In this blog post Mike Read, Product Manager here at The CAI Company, shares his experiences of AI as it starts to become part of every day of working life.
AI at work
The CAI Company has a team of highly skilled Conversational AI specialists. But I am a bit of an outlier in that my area of expertise is product management.
I thought I’d share a little bit of my journey as I get up to speed (rapidly hopefully) on the fascinating topic of AI (artificial intelligence) – something all of us in the Tech industry will need to be doing in the coming months. AI isn’t a topic for one team to investigate in isolation: it will touch every part of our business in the future.
To start with I’ll stick with the more general use of GenAI (generative artificial intelligence) within the day-to-day work that we all do, regardless of our specific fields of expertise or job roles.
These are three examples of my recent experience with GenAI in action in my work day.
Let’s begin with the good…
A few weeks ago, I volunteered to do a workshop on Agile concepts and SCRUM methodology for our broader team. This morphed into a two-and-a-half-hour session on the morning after our Christmas party. Not ideal but I had plenty of time to prepare, so all good.
I’d planned in a couple of days to pull together the materials, create the slides and do some general preparation. I’d even had the idea (I’m definitely not the first person to have this idea…) of running a little SCRUM Bingo. However, as time drew closer certain priorities intervened, in the form of some Chatpulse demos and workshops on future Product direction. All somewhat higher priority than a bit of SCRUM Bingo.
Before I consigned myself to a couple of very late nights and a trawl of the web for my required materials, I decided to give ChatGPT the task of pulling what I needed together. Within half an hour I had all the materials I required. With a further 10 mins over the next couple of days I even managed to generate some pictures to help represent the 4 core values of Agile. The pictures were a bit dubious, but more on that later. Additionally, ChatGPT produced some nice SCRUM Bingo cards…although with the small flaw that they were all identical!
Now there is one big caveat to this experience: it’s a topic that is firmly in my wheelhouse. I know Agile and SCRUM pretty well; my requirement was a quick way to create a succinct summary of all the key points. And there was a definite need to curate the content I was provided with. For example, the initial overview of the core values were not in the format, “while we value <concept x>, we value <concept y> more”. This is an absolutely critical element of the 4 concepts of Agile. However, with one additional prompt ChatGPT came back with the correct format.
So, all in all, it was very beneficial exercise and saved me hours and hours of effort, I reckon easily an 80% saving in my time.
Now on to the distinctly bad…
Also, in the very recent past, we instigated some conversations with a cloud-based marketing solution to support us as we look to further roll out our Chatpulse product, and build some market traction.
We asked for a demo and shared our requirements for CRM (customer relationship management), marketing automation and service desk management; we made it very clear that we initially need base level tooling, which is what we wanted to see, with the ability to scale in due course, and that our team are specialists in Conversational AI. Essentially a demo to highlight core features in the entry level product offerings.
Unfortunately, what we got was a demo of the newly added AI functionality that could generate content for us (not available with the entry level product). Even more unfortunately the results shown were completely unrelated to what we do. The demo produced very dubious content and complex flows we are unlikely to use in the immediate future.
Why is it that so often the most important questions are not asked when it comes to AI? Are features really required and will they actually deliver value? Do they provide the level of quality required?
We decided to explore other products.
And the plain ugly…
OK, this was possibly an unfair request but referring back to the Agile workshop, I asked ChatGPT to create two images, specifically presenting teams that represented the key agile concepts e.g. a team swamped by documentation, compared with a team successfully delivering a Product.
The results were mixed and looked like something out of the Dark Tower by Steven King. Perhaps most worrying of all is that according to ChatGPT I need to grow a beard in order to be in a successful team focused on Software delivery!
However, we did have great fun as a team figuring out what each image was trying to represent.
Generated using AI
So, my conclusion from these experiences of GenAI in action in my work day
- Firstly, if you haven’t started using GenAI to accelerate generation of content I urge you to try it. It could save you a great deal of time. But do make sure you validate the content before you use it.
- Secondly, if you’re buying a product, or you’re starting to look at how best you can utilise AI in your own product, service or business, please start by asking yourself the key question: Why do you need this and what is it going to do for your Business? GenAI is clearly going to be a transformative component of businesses in the future (starting now!) but don’t start off by building/buying something that you don’t need and won’t deliver any real value.
- And finally from what I’ve seen, I’m not sure my creative friends and colleagues have anything to worry about just yet, GenAI can provide some truly spectacular results often in a good way, but also very much in a bad way.
