Earlier this year, an MIT study found that business professionals who used generative AI in their work saw their productivity improve by nearly 60 percent. But that’s not all: According to blind graders, the final product was also of a higher quality when generative AI was involved.
These benefits explain why so many companies across industries—from banking to healthcare to marketing—are looking to harness the power of generative AI. But what does this look like in practice? How can businesses integrate generative AI successfully into existing workflows?
Skilled prompt engineering is the key to getting the most out of generative AI. In Part 1 of our series above, we discussed how to create tailored, specific prompts that yield useful outputs, as well as what to do when things go sideways.
Now you’re ready to do more. Here’s how to effectively incorporate generative AI into your workflow.
Whether you’re developing a research report, crafting a technical presentation, writing an article, or creating a computer program, effectively incorporating generative AI as a tool begins with breaking down each step in the workflow into its constituent parts—just like a prompt.
The greater the specificity, the better (especially for complex tasks).
Suppose a company is trying to write a business proposal for a prospective client. The steps might look something like this:
Now for each step, consider whether a technology tool could help accomplish the task; then ask which tool might be the best fit and how to augment your workflow accordingly. Integrating AI into this workflow might look something like this:
1. Research the client, including its business model, competitors, and relevant markets.
Use generative AI to help research competitors and market trends.
2. Meet with the prospective client to understand its needs.
Input the meeting notes into generative AI to summarize the prospective client’s needs and next steps (make sure to exclude anything confidential).
3. Confer with the team about how the company’s services align with the prospective client’s needs
4. Draft a proposal.
Use generative AI to create an outline or first draft of the proposal, drawing on the inputs gathered in previous steps.
5. Create a scope of work based on client feedback. Prompt a generative AI tool with the proposal and prospective client feedback to create a scope of work.
6. Sign a contract and get to work
This technology is progressing rapidly, with each new iteration bringing new opportunities. And despite the “quick fix” promise of these tools, successfully integrating them into your workflow will take time and experimentation. But the more rigor we apply to the use of these platforms, the more transparent, repeatable, and productive we can ultimately be.
For now, it might be useful to start thinking about a prompt engineer (and the generative AI they prompt) as just another member of your team. In your product development and management teams, you might have a prompt engineer working alongside QA testers, legal compliance, project managers, developers, and others.
Think of these new tools and team members as a part of your creative studio: If you were launching a new book in the past, you’d need to write it, edit it, design it, and market it; with generative AI, you might be able to expedite a first pass on some of these tasks, be it designing a cover using DALL-E or brainstorming character names based on geography and historical context for the book using ChatGPT.
At the same time, start laying the groundwork for the future. While most of the attention right now is on text-based solutions, video, audio, and other capabilities will likely accelerate next. Whatever the case may be, we’ll be here to help.
This is the second article in a three-part series on unlocking the power of generative AI.