April 20, 2024

Profitable, AI-powered companies without employees will arrive “next year”

Last year, DeepMind co-founder Mustafa Suleyman proposed a new type of Turing test (a way to test whether an AI has matched or surpassed human cognitive abilities): Can an intelligent system earn $1 million on its own? only?

It’s an idea that also speaks to one of the most attractive visions of an AI-driven future, where humans will no longer have to work because machines can create economic value on their own.

An increasing number of startups are trying to make these money-making autonomous businesses a reality.

Enter the AI ​​agent

Building an autonomous enterprise relies heavily on a type of AI known as “agents”; The basic concept is that you take all the capabilities of a powerful system like GPT-4, but give it agency. So instead of having to ask an AI every time you want it to give you a result, an AI agent allows you to assign it a big task and then go out and actually do the steps necessary to get there.

As an example, meet Alice. Alice is an outbound sales development representative (SDR) with one goal in life: to generate leads to pass on to her colleague, who will conduct the first meeting with the prospect. After some training on what her company does and who her typical customer is, she will go out and search the Internet for potential people or organizations to work with, and then reach out to them with personalized messages based on their profile and role. .

Alice, of course, is not a real person, but one of four so-called “digital workers” developed by the London-based startup 11x. Founder and CEO Hasan Sukkar tells Sifted that Alice is already capable of surpassing human benchmarks when it comes to successfully converting an opportunity into a meeting.

“It’s actually not difficult to be better than the average worker,” he says. “Early feedback from customers is that this has a huge return on investment. One of our clients is using Alice at scale, she had 10 people running her SDR function in a way that Alice does it on her own.”

It is not the finished article.

11x says it currently has 50 paying customers and an even longer waiting list of companies interested in employing its digital workers. But despite the initial appetite, Sukkar says Alice and her AI teammates still have some shortcomings.

One of them is due to the “hallucination” problem, where GenAI models create results that are inaccurate or not in line with the intended result.

“For example, last week I was reviewing something with one of our clients, where Alice had contacted someone and said, ‘Hey, I noticed you’ve been an accountant for 30 years, that’s pretty impressive.’ A human probably wouldn’t say this,” Sukkar explains.

While the 11x team is constantly working to refine the behavior of its AI agents, Sukkar says it’s issues like this that make it difficult for an AI to outperform the top 10%, as there is “an element of creativity and connectivity that It’s hard to emulate.” ”.

He adds that cost remains a limiting factor for widespread adoption of AI agents, due to the resource-intensive nature of running powerful AI models.

“We have a voice-based product that, to run from start to finish, will cost us $12 an hour, and sometimes this is more expensive than the minimum wage in certain countries,” says Sukkar.

Increase the market for AI agents

11x is just one of a growing number of startups creating AI agents to sell into the digital workforce, alongside others such as Sydney-based Relevance AI, Silicon Valley-based Lindy AI, and AirOps, with headquarters in Miami.

But none of these companies, which typically automate specific, isolated job functions, can yet promise what could be called a fully automated enterprise.

One company that is trying to address this problem and create some sort of company-wide system to monitor several different types of agents is London-based Honu.

Founder Imad Riachi describes technology as a system that “understands the problem space” of your business (that is, the market in which you operate, your competitors, and other elements of the problem the company is trying to solve) and organizes the AI agents accordingly.

“It creates a common language where intelligent agents or applications can collaborate to optimize and solve the most important business problem,” he says.

Honu is preparing to roll out its automated decision-making system to its first clients and Riachi says it will be able to work with a wide range of AI agents in the market. He adds that instead of just doing what he’s told and carrying out tasks decided by humans, Honu will be able to suggest ways to improve his business.

“It may seem esoteric, but it’s as if the ecosystem conspires to make your business run better,” Riachi says. “If you have $1,000 to spend, it will tell you where your money is best spent.”

A new economy

Riachi predicts that by next year we will see the first profitable fully autonomous business, and says it will probably be something like a small e-commerce company that comes first.

“They will be companies that have a fairly simple business model, that are highly digitalized and that have an interesting ecosystem of services around them that can be used to make that business successful,” he says, adding that the role of the CEO will be soon. You will become someone who knows how to run these AI systems.

Sukkar says that agents like those created by his company could very well be used in cost-saving measures (such as mass layoffs) and that, as a society, “we must be smart in the way we think and operate in this new reality.”

Both founders repeat the mantra that all companies will likely need fewer people because of this next wave of automation, but that the number of successful companies will increase significantly.

“When you build a decision infrastructure that uses AI to optimize the business, you make the bar for entrepreneurship much lower, so business creation skyrockets and default rates (of companies) fall by chopped,” he says.

In many ways, the rise of AI agents in the workplace simply continues a trend that has been happening for decades: software and AI automating certain functions of our jobs.

But if autonomous businesses arrive in the coming years, society might also have to start asking what kind of businesses we want to create, and whether we want an economy driven by an increasing number of e-commerce sites. build.

Leave a Reply

Your email address will not be published. Required fields are marked *