5 questions to consider when building an agent

5 Questions Admins Should Ask When Building an Agent With Agentforce

By

It’s possible by now you’ve been to a Salesforce event or went through some Trailhead modules, and you’ve built your first agent in Salesforce. What’s even more possible is that you know about artificial intelligence (AI) and have in some way been involved in company discussions about its use. For me, those types of discussions always include something like, “Does Salesforce have anything like [insert latest tech here]?”, and in this case, yes—Agentforce.

The integration of Agentforce into the Salesforce ecosystem presents a transformative opportunity for Salesforce Admins. That feels like shiny marketing talk, but thinking honestly now about applications and agents in Salesforce and how we build them is changing things. Now that admins can build agents, the next critical step is to ask the right questions within their organizations. These inquiries are essential for aligning Agentforce capabilities with business objectives, improving user experience, and ensuring long-term success.

From today forward, as a Salesforce Admin, you’ll always have applications you built before agents and Agentforce and after. What does that mean? As you start to have AI discussions in your company, you’ll face the decision of, “Am I adding an agent to an existing application or am I building a net-new application with an agent in mind?” 

Let’s tackle some questions we need to start answering as we build agents in our organization.

1. What specific business problems are we solving with Agentforce?

As a Salesforce Admin, you’re really plugged into the organization’s processes, but with agents you must first clarify the organization’s needs. Building an agent just for the sake of technology isn’t productive; understanding the business challenges ensures the agent adds value. You should ask:

  • Which pain points do users or customers currently experience that an agent can address?
  • Which departments will benefit the most from automation through Agentforce?
  • Which key performance indicators (KPIs) will indicate that the agent is providing value?

Understanding these will help you configure Agentforce to handle the most pressing value-add tasks and align with key business outcomes.

2. What data does the agent need access to?

Data is the lifeblood of any intelligent agent. Data Cloud provides the vital customer data that grounds Agentforce, making agents more contextually aware, knowledgeable, and adaptable to customer needs. For example, when a frustrated customer contacts Agentforce Service Agents, Data Cloud’s built-in Retrieval Augmented Generation (RAG) capabilities add real-time insights and context from past emails, support tickets, product photos, voicemails, and any sources as defined by governance policies to help Agentforce better understand the customer perspective. 

Admins should evaluate the data the agent needs for effective decision-making and responses. Important questions include:

  • What data should the agent need to provide accurate responses?
  • How is the data stored, and does the agent have the necessary permissions?
  • Are there sensitive data security concerns (like GDPR or HIPAA) that need to be addressed?

This ensures that the agent operates efficiently while maintaining compliance with data privacy regulations.

3. How will we measure the success of the agent?

Admins need to have clear metrics to gauge the effectiveness of the Agentforce implementation. Performance indicators ensure the agent adds tangible value. Key questions include:

  • Which KPIs will we use to measure the agent’s success?
  • How will we track customer satisfaction and engagement with the agent?
  • What data collection mechanisms do we need to put in place to measure agent performance?

Setting benchmarks before deployment enables teams to adjust strategies and optimize the agent’s contribution to business objectives.

4. How will the agent fit into our current processes and workflows?

A seamless integration of Agentforce with existing workflows ensures a smoother transition and more immediate value. Admins need to assess how the agent will complement existing tools and processes.

  • How will the agent integrate with existing Salesforce features, such as Sales Cloud or Service Cloud?
  • Will the agent handle certain tasks autonomously or will it work alongside human agents?
  • How will the agent assist in decision-making processes for employees in customer service, marketing, or sales?

Understanding the agent’s role in existing workflows ensures that it enhances, rather than disrupts, ongoing operations.

5. What is the long-term strategy for AI adoption in the organization?

The deployment of an agent in Agentforce should be part of a broader AI strategy, not a one-off project. Admins should explore how the organization plans to scale AI usage.

  • What is the roadmap for expanding the use of AI agents in the organization?
  • How will we scale the agent’s capabilities as the organization grows or as new use cases emerge?
  • What additional AI tools (like Einstein or Tableau) will we integrate to provide even greater insights and automation?

A strategic AI roadmap ensures the agent is not just solving immediate issues but is positioned for long-term success.

Bonus question: How will we communicate the benefits of the agent to key stakeholders?

Finally, it’s critical to ensure that key stakeholders understand the value the agent brings to the organization. Admins should prepare to communicate:

  • What business value does the agent provide, and how does it align with strategic goals?
  • What success stories or case studies can be shared with stakeholders to demonstrate early wins?
  • How do we ensure organization-wide buy-in and support for ongoing AI projects?

Clear communication will help build enthusiasm for the agent’s deployment and continued development. It can come in the form of ongoing Chatter or Slack blog posts, informal lunch and learns with Agentforce demos, and staying true to the Salesforce Admin core responsibilities—meeting with users.

For Salesforce Admins who now have the knowledge to build agents in Agentforce, asking the right questions is a crucial next step. These questions will guide the organization in leveraging the full potential of AI agents, ensuring a strategic, secure, and user-friendly implementation that drives business value.

By focusing on aligning AI agents with business needs, data access, training, security, and long-term strategy, admins can ensure that their organizations implement AI and do so in a way that fosters growth and innovation.

Resources

Make Your Sales and Service Reps' Lives Easier with AI.

5 Ways to Make Your Sales and Service Reps’ Lives Easier with AI

Salesforce is ever-expanding its product offerings, and this year, the most fantastic features of the platform are all things artificial intelligence (AI). Have you started exploring capabilities? Are you truly getting the most from the features? In this post, we’ll cover five ways to make your sales and service reps’ lives easier with AI. 1. […]

READ MORE
How to Use Salesforce to Manage Your Documentation

How to Use Salesforce to Manage Your Documentation

No matter the size of your organization, or how many admins you have on your team (if you have a team!), one of the last things you probably enjoy doing is writing documentation. I hear you—it’s hard to start writing documentation, know what to capture, and figure out where to put it so you’ll actually […]

READ MORE
Prepare for the AI Specialist Exam and Agentforce

Prepare for the Salesforce AI Specialist Exam and Agentforce

As Salesforce Admins, we have clear core responsibilities that define our day-to-day roles. We manage users, data, and product, provide actionable analytics, and oversee security. But we’re currently in the third wave of the artificial intelligence (AI) revolution: contextual and generative AI, characterized by prompt-based generative AI, real-time AI applications, and autonomous agents.  This means […]

READ MORE