Definitely the hottest trend right now is generative AI. I already covered my thoughts on Chat-GPT and the future of AI in this post.

Today I’d like to do a short (I hope) and light (I hope) post about generative AI. I’m gonna explain what it is, how you should leverage it as product managers and at the end – I’m gonna pick up some known tech companies and suggest how I’d integrate generative AI into their products.

Ready? Let’s roll.

What is generative AI

From Chat-GPT:

“

Generative AI refers to a type of artificial intelligence that can create new, original content rather than simply responding to input data. Generative AI models are trained on large datasets and learn to generate new data that is similar in style, format, or content to the input data. The generated data can take many forms, such as text, images, music, or even entire videos.

“

Cool. So when talking about ‘generative AI’ we refer to a subset of problems from the AI world which focused on producing new content that wasn’t there before.

We’re talking about the creative aspects of AI.

Chat-GPT, DALL-E, AIVA, ArtBreeder, Midjourney and their similars are all great examples of products which are being driven by generative AI.

Generative AI and product management

We are now in the point in time where generative AI solutions and APIs are becoming a commodity. It should change your state of mind when it comes to some aspects of your work:

Advanced AI capabilities are no longer reserved to companies with huge budgets. It’s no longer a ‘rich-only’ game.

There is a good chance that the generative AI solution you are looking for is just an API away!

And for a reasonable fee – your product can access it and leverage it.

What it means to you as a product:

1. You need to start to think big and not limited by your current data science resources
2. Depending on your product – you may not need a data science team at all
3. If generative AI is relevant to your product – then you MUST leverage it if you wish to stay competitive.

On the other hand, like any other trend, there is a true risk of abusing such a technology just for the sake of ‘being cool’. I do see around me product managers who brag how they integrated generative AI into their products… but when looking deep into this – those integrations are just serving minor features that do not move the needle anywhere.

Look, if you were instructed by the management or the marketing team to integrate generative AI just so your CEO could brag that you are using such a technology – then sure – find the most convenient feature that can leverage it, devote to it one sprint and be done with it. Now everyone in your company can brag that you are part of the game too.

But if you wish to truly make an impact here is where you need to be careful:

1. Don’t use a third party AI solution if this is your core business. For example – if your company sells a CMS (content management system) to digital publishers with the promise of auto-generating blog posts – then you shouldn’t rely on a third party for auto-generating your blog posts. That’s true for reliance on any third party technology and not necessarily for AI. If that’s your main selling point – you need to do it yourself (in-house), because otherwise the third party can easily take you out of business by changing terms, pricing or blocking some of their APIs.
2. Don’t use it for features that do not move the needle, even if it makes sense feature-wise. As reasonable as the pricing may be – it’s still costing money. Why waste money when it doesn’t move the needle for your customers/users?

If you adhere to my advice above – it means that you keep your data science team focused on your core business and claim-to-fame, while leveraging generative AI to shorten time to market and providing additional value using side projects/features.

It may also make sense to leverage such technologies for building quick proof of concepts in case you are unsure that you’d like to put your resources to build some features in house.

How I’d leverage generative AI in known products

Spotify

Spotify is currently not in the business of generating music, but rather in the business of music discovery and streaming. But what’s next for this company?

This product has reached the stage where it didn’t introduce anything new in the last few years. I think it’s time to take it to the next level.

What if Spotify could help you generate your own music and offer it to others under the same business terms?

It means that as a user – if you are into composing music – you will be able to leverage Spotify as a business platform for making yourself a name and money.

For that Spotify needs to provide tools and a framework for composing music. There are already generative AI solutions that can create music tracks for you. I didn’t fall off my chair by listening to these produced tracks, but if those are served as a base and you could add the ‘genius human touch’ to those – you can make some of these tracks a real hit.

So – if Spotify wishes to explore this venue, I’d first connect to such an API and build a quick POC and try it with some users. If this is getting traction and seems to be a promising direction – then I’d either build it in house or buy the company which provides such generative AI solutions.

AirBnb

What a marketplace for short term lease can do with generative AI?

On the face of it, not much.

But again – AirBnb hasn’t made any significant change in the recent years, and it might be the time for opening a new market:

Vacation planning.

Vacation planning is still a pain in the a#$s in 2023. For a new startup to tackle this – it’s a formidable challenge because most of the relevant data you need is behind walled gardens (TripAdvisor, AirBnb and Booking.com).

However, if you are already a walled garden with plenty of assets – why not leverage this and help users plan their dream vacation?

Imagine that when users are looking to rent an AirBnb you show them a short questionnaire about what they are looking for (and not just a bunch of plain filters). The questions in this questionnaire should be aimed to understand what a ‘dream vacation’ should look like for this person/family.

Based on the answers – a generative AI solution can offer such a vacation that may span over multiple AirBnb assets and geographic regions.

Of course – AirBnb will need to collect more data than it has today about the various sites in each region, but that is possible by analyzing the assets’ pages and their reviews using NLP (with named entity extraction).

Anyway – for a company like AirBnb that is fully achievable in my humble opinion and will provide a great value to the users of the platform.

Course Hero

I wasn’t aware of this company before. I stumbled upon it when reviewing a list of unicorn companies. Turns out this company is worth about $4B and very successful.

As far as I understand they offer study materials to students of colleges and universities. Some of these materials are provided by experts but quite a lot of materials are crowd sourced by other students as well.

This company is a classic example of a company that can benefit from generative AI. Given all the study materials they have access to – a good generative AI technology can generate summaries from the most popular resources.

Instead of a student having to go through piles of materials – they can get a clean summary of 4-5 pages fully designated to the specific test at the specific school they are in.

I mean – it’d be nuts if they didn’t kick this project already.

Shopify

This platform for eCommerce can greatly benefit from generative AI as well.

They can auto-generate store pages and also product pages for their customers. This is classic.

A more advanced AI approach would be to auto-generate a short marketing video for each store. Such a service would be of great value to their customers as well.

Any platform which offers web pages (such as store pages on Shopify) can also greatly benefit from a service which offers SEO optimizations.

Now, there are plenty of plugins that do that for CMS platforms (such as WordPress) and for Shopify as well. However, those SEO optimizations tools are following standard rules (setting up a meta-tag for the title with the value of the post title for example).

There is an opportunity here for a generative AI tool which takes SEO to the next level. If you don’t want your shop to look & sound like a big clickbait site – you don’t have to. The AI will help you preserve a dignified look to your site/shop while applying all the SEO techniques behind the scenes only on the metadata while keeping your page ‘clean’ to the human eye.

Summary

Generative AI is becoming a powerful commodity. I believe it’s gonna create an inflection point in the markets for quite a few industries.

You need to start thinking in these terms if your company deals, even remotely, with content. Because if you don’t – the competition will do it for you.

That being said – it’s a tool, and each tool has the right time & place to use it. Not everything requires generative AI, so don’t force it on things that don’t make sense, or won’t make a true positive impact.

That’s it for today.

If you found this post/series useful – please to let me know in the comments. If you think others can benefit from it – feel free to share it with them.

Thank you, and until next time 🙂