Why Your AI Strategy Needs a Solid Data Foundation
LISA HOOPER, SENIOR SALESFORCE CONSULTANT
MARCH 17, 2024
Is your Salesforce org ready for Agentforce? Autonomous AI agents promise to enhance interactions, improve productivity and more. It’s an exciting time to see how AI can positively impact our clients. However, Agentforce needs reliable data as its source. So, before exploring what Agentforce can do, you need to assess the health of your data in the org. Specifically, is your data free of duplicates, relevant, complete and accurate?
Have you ever run a search and come up with multiple records for the exact same person? Frustrating, isn’t it? Now imagine your agent creating summaries based off of all this duplicate data. To help with avoiding duplicates, set policies in place and train those who enter data to run record searches before actually creating new records. In addition, Salesforce has Duplicate Rules/Matching Rules functionality that helps to spot and avoid duplicate record creation. The AppExchange also has additional free and paid duplicate/merge tools. Of those available, we have had good results using Apsona for duplicate management. The fewer dupes you have, the more efficient the agent can be in providing useful and reliable information.
Speaking of useful, another area to consider is the relevancy of your data. Is the data meaningful to those who use your system every day? Or, do they not bother logging into the org because they feel the information isn’t pertinent anymore? Review your data and work with your stakeholders to determine what type of records are worth keeping and enhancing and what can be either archived or removed. You can use reports and dashboards to provide summaries or metrics to better picture trends and patterns in your org. To remove or archive data, you can use tools like Data Loader or Workbench for removal and Salesforce Archive to retain historical records. In some cases, removing unwanted data can help with data storage costs.
If you run reports and dashboards to investigate relevancy, you may also find that key data is missing or incomplete. With incomplete data, the agent won’t be able to accurately provide the response you need. To remedy missing info, you can create reports or run a SOQL query to find incomplete fields. Then, you can use tools like Data Loader or Workbench to update these fields. To prevent incomplete fields from the beginning, utilize validation rules or flows to ensure data population.
Finally, look into whether your data is accurate. Even if the record has all the fields populated, it doesn’t necessarily mean the data is accurate. Bad data can lead to responses from your agent that you don’t intend. You can remedy this through the same resolution and prevention tactics mentioned earlier. Use reports and dashboards to pinpoint records with bad data, then update the data through tools like Data Loader or Workbench. Additionally, take advantage of validation rules and flows to ensure the correct values are being entered. If you have fields that are limited in potential values, use picklist fields with default values.
When I attended Salesforce’s TDX 2025 conference virtually, I was amazed at the potential of Agentforce both within customer facing applications as well as how it can help with basic administrative and development tasks. However, if you have data in your org that is not clean, relevant, complete or accurate, then that potential diminishes. Even if you decide not to utilize Agentforce, having good data stewardship practices will give you a healthier org in the long-term and will improve relationships with your customers and constituents. Your users and your IT staff will thank you.
Clearly Consulting can help with strategies for your data clean up efforts and your Agentforce implementation! Contact us for a free initial assessment!