
Futuro Corporation
Conversational AI Agent For SMB’s
I’m not much of a writer, but I do think that the story of how Futuro was built is a fascinating one, regardless of my inability to give it proper justice due to the lack of my aptitude for writing. When conversation AI first hit the market besides the fact that it sounded completely unnatural. The one thing I noticed was every Company was trying to create a perfect conversation. I’ve always felt that that was the exact opposite of where the industry should go. Human speech is not only imperfect it’s insanely imperfect. With so many people conditions by awful pre-recorded push button menus - a lot of people, especially on the other side will simply hang up if they hear what they determined to be not a person. Therefore, at least in my opinion, the Northstar industry should have always been. How do we replicate how humans speak. The reality is that is it’s quite difficult but the best way I can phrase it is it requires seeking out perfection in being imperfect. It’s a strange dichotomy as it’s the only area I can think of in AI where the goal is to purposefully have errors but without those even with the best voice quality clone, the lowest latency, etc. no one or I suppose I should say very few people will believe they’re speaking to a human during a 10 minute conversation there’s not a single “umm” or a breath after a long sentence or a pause while thinking and saying soooo or riiiighht. I’ve dedicated the last several years to truly finding the exact formula of how to integrate these elements into conversational AI in a way that a client can speak to one of our agents and the vast majority think they spoke to human representative I believe in the next few years entrepreneurship will be one of the few career choices available and the way I view it we can provide a entrepreneur a way to have 24 seven world class customer service, lead generation and even have the AI go through the whole sales cycle without entrepreneur having to put out a tremendous amount of money like they would have if they had to hire an employee. The last double blind that we ran 94% of the thousand people tested, answered absolutely not when they were asked if there was any chance that the representative they spoke to was not human and in fact, AI. I think that kind of speaks for itself to be honest. We’re going to keep pushing and grinding as the goal is 100% and it’s been one hell of a blast so far
About
I believe conversational is the most interesting category of the tech. I’ve always been fascinated by linguistics and technology so creating conversational AI that’s indistinguishable from human speech is fascinating

3 Comments
I like the idea that conversational AI shouldn't optimize for flawless speech—it should optimize for believable interaction.
People don't judge a conversation by whether every sentence is perfectly delivered. They judge it by whether the pauses, hesitation, pacing, and responses feel natural enough to sustain trust over time. That's a very different optimization target than simply making a voice sound more human.