Using ChatGPT to Write a Project Definition for a Salesforce Project

#Salesforce #ChatGPT #GenerativeAI

Authors: Joseph Kubon, Andy Forbes

The opinions in this article are the authors' and do not necessarily reflect the opinions of their employer.?

Starting with a concise statement like "SAP is the system of record for accounts and contacts. Keep Salesforce Accounts and Contacts in sync with the Accounts and Contacts in SAP" might appear straightforward but encapsulates a range of complexities that demand a nuanced approach. Utilizing a tool like ChatGPT can help you refine these simple sentences into a comprehensive project definition, and the process can be and should be collaborative by involving key stakeholders such as the product owner, business analyst, architect, subject matter experts, and end-users.

In the initial phase, ChatGPT can aid the business analyst and product owner in expanding these statements into more elaborate requirements. For instance, the AI could generate questions or considerations about the type of synchronization needed (real-time versus batch), the data fields that need mapping, and security compliance standards. These suggestions could form the basis for stakeholder meetings, where experts can validate or revise the machine-generated content. In a Scaled Agile environment, these can be treated as Epics and refined iteratively throughout your SAFe Program Increments.

For architects and developers, ChatGPT can offer actionable insights into possible architectural designs, suggesting whether out-of-the-box connectors would suffice or if custom Apex triggers and classes might be required. It can also draft API prototypes, discuss data validation approaches, and even throw in snippets of code that adhere to Salesforce best practices. For subject matter experts and end-users, ChatGPT can assist in breaking down technical jargon into easily digestible explanations, providing clarifications on how the integration would affect their day-to-day interactions with either Salesforce or SAP. By presenting these synthesized insights, the tool ensures that both technical and non-technical stakeholders are aligned in their understanding and expectations from the project.

To evolve a two-sentence project definition into a comprehensive blueprint, you can initiate the process by asking ChatGPT specific questions aimed at project elaboration. For instance, with a statement like "SAP is the system of record for accounts and contacts. Keep Salesforce Accounts and Contacts in sync with the Accounts and Contacts in SAP," you might ask ChatGPT, "What are the key considerations or questions we should address to start defining this project in detail?" ChatGPT could then generate a set of targeted queries or discussion points around areas such as synchronization methods, data mapping, security, and compliance. These questions serve as a foundational layer, paving the way for stakeholder interviews and collaborative discussions. By leveraging ChatGPT's capabilities in this manner, you efficiently kickstart the project definition phase, ensuring that it's both technically rigorous and aligned with business objectives.

Prompt:

We are starting with a two sentence project definition - "SAP is the system of record for accounts and contacts. Keep Salesforce Accounts and Contacts in sync with the Accounts and Contacts in SAP" - and would like to write a project definition that will be comprised of two parts: a natural language description of the desired feature and a list of the future state assumptions. Please write the list of questions the business analyst and Salesforce architect should ask the product owner, subject matter experts, and user to begin to gather the needed information.        

ChatGPT Response:

To transition from a high-level definition to a nuanced, actionable project blueprint, the business analyst and Salesforce architect should engage the product owner, subject matter experts, and users with a carefully curated set of questions. The intent is to translate that two-sentence directive into both a comprehensive natural language description and a list of future state assumptions.

1.????? What specific data fields in SAP Accounts and Contacts need to be synchronized with Salesforce?
2.????? What are the business processes that rely on SAP and Salesforce being in sync?
3.????? What is the acceptable latency for data synchronization between SAP and Salesforce?
4.????? Are there any data transformation requirements when syncing between the two platforms?
5.????? What security protocols need to be in place for the data transfer?
6.????? Are there any compliance regulations or internal policies that the integration must adhere to?
7.????? What error-handling and notification mechanisms are required?
8.????? Are there any scalability considerations? How much data volume is expected?
9.????? How will success be measured for this integration?
10.? What are the failover procedures if real-time syncing is disrupted?
11.? How should the system behave in case of conflicting data between Salesforce and SAP?
12.? Who are the key stakeholders who will need access to reconciliation reports?
13.? Are there any specific user roles in Salesforce that need to be considered during integration?
14.? What is the frequency and format for project updates or sprint reviews?
15.? Are there any dependencies, internal or external, that could impact the timeline?

Through a dynamic dialog facilitated by these questions, you can derive the in-depth insights required to draft both the natural language feature description and the future state assumptions. This will ensure that the project definition aligns tightly with technical capabilities and business imperatives, setting the stage for successful delivery using the Scaled Agile Framework.        

While generative AI tools like ChatGPT are poised to revolutionize the approach to project definition, the sheer depth and complexity involved mean that a comprehensive guide would exceed the constraints of a LinkedIn article. For the purpose of this demonstration, we've provided answers to an initial set of foundational questions to showcase what generative AI can contribute. It's important to note, however, that in a real-world setting, the answers to these questions would likely lead to further queries and deeper lines of investigation. The interaction with generative AI in this domain is inherently iterative, with each round of answers giving rise to new, more nuanced questions that help refine the project definition and align it more closely with both business objectives and technical possibilities.

Prompt:

The answers to the questions are below. Please ingest these answers, but it is not necessary to answer them.

