What problem should you use AI to solve with product manager Diana Bui

Product and marketing are closely intertwined. Those who are familiar with marketing theory know that product is one of the 4Ps, with marketing traditionally working to develop products that have a market-fit.

But if you're a product manager or a marketing manager working with product, where does AI come in? And how does it help, whether you're building digital features or a new SKU?

It's why we asked experienced product manager Diana Bui to share her approach to AI, and why she recommends slowing down rather than moving fast and breaking things.

In this episode, Diana speaks to Gary and Christina about:

• Where to get started with AI when you've got a number of different directions you could go in.
• How to harness Copilot properly.
• Thinking user-first when using AI.
• Why you should never ask AI to do something you don't know how to do yourself.

Soc•AI•l Skylines is a Social Hills production.

Hosts: Christina Vetta and Gary Andrews.

Producer: Peter Northcote.

Videographer: Imogen Mabin.

Guest: Diana Bui.

Transcript

SPEAKER_02 0:04

Hi, and welcome to this edition of Social Skylines, a production by Social Hills. I'm Christina Feller.

SPEAKER_00 0:10

And I'm Gary Andrews, and each episode we'll be cutting through the noise and the confusion around AI and getting practical insight from working professionals who are touching these tools day in, day out. And this episode is focused on one key question. What problem is it that do you actually want AI to solve? Now in every organization, there's a lot of problems AI could solve, but building the capabilities to solve all of them is not just a challenge, it's probably quite impractical. So on this episode, we're delighted to be joined by Diana Bui, Senior Product Manager, and you spend a lot of time trying to answer that exact question. Thank you for joining us.

SPEAKER_02 0:48

Thanks for joining. Thanks, Gary and Christina.

SPEAKER_00 0:51

So AIs, LLMs, they're often pitched as a silver bullet for an organization. A lot of businesses go, great, let's speed up adoption. We want to do this really, really quickly. Desire is not the same as implementation. Fairly obvious question, and I think this is one that a lot of people will have experienced. Okay, so where should we use AI? Where on earth do you start?

SPEAKER_01 1:15

It's such an interesting question because imagine I asked you, oh my god, I've just been gifted a hammer. Where should I use my hammer? It's like, do you have any nails to hit? Otherwise, it's just a hammer. And I think AI in a very similar way is a tool that is to be used to solve particular problems. So if you have those problems, you can use them. Otherwise, it's just another thing on the internet, right?

SPEAKER_02 1:40

So, how do you hone in on exactly what problem you are looking to solve with AI?

SPEAKER_01 1:46

That's a great question, Christina. I think much like the hammer analogy, if I didn't know what a hammer does, I wouldn't be able to use it. So the first step is actually understanding what is AI and what can it do for me. I think one of the first most important things for the average person to understand about AI is it's not this magical thing that can do everything for you. It's just a very powerful brain, if you think of it like that, that knows a lot of things, but it's not very good at applying the things that it knows. So for the average person that's interacting with your Chat GPT, your co-pilot, or whatever LLM you've got access to, it's kind of like talking to a very smart but also in some ways very dumb person. Trying to guide them to solve the problems that you have. So if you can understand that aspect about AI, you'll be able to then think, okay, what problems can I apply to this tool here that knows a lot but needs a bit of guidance from me?

SPEAKER_00 2:41

There's a lot of people who kind of understand it, they've got this big picture thinking. We can do this with AI. We know the tool can do this, that, or the other. But whether in a big or small organization, government, private, even if it's just you're a smaller business, where do you start to get into those details and how deep should you try and get into the details that matter beyond that big picture?

SPEAKER_01 3:03

When it comes to details, the most important detail for you is understanding what problems do you have as a person in an organization. So I'll use myself as an example. So I'm a product manager. I always like to illustrate the difference between a product and a project manager. Project managers just organize, you know, making sure that a project runs to course, runs to time. A product manager is someone who owns a product end-to-end its success, the way it's built, and um understands the outcomes it's trying to achieve, and also owns the relationship with the customers. And so in my job, I try to look at where do I find tasks most laborious? What things do I kind of hate doing day to day? At first it was really hard for me to answer that question because a bit of a control freak. Don't like to give up things to a machine. And so I actually I actually sat down with co-pilot. And I sat down and I actually asked it. I said, hey co-pilot, what can you do for me? I actually don't know much about you. And it answered. It's like, well, I can create slide decks for you if you give me some context, I can help you summarize things if you like. And so it's kind of like talking to a person. I want people to use LLMs in that way, by the way. It's the first tool where you can interact with it like it's a human and ask it questions. So when I was interacting with it, what I found was, hey, I actually really don't like sitting and writing things. And so I was like, can you help me with that? And it's like, sure. Give me some context on the problem you're trying to solve in your squad and I'll break them down for you. And it it has so much knowledge about technology and test cases and good practices when it comes to engineering. It does that heavy lifting for me, and all I have to do is just give it the context and the problem I'm trying to solve. Nothing really replaces human interaction, especially when it comes to customer interviews. But one area that I have um found my team uh has used AI for is synthesizing lots of information. So for example, we might in a day interview five different customers, we put all the notes down on a document. It's a bit unstructured, but it's all there in one place. The hard part for you as a human now is to go away and actually look at it, synthesize it, and turn it into something meaningful. AI can actually really help you with that. So you can chuck that document into Copilot, into ChatGPT, and say, hey, can you summarize this for me and pull out some key themes? And it'll do that for you in like 20-30 seconds. So, what are some of the guardrails that can be easy to miss? I have a rule. It is a personal rule. I do not ask AI to do anything that I cannot do myself.

