Capturing the efficiency gains of LLM’s starts with and is centered on your ability to build and implement prompt based workflows…
Simply put, you need to build up your experience around using individual prompts as building blocks to construct the larger effects which you then orchestrate into complete workflows to produce high quality outputs. You do not need private models, specially fine tuned models, or super secure 3rd party services to do any of this type of development for your organization. You could literally have an unpaid college intern do this for you over a few weeks or months with nothing more than a $20/month ChatGPT or Claude account. But what does this development look like?
The point is to find the prompts, test the effects of combination, and evaluate the combined workflows. It starts with dissecting your target business process into its fundamental action steps. Building out, testing, and refining individual prompts around each of these steps. Combining these prompts together into ever more complex structures, measuring effectiveness, and revising along the way. Ultimately, the task is complete when a minimum quality threshold is met by a sufficient output from a total workflow of all the afore-mentioned pieces. Remember this is all done in just plain English, so data analysis backgrounds help, but basic grammar fluency and many reps with the tools is the most beneficial. Hence a person more apt to wilding enjoying crossword puzzles would be a better choice than a data scientist for this task.
Remember as unique as your business’s products or services may truly be, all your normal internal business processes are not. These boring but necessary processes are the perfect kind of tasks to start with, with many examples and easily understandable components for anyone to try automating with LLMs. They also carry the benefit of wide horizontal integration in your business when you achieve them. The point being to start somewhere, do the development as cheaply and efficiently as possible, then scale up and integrate your successes. This is the point where you might want to look at all those other model and tool options if they are applicable to additional scaling of efficiency gains to your successful workflows.
This is what GenAI integration looks like for your business in the most cost efficient and business advantaged way. Go buy off the shelf products if you want, but learning and implementing your own solutions is the key to long term success. If you need help doing or setting up this kind of work, I also offer training and consulting services.
