Everyone Will be a Business Analyst
Everyone Will Have Einstein
At the Dreamforce 2023 opening keynote, Salesforce CEO Marc Benioff introduced the Einstein 1 Platform, a generative artificial intelligence foundation for Salesforce products. The keynote included the tagline, “Now everyone is an Einstein – Data + AI + CRM + Trust.” It really means everyone will have an Einstein as Salesforce builds AI into its services.
Dramatically More Capable Solutions
Shortly after his Dreamforce keynote, Benioff interviewed Sam Altman, Founder and CEO of OpenAI, the leading developer of generative pre-trained transformers (GPT). Altman predicts that GPT models will become “dramatically more capable” using software development as an example. He said that code-generating AI models will progress from writing 25% of a developer’s code up to 95%. Altman speculates this will lead to a qualitative shift:
If you can operate a higher level of abstraction [and] you can keep more of the big picture problem in your mind at one point, you can just do dramatically more.
What does “dramatically more” mean? It means that AI-generated code will widen the development bottleneck, enabling more ambitious projects and strategic solutions.
AI as a Developer
Code-generating AI will need business analysis deliverables to produce high-quality code:
The data model
Process specifications (from a process map)
Use cases
The glossary of terms
Training a code-generating AI model on this content will take far more than prompt engineering. Business analysts and architects will need to train the model with their deliverables. Training AI models will eventually become easy enough to do that. Ideally, expert users will specify requirements that a well-trained code generator can develop into a prototype solution.
Patterns of Innovation
If this sounds fanciful, there’s a historical analogy from 40 years ago. Personal computers were a novelty in the early 1980s until spreadsheet software made PCs useful to business. Technically inclined users turned into “spreadsheet jockeys” until spreadsheets became easier to use. Generative AI is at this early stage, requiring machine learning experts to train AI models. Similar to spreadsheet jockeys morphing into power users, technically inclined professionals like business analysts and architects will lead the way in training AI to generate code from specifications.
Everyone Will Need Business Analyst Talent
Code-generating AI could evolve to build prototype solutions from requirements provided by AI-inclined domain experts. These experts will need business analysis talent to develop the requirements. We could reach a point where everyone needs business analyst talent, similar to Salesforce’s claim that “everyone is an Einstein.” This means everyone will have access to powerful AI-driven tools, enabling them to create dramatically capable strategic solutions.
Once AI generates code, more people will need business analyst talent to specify what it will develop.