Hey! Charlie here 👋

This last month I’ve surveyed over 50 products that use AI and compiled a list of the best strategies I’ve seen put to use that aren’t yet another (😩) chatbot.

Let’s get straight to it… ⬇️

Join 400+ tech peeps learning to build SaaS products every week for free on IdeaHub.

1. Augmenting data

A common use-case of generative AI is to simply take existing data, and make more of it, optimising for a goal.

TweetHunter uses this strategy to generate new tweets from its customers’ account history, based on what’s performing best.

But in practice, this approach would work well in any SaaS product where the customer’s goal is to optimise content.

2. Intelligent search

If your product manages text, image, audio or video content, then adding a generative search could help users find what they’re looking for quickly.

Nuclia is offering an as-a-service solution for AI search across multiple mediums where you can upload your own data, and use their SDKs to query it.

3. Complex form completion

Any SaaS product that requires users to use a “long enough to be a bit annoying” form could benefit from an AI integration to take some of the pain away.

Working on risk management SaaS myself, I’m particularly excited about the idea of having AI fill-in insurance forms for accidents in the workplace - a huge time saver!

4. Summarising data

We’ve all seen the hundreds of SaaS products endlessly offering to summarise our meeting notes.

But there’s actually so much more value in this concept if applied to your niche correctly.

Take AppRadar, which uses AI to summarise reviews for mobile apps.

It allows its users to keep tabs on competitors without trawling through reviews themselves.

Often the value gained from this kind of AI solution is in time saved, so think about where this might be of most value to your customers.

5. Predictive generation

The use case for AI in predictive generation is literally endless, but here are 2 of the best practical examples I’ve found that you could take inspiration from.

Sentry, an application performance monitoring tool, uses AI to suggest solutions to errors that are detected in its customers’ applications.

SocraticWorks uses predictive modelling to forecast how long projects might take to complete based on their massive data set of completed tasks.

6. Data visualisation

If you’ve ever added an analytics dashboard to a SaaS product, you know that defining exactly what customers want to see can be challenging.

Cumul.io are building an ‘insights miner’ based on GPT-3 that allows you to simply generate a dashboard from an existing dataset.

In a nutshell, you could expose data to a pre-trained model and have it generate charts and metrics that are optimised specifically for each customer.

7. AI recommendations

The ability of AI to suggest a user’s next action is already prominent on social media platforms, video streaming applications and E-commerce websites.

But since the explosion of AI, this wave has started to hit many more markets.

Recombee is a big part of this with its AI recommendations-as-a-service engine that is already serving multiple industries like gaming, music, travel, marketplaces etc.

8. Roll your own!

If you have an idea for an innovative AI integration for your product, then there’s no reason you can’t build this yourself.

Plus, you don’t need any advanced AI knowledge, just a prompt and an integration with an LLM.

Here’s the simplest way to achieve this:

• Design a prompt (guide)
• Enrich the prompt using your customers’ specific data
• Hook up your SaaS to ChatGPT using Pipedream or Zapier
• Interpret the output in your product

That’s it for this week! 👋

Join 400+ tech peeps learning to build SaaS products every week for free on IdeaHub.