Start With the Problem: A Practical Guide to Salesforce AI

Start With the Problem: A Practical Guide to Salesforce AI

For many organizations, the question is no longer whether AI has potential. The real question is where it can create practical value without adding unnecessary complexity or distancing the business from its customers. My answer is to start with a real business problem, not with the technology. Later in this post, I’ll share how one of our clients did exactly that, using an AI agent to take routine appointment calls off its contact center’s plate.

Key AI Technologies Used by Salesforce

Let’s take a closer look at some of the key AI technologies Salesforce has brought to the platform over the past few years. Each of these reflects Salesforce’s broader vision for how AI can help organizations work smarter, make better decisions, and create more personalized customer experiences.

1. Einstein AI

Einstein serves as Salesforce’s original AI layer. It analyzes customer data and provides actionable recommendations across sales, marketing, service, and commerce operations. Its conversational assistant, originally launched as Einstein Copilot, is now part of Agentforce.

Key capabilities include:

• Lead and opportunity scoring
• Predictive forecasting
• Customer behavior analysis
• Automated data entry
• Personalized recommendations

By leveraging machine learning models trained on CRM data, Einstein helps organizations identify high-value opportunities and prioritize customer interactions.

2. Agentforce

Agentforce brings Salesforce’s AI capabilities into action by helping users work with CRM data, automate routine tasks, and support customer-facing workflows through natural language and autonomous agents.

Employees can ask questions such as:

• “What are my top opportunities this quarter?”
• “Summarize this customer account.”
• “Create a follow-up email for this prospect.”

The AI assistant retrieves relevant information, generates responses, and performs actions directly within Salesforce workflows, reducing manual effort and improving efficiency.

Agentforce also brings generative AI into everyday work, helping employees create high-quality content faster while keeping customer communications consistent. Common uses include:

• Drafting emails and sales proposals
• Writing customer service responses
• Developing marketing campaign content
• Generating knowledge articles

Agentforce agents can also complete tasks on their own, such as handling a customer request from start to finish and passing it to a person when needed. The call center example later in this post shows what that looks like in practice.

3. Data Cloud and AI

Salesforce’s Data Cloud serves as the foundation for AI-driven decision-making by unifying customer data from multiple sources.

By creating a comprehensive customer profile, Salesforce enables AI models to:

• Deliver personalized recommendations
• Predict customer needs
• Improve segmentation
• Enhance customer engagement strategies

The combination of Data Cloud and AI allows businesses to generate more accurate insights and provide highly tailored customer experiences.

4. AIforce and Headless 360

Salesforce announced AIforce in September 2026 with a focus on making Salesforce capabilities available in more places, not only inside the traditional Salesforce interface. For businesses, the broader implication is that CRM intelligence can become easier to access from the tools employees already use. AIforce builds on Headless 360. Headless 360 opens the platform’s data, workflows, and actions to AI agents and developers through APIs, so they can work with Salesforce without going through the traditional browser interface.

One of the first AIforce experiences is Claudeforce, a strategic partnership between Salesforce and Anthropic that connects Claude’s advanced AI reasoning capabilities with Salesforce customer data, CRM workflows, business logic, and security controls. For business users, the value is not simply having another AI assistant. It is the ability to work with Salesforce information in a more natural way while still respecting the governance and context that enterprise systems require. Salesforce and Anthropic first announced Claudeforce in August 2026, and it is currently available to customers in beta.

Real-World Example: Using AI to Free Up the Call Center

One of our clients faced a challenge many service organizations will recognize. A large share of the calls coming into their contact center weren’t about complicated service issues. Customers just wanted to cancel an appointment or move it to another day. Each call was simple, but together they took up a significant amount of agent time, caused long hold times during busy periods, and kept the team from focusing on customers who needed real help.

To solve this, the client put an AI agent in place to handle these requests on its own. When a customer reaches out, the agent confirms who they are, finds their upcoming appointment, and asks what they would like to do. If they want to cancel, the agent confirms and updates the record. If they want to reschedule, it checks availability in real time, offers open time slots, books the new appointment, and sends a confirmation – all without a person getting involved. If the conversation goes beyond a simple change, such as a billing question or a complaint, the AI passes it to a live agent with a summary of the conversation, so the customer doesn’t have to repeat themselves.

The impact on the call center has been substantial. Routine cancellation and rescheduling calls no longer pile up in the queue, which has freed agents to spend their time on more complex conversations and shortened wait times for customers who need a person. Customers also benefit from being able to change their appointments at any time, including evenings and weekends when the call center is closed. And because cancellations are recorded right away, those open slots go back on the schedule quickly and can be offered to other customers.

What I like most about this example is that it follows the approach I described earlier. The client didn’t start with AI. They started with a clear business problem. AI was the right fit because the task is high-volume and predictable, and it relies on information the business already has. That combination is often the best place for an organization to start its AI journey.

Implementation Considerations for Businesses

Salesforce AI can deliver real value, but getting there takes planning. Before moving forward, businesses should consider:

• Data quality – Inaccurate, outdated, or duplicate data limits the value of any system, and AI is only as good as the data behind it.
• Integration – Salesforce needs to connect with the other systems the business already uses so information can be shared easily across the organization.
• Security and governance – Sensitive information must be protected and data managed properly.
• AI guardrails – Define what an AI agent is and is not allowed to do, test it thoroughly before go-live, and plan a clear handoff to a person for anything outside its scope.
• User adoption – Even a great system won’t succeed if employees don’t understand how to use it or don’t see its value.
• Overall cost – Factor in licenses, customization, data migration, training, and ongoing maintenance.

Looking at these areas ahead of time helps businesses avoid unexpected challenges and get the most value out of their Salesforce investment.

Looking Ahead

When I look at where Salesforce is heading, I see AI playing an increasingly central role in the platform. Having followed Salesforce’s AI journey over the past few years, what excites me most is how quickly the technology is moving from something organizations experiment with to becoming an integral part of how they operate.

I expect Salesforce to keep making AI more proactive and deeply integrated in everyday business processes, from anticipating customer needs before they arise to automating increasingly complex workflows. For organizations getting started, my advice is to pick one high-volume, predictable process that relies on data you already have, prove the value there, and then expand.

For me, the real value isn’t the technology itself. It’s about how AI helps people spend less time on repetitive tasks, make better decisions, and build stronger relationships with customers. We are still in the early stages of what AI can bring to the Salesforce ecosystem, and I’m genuinely looking forward to what comes next.

Interested in exploring how Salesforce AI could benefit your organization? Our team can help you evaluate opportunities, identify practical use cases, and develop an approach aligned with your business goals. Contact us to start the conversation.