Author: Jeff Young
Director, Client Success
Spring has finally sprung! It's the perfect time to get outside, soak up the sunshine, and tackle some much-needed yard work. After a long winter, I eagerly dusted off my trusty lawn mower, quickly realizing that it needed an oil change. While I could probably ignore it for another season, I will pay for it dearly when my lawn mower is out of commission, and my lawn is overgrown. As with any oil change on a small engine, the requirements are simple – oil, an oil filter, and a little manual work.
I jumped on my computer and searched for the parts I needed. I quickly discovered the oil filter I need is sold out everywhere locally and would be a week if I ordered it online. My Sunday morning routine quickly went from cutting the lawn while drinking beer soda to vigorously searching online for a fifteen-dollar oil filter that does not seem to exist anywhere on planet Earth. It then dawned on me… Why not dispatch a GPT agent to search the web for me so I can return to spending time outside? In about ten minutes, I had the GPT agent deployed, goals for the AI engine established, and I could get back to what I was doing, including not staring at a computer screen when it was 75 and sunny outside.
After thirty minutes outside doing some string trimming, I returned to my computer to find a nicely packaged report. The report included links to websites selling the required oil filter and phone numbers for a few local dealers. Unfortunately, the local dealers do not have a strong e-commerce presence with a detailed inventory. Still, the GPT agent was smart enough to call them out simply because they are partners with the specific oil filter brand. Sure enough, the second local dealer I called had one in stock! As it turns out, this is a place I drive past regularly, but I would have never guessed they sell lawnmower parts. So before long, the oil change was done, and my lawncare routine was complete.
Ultimately the GPT agent achieved three things for me:
- True AI Delegation – I dispatched a GPT agent to research on my behalf while spending time on higher-value work. This delegation helped me prioritize more important things while the GPT agent handled the lower-value task of searching the internet.
- Wait Time Savings – I could have ordered the part online, but it would have taken up to a week. My wait time was cut down to minutes to what could have been multiple days.
- Faster Delivery of Value – It is doubtful I would have called the place that ended up having the part I needed since I would have never thought to look. However, because the AI scrapped their website and identified a match to the specific brand, it brought it to my attention. This resulted in me getting the job done much faster than initially anticipated.
Something our daily lives and the business world have in common is trying to drive value in the most efficient way possible. A common practice used across multiple industry verticals is the five lean principles. While lean principles originated from the manufacturing industry, these principles can be applied across any business or process across all industries.
The lean principles are:
- Value – What is it the customer is looking for, and what are the requirements
- Value Stream – What steps need to be taken to deliver on the requirements and the value
- Flow – Ensuring the process is smooth end-to-end with no interruptions, delays, or bottlenecks
- Pull – Managing time to market (flow) against demand from customers (pull)
- Continuous Improvement – Continued iteration of processes to increase overall flow efficiency
If we map out the oil filter situation "flow" in a simple value stream map before augmenting it with a GPT agent, it looks something like this:
Because all of the steps in the process must be carried out in serial, this increases the work time and wait time.
- Work Time: 3 hours and 20 minutes
- Wait Time: 5 days and 4 minutes
After we have augmented this process with the GPT agent, the value stream map looks like this:
By augmenting the flow with the GPT agent, the work time and wait time results are the following.
- Work Time: 2 hours and 55 minutes (25 minute time savings)
- Wait Time: 12 minutes (4 days, 23 hours, and 52 minute wait time savings)
While we added a process to the overall flow, we saved work time and significantly reduced our wait time. The other significant change was performing the "String Trim" process earlier in the overall flow. This ensured wait time could be as efficient as possible.
The lesson here, from a business perspective, is:
- Understand and Evaluate Flow – If you do not adequately document and map your current flow, augmenting the process with AI could increase work and wait time. AI might sound like the answer to flow efficiency, but only when the flow I well understood.
- Consider Adjusting Flow to Accommodate AI – You might need to adjust your current flow to allow an AI model time to work. You must be mindful of where an AI solution is inserted in the flow so humans are not in a "waiting" pattern for the AI to do its job. This is where the parallel execution of processes is of the most benefit.
- Additional Processes Might Be Okay – We want to strive for fewer processes in a flow, but we need to be mindful of the ultimate goal – reduce work time, wait time, and potential errors/bottlenecks. A GPT Agent is a low-cost resource that runs at machine speeds. When considering the first two points in this list, adding an extra "process" to the flow is acceptable since we can deliver value more efficiently.
From value stream mapping to augmenting your existing processes with AI, Long View is here to help! Please reach out if you want to continue the conversation about how AI can accelerate your business (or if you need help finding an oil filter for your lawnmower.)