Unleashing the Power of Generative AI for Salesforce Customization

#ai #salesforcedevelopers

The opinions in this article are my own and do not necessarily reflect the opinions of my employer.

The Three Parts of a Salesforce Developer Role

There are three parts to a Salesforce developer role:

• Prediction – How can this task be addressed?
• Judgment – Which prediction will provide the best result?
• Production – Doing the actual work.

For many Salesforce developer tasks, it is already clear that Generative AI can assist with the prediction and production activities and will only continue to get better as the specialized knowledge bases and feedback loops are established. The work for humans will be judgment e.g. what prediction will give the best result?

Leveraging Generative AI for prediction can help provide clarity, reduce complexities, and make apples-to-apples comparisons for the different options ideated during development. For example, for user stories to integrate Salesforce Sales Cloud with an Oracle ERP system, Generative AI can “predict” multiple solutions. These predictions can be based on past patterns, the architecture of both systems, standard practices for integration, project-specific architecture and design, and the API documentation for Salesforce and Oracle ERP.

For user stories integrating Salesforce with Oracle ERP, Generative AI predictions could include:

• Using a middleware solution like MuleSoft that Salesforce itself suggests for complex integrations. Mulesoft can act as a bridge between Salesforce and Oracle ERP, effectively managing data flow and transformation between the two systems.
• A direct integration using Salesforce Connect. This option leverages the built-in capabilities of both systems to allow Salesforce to pull data in real-time from Oracle.
• The use of an ETL (Extract, Transform, Load) tool to regularly import Oracle ERP data into Salesforce. Depending on the volume and complexity of the data, this could be done in batches. The ETL tool would extract data from Oracle, transform it into a format Salesforce can consume, and then load it into Salesforce.
• An event-driven architecture where Salesforce and Oracle ERP are linked through microservices and event-driven APIs. This approach is modern and scalable but might be more complex and require more infrastructure support.
• A custom-built integration. This would involve writing specific code to directly connect Salesforce with the Oracle ERP system, an option that may provide the most flexibility and, with the AI helping, could be done very quickly.

The human in the loop would look at these options and know that:

• The project doesn’t have access to middleware like Mulesoft, and the procurement lead time and license costs make middleware a non-starter.
• The project doesn’t have access to an ETL tool, and the procurement lead time makes an ETL tool a non-starter.
• The organization has not started down the microservices path and, at this point, does not have a microservices DevOps team to take on this work.

This leaves a direct integration using Salesforce Connect or a custom-built integration. Assuming the availability of Generative AI to assist with the custom-built integration, the total cost of ownership (TCO) landscape shifts favorably towards a custom-built integration. This is because the main downside to a custom-built solution - the time and effort required for development - is mitigated by the Generative AI's capabilities. A custom-built integration also bypasses the need for Salesforce Connect licenses, eliminating the procurement lead time and ongoing license costs associated with Salesforce Connect. Therefore, while even with the use of Generative AI, some upfront development is required for the direct integration approach, the total lifecycle costs will be significantly lower.

In this scenario, the remaining options can be summarized and ranked as follows:

1. Custom-Built Integration Built Using Generative AI: Given the assistance of Generative AI during development and maintenance, this becomes the recommendation. It is likely to have reasonable upfront development costs due to the efficiency of AI, and its ongoing costs will be lower without the need for Salesforce Connect licenses. Plus, it offers the most flexibility to meet specific integration requirements.
2. Direct Integration Using Salesforce Connect: This approach, while still viable, moves down to the second position due to the ongoing cost of Salesforce Connect licenses. While it might require less initial development work, the long-term cost could be more significant due to the licensing fees, particularly for a larger user base.

With Generative AI for prediction and production along with human judgment, the best option is a custom-built integration. In fact, as Generative AI custom development becomes more approachable, it will, over time, lead to faster development and higher quality code, making it a more cost-effective solution than it is today.

Where the Three Parts are Taking Us

Generative AI is poised to make a significant impact in the realm of Salesforce project design and development. By aiding in the prediction phase, Generative AI can propose multiple feasible solutions, help understand the architectural options, and offer informed insights to facilitate decision-making. From integrating Salesforce Sales Cloud with Oracle ERP to customizing specific functionalities, AI can generate a range of strategies, contributing to a more robust and efficient design process.

As the AI presents various architectural solutions, human developers can review these predictions to make a judgment about the best fit based on factors like cost, timeline, complexity, and long-term sustainability. This AI-supported approach enriches each step of the project lifecycle, from ideation to production, allowing for streamlined decision-making and reduced effort in complex tasks.

Furthermore, Generative AI's role in the production of custom development changes the dynamics of project implementation. Previously, custom solutions were often sidelined due to their extensive development timelines and high initial costs. However, with AI providing rapid and reliable custom development, this option becomes more feasible and attractive. It potentially reduces the reliance on off-the-shelf solutions and licensing costs, leading to more efficient and cost-effective Salesforce implementations. Generative AI's capabilities will revolutionize Salesforce project design, implementation, and maintenance, favoring more custom development for better-tailored, efficient, and cost-effective solutions.