Deep: AI Trends and Technologies
Your guide for getting up to speed in AI - quickly. Resources, trends, features and more.
🔒DoP Deep goes deeper into the concepts and ideas that are covered in the Weekly Briefing to help you learn lessons from the experiences of top tech companies. Paid subscribers get ongoing access to new reports every month.
Hi product people 👋 ,
It’s been a little while since we last took a step back and gave you a curated look at how to get up to speed in AI - quickly.
In the months since the last AI memo, we’ve had new models released from all of the major tech companies, new funding rounds for AI startups and a whole series of new features and capabilities released by top tech companies and startups.
Keeping on top of all of this is pretty exhausting.
In this Deep Dive report, we’ll share a collection of insights, trends and resources, designed to bring you up to speed with everything you need to know today.
Coming up, a snapshot of where we’re at including:
AI Monetization strategies - how AI products and features are monetized
Trends, new features and technologies
The rise of Agentic AI and what this could mean for product teams and UX
How one company’s AI feature has boosted revenue by 40%
How AI is powering international growth at companies like Reddit, Spotify and others
Essential free courses and resources for staying up to date with fundamental AI concepts
Model comparisons
Recent reports have suggested that progress across some of the large AI models is starting to flatline. And as a result, AI companies are refocusing their attention away from the models themselves and towards different value propositions built upon them.
This may be the case, but it’s still worth keeping tabs on how each of the major models is performing.
Here’s a snapshot of the latest models and their respective performances.
Chatbot Arena (lmarena.ai) is an open-source platform for evaluating AI through human preferences and feedback, developed by researchers at UC Berkeley SkyLab and LMSYS. With over 1,000,000 user votes, the platform ranks the best LLM and AI chatbots using the Bradley-Terry model to generate live leaderboards.
There are plenty of other comparison models available but for ease, we’ll stick to this one.
The evaluation framework ranks each of the chatbots across criteria including hard prompts, coding, math, creative writing and more. Google Gemini ranks top.
Using the benchmarks from this tool, Google’s Gemini is currently in the lead, in joint first place with OpenAI’s ChatGPT. This may change in the near future but it’s an impressive comeback for Google who were seen as laggards just a few months back.
Open source Llama comes in at 11 on this chart and just misses out on the top 10. Keeping up to date with the scores of each of these models is helpful for folks who are considering building out their own LLM-powered features but factors including API access and cost are also major considerations; some companies will happily trade off one factor against another depending on their objectives.
And speaking of cost, let’s now take a look at how monetization strategies are evolving in the AI age.
AI Monetization Strategies - how AI products and features are monetized
For this analysis, we’ll look at monetization in two distinct ways: first, we’ll explore how chatbots are monetized and second, we’ll look at how companies are monetizing the AI features they’ve built.
The monetization of chatbots
For this, we’ve included the top chatbots including Claude, Gemini, ChatGPT and Perplexity.
Premium, paid-for subscriptions and API access are the two core monetization strategies deployed by mainstream chatbots for now. And the extent to which companies rely on each of these revenues is different for each company.
Here’s a recent piece of analysis which breaks down the revenue share for API vs paid subscriptions for Anthropic and OpenAI:
As it stands, Anthropic derives 85% of its revenues from its API products vs OpenAI which is currently generating over 70% of its revenues from its chatbot premium subscriptions and only 27% from its APIs.
ChatGPT has far greater consumer awareness than Claude but with Claude hiring the ex-CTO of Instagram as its new CPO earlier this year, we could start to see Claude become a more widely known brand.
Premium subscriptions for new models - but for how long?
We’re also starting to see companies like Perplexity experiment with alternative monetization methods.
Recently, alongside its premium subscription product, the company announced that ads would start to appear in search results as a suggested follow up question. And last week, the company unveiled Perplexity Shopping which embeds products directly into search results along with clickable buy-now buttons. But the 1-click purchase feature with free shipping is only available to premium subscribers.
Originally, the main benefit of Perplexity’s Pro offering was access to the latest advanced AI models but as models start to become commoditized, earlier access to more complex models is becoming less powerful as an upsell. In this case, Perplexity is starting to add additional valuable bolt-ons for premium subscribers including 1-click ecommerce purchases.
We could also start to see other companies / chatbots follow suit and start to bundle additional offerings to Pro users.
Monetization of AI features
Given the costs involved in powering AI features, most companies have chosen to charge additional fees for accessing AI features. This is only a small snapshot but it includes companies across B2B and B2C including DuoLingo, Notion, Airtable and others to get a feel for how they’re integrating AI into their products: