Alphabet (GOOGL -0.23%) (GOOG 0.29%) released a terrible earnings report on Oct. 25. And since the company has few good things to say about its near-term prospects in the online ad market, its primary source of revenue, CEO Sundar Pichai chose to talk about the company's best long-term opportunities. 

Here is one opportunity that he thinks is one of its best bets over the long term.

Alphabet built an AI-first company

Google has dabbled in artificial intelligence (AI) since Larry Page and Sergey Brin founded the company in 1998. However, it took until 2011 before hardware architecture was powerful and cheap enough to make widespread use of AI possible -- sparking an explosion of AI use globally. And it did not take long before Google became very active in building its AI capabilities. 

According to RS Components rankings, Alphabet acquired the most AI start-ups between 2009 and 2020, possibly assembling the best collection of AI talent worldwide. Included in the acquisition spree was its 2014 acquisition of DeepMind, a company known for employing some of the best AI researchers in the world.

In 2015, Google started using its homegrown Tensor Processing Unit (TPU) within its data centers to power AI applications for over 100 Google products like Street View. A TPU is a custom-made chipset purpose-built for machine learning. These TPUs ran on TensorFlow, a software platform able to prepare data, build AI models, deploy models, and run models in production.

Consequently, Alphabet became one of the first companies with the technologies and skills needed to complete an AI project by itself, from hardware to software to algorithms to data -- a full-stack AI company.

Alphabet soon publicized its change from a mobile-first company to an AI-first company when current Alphabet CEO Sundar Pichai announced the change at his Google I/O 2017 keynote.

How AI improves Google's products

In its third-quarter 2022 earnings call, Pichai gave several examples of how Google uses AI within its products. In particular, voice and image searches are crucial to the company's plans.

According to several surveys, many consumers prefer to search by voice instead of typing. So the demand is there for a good voice-to-text product. But, unfortunately, many voice assistants still leave much to be desired.

Pichai highlighted two efforts to improve voice chat during the call, Sparrow and Language Model for Dialogue Applications (LaMDA).

Google designed LaMDA to understand and respond to a user in a natural, conversational way that the average chatbot using narrow, predefined scripts cannot accomplish. The goal is to build a virtual assistant that can hold a conversation nearly indistinguishable from a human. You can judge LaMDA's progress by downloading and using AI Test Kitchen, a Google app that contains LaMDA.

Sparrow is a product developed by Google subsidiary DeepMind designed to train conversation AI, helping the AI avoid generating inaccurate or invented information, using discriminatory language, or encouraging unsafe behavior.

In addition to developing voice products using advanced AI techniques, Alphabet is developing Google Lens, a product designed to perform visual searches on images or using a phone camera. For example, Lens can identify inanimate objects, plants, and animals. It can also search the web for assistance in solving schoolwork problems in math, history, chemistry, biology, physics, and more simply by pointing a camera at a question or equation.

Lastly, Lens has a translation feature. It can translate text within an image from over 100 languages and instantly display the translation. So, for instance, Lens can translate text on menus or street signs in 100 milliseconds and uses AI to swap out the original text to display the newly translated text within the image, matching the original style. 

Should you invest?

Some analysts estimate the total addressable market for AI to be well past $1 trillion by 2030. And Alphabet is one of the best-positioned companies in the world to profit from the opportunity.

However, the company's near-term prospects look awful, despite AI's fantastic long-term promise. Third-quarter revenue growth decelerated to 6% from 41% in the year-ago period as Google remains challenged by the poor economy. In addition, its diluted earnings per share dropped 24% year over year to $1.06. Considering that it plans to continue slowing head count growth into 2023, Alphabet appears to be preparing for things to worsen.

Currently, Alphabet sells for a price-to-earnings ratio (P/E) ratio of 17, well below its median P/E ratio of 27.02 over the past 10 years. So, you could make a strong argument that once the downturn in the ad market ends and revenue growth reaccelerates, Alphabet will sell at a much higher valuation. 

If you can withstand short-term downside, you should squirrel away a few shares ahead of the oncoming winter.