The #1 predictor of Agentforce success: Use-case selection
In my first 3 months back in a Forward Deployed Engineering (FDE) role at Salesforce, I’ve been hands-on with the tech and up close with customers and partners rolling Agentforce out.
Pattern observed: the #1 predictor of success isn’t budget, skills, or hype — it’s use-case selection.
It doesn’t need to be flashy. It probably shouldn’t be complex. Pick something straightforward to build, delivers value quickly, and has a business case.
1️⃣ Use-Case Qualification Framework
Ask three questions before you start: - Is the use-case clearly defined with a measurable business outcome? - Is the required data available & accessible now (including licenses)? - Is there a committed timeline or clear trigger for action?
2️⃣ Start Small to Win Big
Big-bang automation sounds great — real wins come from frequent, simple, low-risk tasks.
- Avoid: complex edge cases, high-error scenarios, unclear ownership
- Prioritise: frequent actions, low variance, leveraging existing data
Upcoming article: an impact/effort matrix to rank Agentforce use-cases.
3️⃣ Build the Business Case & Get Stakeholder Support
Good use-cases are common; great ones deliver measurable business value (increased profit or reduced cost).
- What’s the measurable impact?
- Can you quantify time saved, errors reduced, productivity gained?
- Who is the executive sponsor accountable for outcomes?
Yes, it’s work — but in the AI era, the first draft can be automated.
I’ve built two custom GPTs for Use-Case Evaluation & ROI Calculation.
Disclaimer: opinions and tools are my own; not endorsed by Salesforce.
Want to validate your Agentforce use-case? DM me to try the GPTs.
Got feedback or a better way to prioritise? Comment or reach out.