Agentforce went GA October 25, 2024. We’re now six months into what has been one of the most transformative periods in the history of Salesforce. In that time, we’ve seen real wins, real roadblocks, and a clear pattern emerge for financial institutions.
What we’ve learned about Agentforce for Financial Services six months in:
- The rise of the “digital workforce”: The platform’s future centers on enabling digital workforces to drive internal cost and operational efficiencies.
- Stop building—start testing: We pivoted hard from DIY your AI to leveraging off the shelf solutions like Agentforce to start experimenting with available data sets faster.
- Compliance still rules the roadmap: Financial Services compliance and regulatory measures are limiting initial use cases to truly determine the value of agentic AI given data security requirements. We knew this already, but engagement in the last six months has confirmed it.
- Strategy before agents: Everyone needs to identify their first agentic use case (this is every second conversation we have these days). However, most Financial Institutions do not currently have operating models and processes to effectively curate ROI driven use cases.
The Age Old Adage of People, (Data), Process & Technology Still Ring True.
The tech is exciting and it’s not all that hard to implement; we’ve deployed our Agents to production in under four weeks. But adoption across the ecosystem hasn’t met expectations. Why?
PEOPLE
Challenge: You’ve invested in the tech, not change management, impacting adoption.
This is most often the case. We roll out new technologies and don’t emphasize or plan for the rollout effectively. Now, to be fair, early adopter customers are experimenting, so disruption to core operating processes likely still a ways out.
Solution: You need executive buy in and support from a limited set of human agents.
We’re seeing the most adoption success where a key executive stakeholder is both a champion for AI/Innovation and accountable for outcomes/KPI’s. Even though most of these initiatives are just proof of concepts (POCs), teams must provide basic training, change management, and internal advocacy to get POC users engaged and using the new functionality.
DATA
Challenge: Your data quality isn’t what you thought it was.
Most organizations don’t have a data volume problem—they have a data quality problem. Agentforce enables faster experimentation with CRM data, but if teams don’t follow clear processes or data governance frameworks, digital agents won’t have the inputs they need to operate effectively.
Solution: You need accurate, accessible data—structured for speed and decision-making.
Start by understanding how your CRM is set up, where data enters, and how it flows. Prioritize consistent inputs, clear ownership, and scalable integration points. Use automation to assess quality and identify gaps early. Flexible support for platforms like Salesforce Data Cloud, Databricks, Azure, and Snowflake makes it easier to unify sources, enforce governance, and get clean data into the hands of digital agents—fast.
TECHNOLOGY
Challenge: The tech is evolving faster than your roadmap.
With 4,000+ Agents sold and over 1,000 in production (as of March 1, 2025), early pilots are showing promise. But the technology is still developing. Salesforce is releasing features in phases. Many teams don’t know what’s possible today, what’s coming next, or how to plan around it.
Solution: Build a clear, realistic roadmap tied to what exists, not what’s promised.
Start by aligning with your consulting team and Salesforce account reps on current capabilities and rollout timelines. Draft a phased view of Agentic functionality to help executives see the long-term value beyond the POC. A shared roadmap improves planning, avoids mismatched expectations, and keeps everyone focused on what’s feasible now.
PROCESS
Challenge: Your first use case sets the tone.
You’re expected to move fast. Test. Iterate. But if your first use case falls flat, you’ll struggle to get support for what comes next. Most teams don’t have a clear framework for picking high-impact, feasible pilots.
Solution: Start with use cases that prove value and build momentum.
Align with executives on an AI strategy that connects today’s pilots to tomorrow’s goals. Many successful first use cases focus on middle and back office operations—feasible for POCs, but not always exciting to leadership. Pair feasibility with a clear vision of what’s next. Use workshops and analytics to map dependencies, uncover opportunities, and shape a roadmap that makes sense technically and strategically.
Next steps: Secure executive alignment and build your Agentic AI roadmap
Based on our learnings from the past 6 months, organizations who have these two things are most likely to succeed:
- A roadmap that aligns AI efforts with business goals
- A data-driven method to identify the right first use cases
Without these, you risk running into adoption gaps, poor data quality, and unclear ROI.
Start with a Data and AI Advisory Workshop
A Data & AI Advisory Workshop brings together your key stakeholders to align on priorities, assess organizational and data readiness, and identify which processes are best suited for Agentic AI. During the workshop, we introduce a practical framework to evaluate your current state and guide decisions around where and how to start.
See the full Data & AI Workshop agenda →
Shift the way you identify use cases
Too often, teams rush into pilots without fully understanding the workflows they’re trying to improve. Process optimization and Agentic AI must go hand-in-hand. Before deploying agents, you need to map your workflows, uncover bottlenecks, and surface the areas where automation will drive measurable outcomes. We recorded a walkthrough of how this assessment works in practice.
