- How to compete with ChatGPT, capable of providing an “objective” answer to almost any question? Create an even smarter machine capable of delivering even more objective answers? No, quite the opposite.
- You need to create a machine capable of providing “subjective” answers 😉 For example, opinions, advice, and recommendations from real people you already trust, or are willing to start trusting.
- Strangely enough, but this psychological trick works! Otherwise, this platform wouldn’t have been able to gain popularity without advertising and raise $6 million in investments in the very first round:
Project Essence
Dexa describes itself as a place “where curiosity meets authority.” On their website, you can find authoritative opinions from experts on health and fitness, AI and technology, business and finance, as well as personal development.
The sources of information are expert podcasts. Currently, Dexa has already collected and continues to add new podcast episodes from “hundreds” of selected experts to its database.
By visiting an expert’s page, you can see the latest episodes of their podcast—which you can watch or listen to in full or familiarize yourself with their brief content, compiled by the platform’s AI engine.
In addition, on this page, there is a list of the most popular questions that the expert answers in their podcasts. The list of questions and the texts of the answers are also generated by the AI engine.
You can not only view the site and pages of favorite experts but also utilize Dexa’s most useful feature—asking a question in the search bar and getting an answer to it.
You can ask a question “out of the blue”—then the AI engine will find the best answer from one of the presented experts. Or you can specify a specific expert—and then the AI engine will search for an answer only in their podcasts.
Answers are provided in the form of a short excerpt from a long video or audio track, in which the expert says something that can be considered an answer to the question asked by the Dexa user.
Dexa was created in February 2023, and its website is currently visited by about 50,000 people per month. Despite these fairly modest achievements, the startup recently raised its first round of investments, which amounted to $6 million.
What’s interesting
The founder initially created a tool for himself. He has a favorite author whose podcasts he listens to—and in one of them, this author talked about how magnesium helps to sleep better and get enough rest, and gave recommendations on how to take it for this purpose.
The guy got inspired, bought magnesium… but then realized that he didn’t remember how, according to the podcast author, he should take it. As a result, he had to spend several hours first finding the right podcast and then the place in the podcast where the author talks about the recommended method of taking magnesium. The problem was compounded by the fact that the author’s podcasts can last up to three and a half hours 😉
So the guy created a simple AI tool that would quickly find answers to questions in the podcasts of a specific author and shared it on Twitter. This tweet caught the attention of the podcast author, who reposted information about the tool, and the guy’s inbox exploded with requests to create a similar tool for searching for information in the podcasts of their favorite authors. And that’s how Dexa was born.
The founder of Dexa claims that he has created a new type of search engine.
“Google is good for finding facts, but not so good for finding advice and recommendations,” he explains his position, “Because in this case, the user wants to get not impersonal text, but an authoritative opinion of a person to whom he already trusts or can trust. Moreover, such opinions are usually hidden inside podcasts and videos that Google doesn’t even index.”
“Dexa’s ambitions go far beyond replacing 10 links [in a Google search] with two short paragraphs with an answer to a question [as ChatGPT or Perplexity does],” says one of the investors who participated in the current round. “They are creating a new principle of content indexing based on the opinions of specific people, not on the content of web pages. After all, people seek experience and wisdom, not just a digest of content.”
Essentially, this is a manifestation of a broader trend that people are most interested in other people—especially their personal opinions on various subjects. Without this, social networks, blogs, and podcasts would not have become so popular.
Searching for dry news or objective facts is one thing. Here, people usually turn to authoritative but therefore faceless “sources of information,” which can be news outlets or databases.
However, when it comes to interpreting news and facts, let alone advice and recommendations, people prefer to turn to other people. Because there is no objective truth here, and the only basis for accepting or rejecting information is the level of trust in the person presenting it.
True, when there were few real live people with their own opinions in the information space, many believed that “if it was said on TV,” it was true. Then the same people began to believe that “if it was the first thing found on Google,” it was the most authoritative and correct answer to the question. Now, those same people are beginning to believe that “if ChatGPT said so,” then it must be true 😉
However, today such people don’t even look for answers if their favorite blogger has already given them a ready-made opinion on the subject in their blog or podcast. So, the blogger seems to be still more authoritative than ChatGPT 😉
Summing up this lyrical digression, it can be said that there is definitely a rational grain in the general idea of providing “live human wisdom instead of a digest of content.”
Another startup that noticed this trend is Quench, which I wrote about in the summer of 2023. Back then, they called themselves “a learning tool for busy professionals.”
The essence is that busy people don’t have time to watch long educational videos. Instead, they prefer to quickly get short answers to their questions. Technologically, Quench did the same thing as today’s Dexa—Quench’s AI assistant shows short excerpts from long videos uploaded to the platform by experts that contain the answer to the student’s question.
Currently, Quench technically continues to do the same thing but has reformulated its offer more broadly into something like “Turn content into an AI assistant.”
Where to Run
I’ll start by saying that I don’t see any technological novelty in Dexa either. But more often than not, the key to a startup’s success lies not in unique technologies, but in how they apply already known technologies.
In the case of Dexa, it’s the replacement of “summarizing common opinion in one’s own words,” which is essentially what ChatGPT and its counterparts do, with direct quoting of personal subjective opinions.
It’s interesting that in different forms, the same principle can find application even in unexpected areas.
Take, for example, the topic of pet care and treatment. Right now, you can find a bunch of answers online to any question related to this. However, there’s a startup called Dr.Tail, which I wrote about last summer—and it charges $9.99 per month for its consultants to provide answers to such questions.
The coolest part is that the consultants themselves use an AI tool that searches for answers to these questions online 😉 But people are willing to pay for the fact that live consultants give them “authoritative” unequivocal and confident answers to their questions.
So, it turns out that this principle can be applied in many places. Dexa applies it to find advice and recommendations from experts in podcasts, Quench—for finding answers to students’ questions in educational videos, Dr.Tail—for providing veterinary advice.
So, what other product can you create using the same principle? In what areas are personal opinions, advice, and recommendations of specific people important to you personally, rather than generalized answers you can get from ChatGPT? Where can you get the initial information for this? How can you best convey to other people the difference and value of such an approach in a specific area?
About the Company
Dexa
Website: dexa.ai
Last funding round: $6M, 02/05/2024
Total investments: $6M, rounds: 1