How to Make the Case for Artificial Intelligence
“ChatGPT turned the work of days into the work of minutes.”
This wasn’t a marketing pitch; it was a gushing review from a senior leader in an organization I’ve worked with. She was describing the impact the newly released ChatGPT had on her understaffed team. With early rave reviews like that I continue to be surprised by how many organizations have slow-walked their rollouts of LLM tools. As time goes by, I’ve also begun to hear arguments that ChatGPT, CoPilot, and other large language models haven't delivered on their value proposition. The price tag, hallucinations, content ownership, budgets, and other challenges coupled with frustration from some early adopters seem to have contributed to delays. Most often, I hear people repeat that they need to make the business case.?
Who needs a business case with results like that??
Admittedly, as professional writer I don’t get a lot of use out of the copy generation capabilities. However, every time CoPilot summarizes a meeting I was double booked for makes?it?worth the $30 (Canadian) a month it costs on top of the base O365 license. Finding files quickly has been another winner for me, as well as ideation support when the creative juices aren’t flowing or the coffee hasn’t kicked in yet. That said, if it was easy to do a Net Present Value calculation on the value of having an assistant who’s available at the drop of a hat, we’d all have an admin to back us up. ?
Business Cases are easy... for the big stuff.?
I’ve written a few business cases for technology in my time. They are fairly straightforward when you are tackling solutions that have a repeatable measurable effect on a quantifiable set of processes with a large number of users. As the BABOK says, “A justification for a course of action based on the benefits to be realized by using the proposed solution, as compared to the cost, effort, and other considerations to acquire and live with that solution.”?Where the traditional approach breaks down is when you have to nail down the quantifiable benefit to individuals using the tool to solve wildly different set of problems. ?As a result, I’ve seen specialized AI solutions appearing in application portfolios before LLMs. This is true even when there are significant concerns, as I discussed in my article about the brouhaha over the implementation of facial recognition software by two Ontario police services.?
So how do you do make the case for LLMs??
领英推荐
The first step is realizing that you realize that the value proposition is fundamentally different: ?
Buying a LLM license for a knowledge worker is like buying a carpenter a table saw. The handsaw is way cheaper, and who doesn’t love bespoke wood shavings? On the other hand, they can cut a lot more boards feet in a day if you give them the right tool for the job. While LLMs might look like a large-scale platform license because they sit on top of your existing productivity software. ?
Make the case for?personal productivity.?
Tiff Macklem, the head of the Bank of Canada, has called productivity Canada’s Achilles Heel. If you’re struggling to make the case for an LLM in your organization, think about starting with a case of one. Maybe even just yourself, and the personal use cases it can help you power through.? Work out from there to different types of roles and personas and you’ll quickly find the case will start to make itself.?