SPEAKER_02 5:31

Because if I don't know how to do that task, I can't fact-check it. But that's one of the first times I've heard someone say that in all the interviews that we've done. I love that.

SPEAKER_01 5:41

Like where it is today, AI still is very prone to what they call hallucination, right? Yes. And so we hear that term thrown around a lot. It just makes stuff up. And so now think of it in the context of that very smart human, right? You have this very smart human who kind of struggles to apply what they know, who's then has a propensity to lie to you sometimes.

SPEAKER_02 6:00

Yeah.

SPEAKER_01 6:01

Probably not the best employee or the best colleague to have on your team, but think about how you would manage someone like that. You probably want to check their work, right? Before you push something out, if they've given you something. So, much in the same way you might check the work of a junior colleague. Check the work co-pilot or chat GPT is given back to you.

SPEAKER_00 6:17

One of the things that it sounds like we're coming to a little bit, and has been actually a theme through some of the other interviews, is that a lot of people think, oh cool, I can use AI and it will we can go much quicker. It feels like there's a lot of benefit in slowing down the process. So, you know, what for you would be a benefit? Is it a benefit to slow down the process?

SPEAKER_01 6:40

For me personally, yes. And I go off the fact that I I believe that AI is not going anywhere anytime soon. So therefore I'm not in a rush. It's gonna be there regardless of whether I act quickly or slowly. I'm also not missing out on anything in particular because it's I see it as a productivity tool. If I have eight hours in my day and I can already get whatever I can get done in eight hours, great. I now need to start looking at little use cases where I can improve my productivity so that eight hours is more meaningful incrementally. I don't need to double that eight hours, you know, or the output in that eight hours overnight. But if every day I can find a little use case for it, test and learn with it, and increase, you know, my productivity by 1%, 0.5 of a percent, over time that adds up. And that's the beauty of it, is that you have the time to figure it out. You don't need to, I think in tech we call it like build fast and break things. Like also work smarter, not harder. Yes, in many ways that too. That too. Like I think we we put a lot of pressure on ourselves to figure it all out right away. But I promise you right now, even the people right at the top of some of these biggest companies don't know or don't fully understand all the things AI can do for them. So don't put that pressure on yourself as just a regular person.

SPEAKER_00 7:55

It almost sounds like some of the best ways of trying to solve the little problems rather than the big problems, first off. And if you can understand the little problems, you can work your way up to a bigger problem.

SPEAKER_01 8:07

Exactly. I mean, think about it. Wouldn't you, as the primary person working on a project or a task, you wouldn't you want to be the one to work on the big problems as the human? Give the AI the little things that take up your time but are not very valuable.

SPEAKER_02 8:21

So it frees you up to do the harder tasks, the tasks that need more strategic work.

SPEAKER_01 8:26

Correct. That's your strength as a human. Lean into that and then let the AI do the stuff that frankly isn't worth your time.

SPEAKER_02 8:32

So, what would your advice be to somebody that is looking to implement AI? How can it benefit them if they haven't, you know, used an AI tool to date? How can they kind of dive in and start getting it to help them free up their time to do the bigger, harder tasks themselves?

SPEAKER_01 8:50

Well, my first not so much advice, but maybe sort of helping words is to say don't be afraid of it, but it's okay if you are afraid of it. I think there's a lot of noise at the moment in the media, and I think many people across all professions feel this pressure to get it right away. Don't put that pressure on yourself, you're just human. Like no one understood the internet right away when we had our big dot-com boom. You're not gonna understand AI right away. But just pick something that you might already have for free on your phone, download the free version of Chat GPT or use Gemini, whichever one's available for you, and just slowly start to ask the questions and work out how it might help you. I think the one thing that's that sets AI apart from other technological advancements is that it is a very clever tool that can interact back to you as a human, like it is human. If we didn't have software engineers and people like that, you can't really do much, it's just an empty shell. But this tool on its own is like this kind of blank slated human that can talk back to you and knows all these things. So you can sit there. Soundboard. Yeah, you can soundboard with it if you really wanted to. You have the time and you have basically a tool that understands you or has the capacity to understand you and ask you questions. So sit down with it, don't rush yourself and talk to it and see what it ask it what it can do for you.

SPEAKER_00 10:10

I love that. Everything is a problem, and there's lots of problems to be solved within there. Whether you're a small business or you're a big governmental organization, I guess you can just ask the question well, what problem do you think is most important to solve? Because everybody's got an opinion on the internet and it's very good at searching on there.

SPEAKER_01 10:29

That's right. My job as a product manager is to prioritize problems to solve. Realistically, I can't solve all my problems at the moment. What are the biggest, chunkiest problems I can solve right now to deliver business value quickly? If you already have that, which you should, by the way, regardless of whether or not AI is there, if you already have that mapped out somewhere, start brainstorming how you might apply or integrate AI into some of those processes to make it better. Good advice. Thank you so much for joining us today.

SPEAKER_00 10:56

So thank you so much, Shanna. That was really great insights. Probably gives a lot of people pause for thought as well. Uh, you can find more episodes of Social Skylines. Don't forget to the Conversocial Group's website where we'll collate everything that we've done.

SPEAKER_02 11:08

Thanks everyone for joining us today. We hope that you follow our journey and we look forward to seeing you on our next episode.