1.????? What specific data fields in SAP Accounts and Contacts need to be synchronized with Salesforce? Answer: Account name, address, phone number, URL, D&B number, and SAP “Account Identifier”. Contact last name, first name, email, phone, and SAP “Contact Identifier”.
2.????? What are the business processes that rely on SAP and Salesforce being in sync? Answer: Sales, Service, Ordering, and Finance.
3.????? What is the acceptable latency for data synchronization between SAP and Salesforce? Answer: Minutes, preferably not hours.
4.????? Are there any data transformation requirements when syncing between the two platforms? Answer: Yes, but minor. Phone numbers are stored formatted in one system and unformatted in the other.
5.????? What security protocols need to be in place for the data transfer? Answer: Will need to be defined during the project.
6.????? Are there any compliance regulations or internal policies that the integration must adhere to? Answer: Yes, GDPR.
7.????? What error-handling and notification mechanisms are required? Answer: Errors are logged locally to each system and in batch mode sent to Splunk.
8.????? Are there any scalability considerations? How much data volume is expected? Answer: Average Accounts created / updated each day is 50. Peak could be 50,000. Average Contacts created / updated each day is 500. Peak could be 500,000.
9.????? How will success be measured for this integration? Answer: Recurring audit finds less than 10 accounts and contacts that are out of sync.
10.? What are the failover procedures if real-time syncing is disrupted? Answer: None. Will wait until systems are back online.
11.? How should the system behave in case of conflicting data between Salesforce and SAP? Answer: SAP is the master.
12.? Who are the key stakeholders who will need access to reconciliation reports? Answer: IT Dept.
13.? Are there any specific user roles in Salesforce that need to be considered during integration? Answer: Sales Rep, Sales Manager.
14.? What is the frequency and format for project updates or sprint reviews? Answer: A program iteration is comprised of five (5) two (2) week sprints.
15.? Are there any dependencies, internal or external, that could impact the timeline? Answer: While not currently in use, there is a project to stand up a middleware application, and the SAP team may need to support that project.        

ChatGPT Response:

Your answers have been ingested and will serve as the bedrock for our ongoing discussion. They provide a granular understanding of the requirements, constraints, and objectives that need to be considered for the Salesforce-SAP integration project. From data fields and latency requirements to compliance regulations and error-handling protocols, this information gives us a detailed snapshot that will be invaluable for shaping the project's trajectory. I'm prepared to delve deeper into any technical or business aspect, so we can construct an integration strategy that aligns with both your business goals and the technical landscape.        

It's crucial to understand that the utilization of ChatGPT for crafting a project summary and defining future state assumptions is inherently iterative. As the project progresses and more nuanced information is gathered from stakeholders or as a result of technical evaluations, the brief can be revisited and refined using ChatGPT. Each new round of iteration would lead to a more precise, actionable, and context-aware project description and set of assumptions. This iterative approach ensures alignment with both tactical and strategic objectives, thereby minimizing risks and enhancing the quality of the integration. For the scope of this article, let's proceed with synthesizing a project summary and future state assumptions based on the current set of answers we have.

Prompt:

Using what has been provided so far, please write a project summary and the future state assumptions.        

ChatGPT Response:

Project Summary:

The core objective of this project is to integrate Salesforce with SAP, specifically targeting the synchronization of Accounts and Contacts. SAP will function as the system of record. The integration aims to support key business processes such as Sales, Service, Ordering, and Finance by ensuring a near real-time data sync, with an acceptable latency of minutes rather than hours. Custom Apex code and standard configurations will be developed to map specific data fields like Account name, phone number, and SAP identifiers. Compliance with GDPR will be mandatory, and all data transmissions will be secured following protocols to be determined. Error logs will be generated locally and batch-processed to Splunk. The project will be executed following the Scaled Agile Framework, with program iterations consisting of five two-week sprints.

Future State Assumptions:

1.????? The integrated system will manage an average daily volume of 50 Accounts and 500 Contacts, with the capability to handle peak loads up to 50,000 Accounts and 500,000 Contacts.
2.????? Data transformation requirements are minimal, primarily focusing on phone number formatting.
3.????? Error-handling mechanisms will log issues locally and batch-send them to Splunk for further analysis.
4.????? GDPR compliance measures will be strictly enforced, ensuring data integrity and security.
5.????? A real-time sync will be the primary data exchange mechanism, without a defined failover procedure. Should real-time syncing be disrupted, the system will await reconnection.
6.????? Given that SAP is the system of record, any data conflicts will be resolved in favor of SAP.
7.????? Access to reconciliation reports will be restricted to the IT Department.
8.????? Success will be evaluated based on recurring audits, with the aim of finding fewer than 10 Accounts and Contacts out of sync between the two platforms.

This multi-faceted approach ensures that we are not only solving for immediate needs but also establishing a foundation that is scalable, secure, and adaptable to future challenges and changes.        

In this article, you have learned the process of using ChatGPT to define a project for Salesforce-SAP integration, starting from a mere two-sentence requirement. Through a series of questions and answers, you have gained an understanding of how ChatGPT can be employed to distill complex project information into a cohesive and comprehensive project summary and future state assumptions. The process demonstrates that the role of generative AI is highly iterative, allowing teams to refine and adjust their project scope and requirements as new information is gathered. Importantly, the article also emphasizes that while AI like ChatGPT significantly streamlines the project definition phase, the task still involves much more than can be covered in a single article. This underlines the idea that the real-world application of generative AI is a collaborative, ongoing process that complements human expertise rather than replacing it.