cases; Holistic, transformational approach to customer engagement
Organizational readiness Focus on basic automation, conversation flows, and cross-functional alignment, requiring minimal AI or data maturity Require advanced AI infrastructure, robust data governance, continuous optimization, and enterprise-wide stakeholder engagement for complex, adaptive workflows

Delivering value across the enterprise

Knowing when and how to progress to AI-enabled chatbots and agents requires the ability to recognize areas that can offer the greatest potential for value generation. PwC and Salesforce have guided companies in a wide range of industries to achieving greater cost savings, product innovation, productivity, and performance.

This often involves PwC and Salesforce industry specialists working with client teams to pinpoint where they sit on the AI continuum, outline stakeholder and customer needs, then identify the ideal technologies, providers, and architectures for keeping pace with AI as it continues to rapidly advance. Together, we lay out use cases and opportunities to automate interactions via Einstein Bots or Agentforce agents. Many companies move into AI agents internally at first, allowing the combined team to test, refine, and prove value before scaling to external-facing agents in line with a client's tech environment, architecture, culture, risk factors, and brand.


Industry use cases

  • Hospitality: A large hospitality company streamlined management of their brand standards across their global portfolio by deploying agentic workflows. Intelligent agents now automate updates, approvals, and compliance tracking—reducing review times by up to 94%.
  • Finance: Agentforce is generating transcripts, keyword insight analysis, and voice/video call analysis to improve interactions and productivity.
  • Utilities: AI-powered agents help customers understand usage patterns, identify cost-saving opportunities and identify opportunities for energy efficiency in real time.
  • Healthcare: Autonomous agent support drives increased self-service and call deflection, allowing more thoughtful, personal engagement for complex issues and improving registration and retention of donors.
  • Consumer goods: Operators are assisted by AI agents in managing orders, profile changes, and other simple yet time-consuming requests, lowering high call volumes.

 


Because Salesforce solutions are purpose-built to serve the holistic continuum of AI innovations, Einstein Bots and Agentforce readily leverage connected data to inform customer interactions. In line with the other Salesforce AI solutions, Einstein Bots and Agentforce also can provide a thoughtful, measured approach to risk, benefitting from holistic governance frameworks, seamless employee handoff protocols, integration with existing security infrastructure, and thorough audit trails.

Whether through chatbots or intelligent agents, AI solutions can provide greater efficiency for employees and teams, as well as the ability to shift their attention to more high-value tasks. At the same time, immediate, context-aware responses significantly help increase customer satisfaction.

Together, PwC and Salesforce empower organizations so they can effectively navigate the new AI-enabled landscape, thanks to innovative Salesforce technology coupled with PwC's implementation experience and industry-leading insight.

Contact us

Timothy Watt

Timothy Watt

Principal, Salesforce Practice, PwC US

Deepa Iyer

Deepa Iyer

Principal, Salesforce Practice, PwC US

Follow us