GPT Dreamin' About Business Analysis
This is When it all Began
AI Force and Elements hosted the first GPT Dreamin’ Conference on August 7-11 to “share, challenge, learn and grow by solving real-world problems using GPT for managing Salesforce.”
Vernon Keenan, CEO and Founder of Keenan Vision, offered a long-term perspective in the keynote presentation, “The GPT Revolution and Implications for the Salesforce Ecosystem.”
Vernon presented a technology revolution timeline, approximated below:
Early 1980s: Personal computers
Mid-1990s: Worldwide web
Mid-2000s: Smartphones and cloud computing
Early 2020s: Generative artificial intelligence
Each revolution fueled the next, with personal computers initially accessing the worldwide web, then smartphones accessing services, like social media over web-based cloud computing. We’ll look back at 2023 as the year generative AI began changing everything.
Though artificial intelligence has been around for decades, ChatGPT sparked an explosion of interest in 2022. It recalls the reaction when PCs first burst on the scene:
Wow! This is awesome!
What can we do with it?
How will this promising new technology work with what we have in our organization?
GPT Dreamin’ demonstrated how the latest promising technology, generative AI, can integrate safely with Salesforce.
Business Analysis is THE Superpower
In the session “Ride the Wave of GPT Without Drowning!” Ian Gotts offered rich alternative meanings for AI:
AI = Augmented Intelligence
AI = Assisted Intelligence
AI = Actionable Intelligence
Like PCs, the web, smartphones, and cloud computing, AI will enhance how we work and live, especially for business analysts.
Ian went so far as to declare, “Business analysis is THE superpower.” Generative AI, like GPT, can assist business analysts in many ways, such as:
Organization and domain research
Stakeholder surveys and interviews
Translating business needs into solution requirements
Writing user stories and acceptance criteria
Augmented intelligence will relieve business analysts from collecting and transcribing content, enabling them to focus on creating high-quality requirements, user stories, and acceptance criteria.
Delegate to AI, then Verify
Ian Gotts recalled a vital management lesson that especially applies to assisted intelligence:
Delegate, don’t abdicate
ChatGPT has generated a lot of excitement, as well as content, leading some to believe they can hand off any content creation task to it, then pass generated content on to the next stage of the process without checking it.
Generative AI is in its infancy and can make mistakes. While it can produce vast amounts of content from a few prompts, the quality may not match the quantity.
The late U.S. President Ronald Reagan once said, “Trust, but verify.” When it comes to AI-generated content, that becomes:
Don’t trust, always verify
Generative AI can produce a lot of eloquent - and inaccurate - content. Business analysts should check all generated content to ensure it expresses the intended message. They should thoroughly test any interviews or surveys where generative AI converses with stakeholders.
AI = Assured Intelligence
Most GPT Dreamin’ sessions emphasized security requirements when using GPT in any form. No one wants to leak their company’s proprietary information into GPT or worse, train it with the information.
Imagine using GPT to write user stories for a process that gives an organization a competitive advantage. If GPT learned that process, it could unintentionally share it with a competitor.
Anyone using generative AI, such as GPT, should understand what it retains from prompts and responses, who has access to that content, and what they can do with it. Michael Leach’s session “Building GPT Apps on Salesforce” included Microsoft Azure’s Open AI security pledge:
Your prompts (inputs) and completions (outputs), your embeddings, and your training data:
are NOT available to other customers.
are NOT available to OpenAI.
are NOT used to improve OpenAI models.
are NOT used to improve any Microsoft or 3rd party products or services.
are NOT used for automatically improving Azure OpenAI models for your use in your resource. (The models are stateless unless you explicitly fine-tune models with your training data.)
Your fine-tuned Azure OpenAI models are available exclusively for your use.
Michael strongly recommended reading the documents in the OpenAI Trust Portal, covering OpenAI’s “commitment to data security, privacy, and compliance.”
Salesforce takes generative AI security a step further with its EinsteinGPT Trust Layer. It prevents customer data from going to an external generative AI service, checks responses for toxic content, and filters it out.
While generative AI provides tremendous leverage to research and content creation, all stakeholders need assurance that confidential content and data stays safe.
The Long and Thrilling Road Ahead
The 2023 GPT Dreamin’ Conference spotlighted the transformative role of generative AI in the Salesforce landscape. The keynote session provided perspective, reflecting on the PC, web, and cloud computing revolutions. Thinking back on how far those technologies have come, we can see we are only at the beginning of the generative AI revolution and have a long, exciting journey ahead.
A couple of themes emerged from GPT Dreamin’. First, generative AI will become a valuable assistant for business analysts (BAs), streamlining processes from stakeholder interviews to creating requirements and user stories.
However, BAs should exercise caution with AI-generated content, verifying all its responses. As GPT's capabilities continue to grow, so will concerns about security and privacy. Salesforce's EinsteinGPT Trust Layer and Microsoft Azure's OpenAI security pledge assure that an organization’s information will stay safe in generative AI interactions.
Generative AI, with its transformative potential for business analysis in the Salesforce ecosystem, should balance enthusiastic adoption with vigilant verification to ensure data security and accurate content generation.
Learn more about GPT Dreamin’ here.