Real AI Use Cases
In this memo, we describe the current state of artificial intelligence practical use cases by consumers and the enterprise.
While the idea of artificial intelligence has taken the world by storm since OpenAI launched ChatGPT, and Big Tech is investing heavily in NVIDIA chips, the practical applications of this new AI methodology have left many executives and practitioners scratching their heads.
What do we use AI for?
In years past, the most compelling business use case was automatically closing support tickets that had been re-opened when a customer replied, “Thank you.” This is not lighting the world on fire, or accelerating free cash flow growth.
So, we've been surveying, sitting in, researching, and speaking with sales leaders, business owners, product and technology people, investors, and founders to get a sense of the industry's current state.
In short, LLMs have opened a path to new and valuable use cases that we have had an early glimpse at, and some we've helped architect, patent, create, and commercialize.
Here's an example of a requirement set at a planetary scale that we feel confident will be realized over the coming years:
Planetary-Scale Commercialized Solution
AI Consumer Agents
On the consumer side, we expect “AI Agents” to do work on your behalf. For example, your Amazon AI agent may search the web or negotiate the best price for your regular grocery and other home goods purchases as part of your Prime subscription—better prices without you spending all the time.
Personalized Email Lead Gen
On the business side, the big idea right now is personalized email lead gen for sales teams, including automated sequences at the moment of buying intent. Two companies capture intent signals based on when customers visit certain types of websites more than normal. One public company just released a better product to realize this vision, and there are other smaller startups doing the same. Reach out if you'd like their names.
Content Creation: Faster First Drafts
On the content creation side, there are many startups executing all manner of media formats and smaller use cases for communication, entertainment, or marketing. In general, AI gets you faster first drafts but still requires experts to review, make edits, and distribute the final output. Humans are still the AI guardrails (jobs don't go away, they just move to higher-order tasks).
领英推荐
AI Business Agents
There are also AI Agents on the Business side that are beginning to string tasks together in a chain to help complete standard workflows. One of the things we’ve heard from SAAS users over the years is, “There are too many clicks.” Our workers have turned into mouse clickers instead of thinkers, so this should free up more strategic time instead of low-impact execution clicking time.
Knowledge Management
Knowledge management is another use case creating large amounts of value in bigger enterprises like Salesforce. Upload policy and procedures, training and learning material, and then let employees ask questions, saving time for humans.
Prevalence
We had a list of 170 companies that used ChatGPT to create new businesses. We then used ChatGPT to categorize and provide %s in each category to understand market trends. As you might expect, most of the use cases are business, not consumer:
Planetary-Scale Marketplace
Finally, we leave you with what the internet actually is. The internet is a two-sided marketplace. On one side are people who need things. On the other side are people who have things to fulfill that need. Google and Facebook sell ads using keywords and interests that attempt to fuzzy match buyer and seller, and creator and audience, but they are blunt tools because they are not accurate and are not presented at the precise moment of need.
AI solves that (wait: I though Bitcoin solved that :).
Jokes aside, we believe the most fundamental AI use case of all is inputting your needs and what you have to offer, then letting your AI agent (in work and personal life) go out to the internet, negotiate, act, pay, and complete things on your behalf using streaming payments to realize it.
If you want more insight on how to think from first principles on designing products that reach 5.4 billion people and 17 billion machines, have a look at our book.
It's $1. Stripe and the old-school payment systems take 35% of that dollar, which is insane. So we also offer the new-school way to pay, called the Lightning Network over Bitcoin, which only takes 1% of that. This is what Bitcoin solves.
You keep 99% of what you're selling instead of 65%. This is the moment where emerging tech stops being about something engineering focused and starts being about something that drives profitable growth. That's just smart business.
--Sean
Senior Managing Director
6 个月Sean Everett Great post! You've raised some interesting points.