Hearing aid costs and performance are a major source of complaint for many Americans who use, or should use, hearing aids. Probably the most intractable challenge facing manufacturers is how to make speech audible and clear in a noisy environment. It’s such a stubborn problem that it has its own widely used shorthand: Cocktail party syndrome.
Today’s digital hearing aids usually include features to dampen background noise, including directional microphones. Theoretically, it’s no longer necessary to take your hearing aids off in a noisy restaurant, as it was in the days of analog hearing aids. Analog aids primarily amplified sound, both sound you wanted to hear and sound you didn’t want to hear. Modern digital hearing aids filter background noise and direct the hearing aid microphones towards the sound you want to hear. For the most part they do it imperfectly.
But it’s possible that help is on the way. In December 2025 a new hearing aid came on the market. Called the Fortell, it’s made by a startup of the same name. The hearing aid looks like a conventional hearing aid with a behind-the-ear processor and an in-the-ear component, which can be a dome or a custom-made mold, housing a speaker. It acts pretty much like a conventional aid as well, with one significant difference. The Fortell has a program called Front Voices, which is activated manually and which is remarkably successful in isolating background noise, including other voices, from the speaker you want to hear. In a Zoom interview, Matthew de Jonge, one of Fortell’s founders, referred to this as “turbo” mode.”
The mission: An AI algorithm
Fortell was founded by de Jonge and Cole Morris, a college roommate and friend from an earlier startup, which was funded by the investor Joshua Kushner. When de Jonge and Morris approached Kushner about switching gears and trying to develop a better hearing aid, about five years ago, he immediately agreed to fund it. As a third founder, they brought on Igor Lovchinsky, an artificial intelligence (AI) expert de Jonge had worked with at a previous startup, who is now the company’s chief scientific officer.
Lovchinsky’s mission was to develop an AI algorithm that would identify the aural fingerprint of the primary speaker, and delete not only background noise but also other speakers’ voices. Once they figured out the algorithm, they needed to develop a chip that could deliver the information to the listener’s hearing aid. They recruited Andrew Casper, now the company’s chief technology officer and fourth co-founder, who was a lead engineer on a Google team making AI chips and, like Lovchinsky, a colleague from a previous startup.
They had the team, they had the plan, all they needed was the money. Kushner’s original $9 million investment might get them there but it would take time. As they began to achieve their goals for the novel AI program, and the chip to run it, other investors stepped in, bringing the funding to $150 million and expediting progress.
Fortell’s AI-powered hearing aid focuses on improving speech clarity in noisy environments. Its success depends on two factors. First is its proprietary AI algorithm, Spatial AI, which is trained to isolate voices in front of the speaker (as do conventional spatial AI algorithms) but also to recognize competing talkers in the background and nonspeech noise in all directions. A key step is source separation, separating speech from noise, something the brain does easily but was a challenge to develop for a computer. They then had to figure out how to separate the voice the speaker wanted to hear from other voices. Once they developed this program, under Andrew Casper’s direction, they developed a proprietary chip to administer it.
Testing efficacy
They’ve tested its efficacy in two studies, both of which have been submitted for publication at peer-reviewed journals. Their titles are confusingly similar, and I’ll refer to them as the Intelligibility study and the Preference study. The first author on each was Cole Morris of Fortell, with others named as co-first authors.
The Intelligibility study was completed in May 2025. The title is “Spatial AI Improves Speech Intelligibility for Hearing Aid Wearers in Challenging Multi-talker Noise.” The supervising consultants were William Shapiro and Mario Svirsky, both of N.Y.U. Langone. The second, the Preference study, completed in July 2025, is titled “Spatial AI Consistently Preferred to State of the Art Hearing Aids in Multitalker Noise.” The supervising consultants on this one were Joseph Montano, of Weill Cornell Medicine, and Barbara Weinstein, from the Graduate Center of the City University of New York and N.Y.U. Langone. They are all well known and highly respected. Their imprimatur is significant.
The Intelligibility study was a double-blind randomized controlled trial, with 30 hearing-aid users, comparing the Fortell to the Phonak Audeo Infinio Sphere, rated Best Overall Hearing aid of 2026 by Hearing Tracker. (Phonak is part of Sonova, which also makes the Advanced Bionics cochlear implant.) The Front Voices mode was activated on the Fortell. The control aid, the Phonak, was also set to maximum directionality and noise reduction. The Fortell ran the proprietary Spatial AI program. The Phonak ran a real-time neural network noise reduction algorithm, which also used source separation. (The ReSound Vivia and the Signia IX use a similar AI program but have not made the progress Fortell has.) The study used standard speech in noise testing (Quick SIN) to compare programs.
The Fortell hearing aid performed beyond expectations, significantly improving intelligibility of speech in noise. In a December 2025 article in Wired by Steven Levy, Mario Svirsky explained that the test goals were set out in advance. If the test showed a 4 decibel increase in boosting speech intelligibility, it would be a “home run.” The testing showed an increase of 9.2 decibels over the 3.1 decibel increase in the control. “The results were overwhelming,” Svirsky told Levy. “I’ve never seen such a categorical result in my career.”
Configurations and comparisons
When I spoke to Svirsky recently, he was still enthusiastic but a bit more reserved: The configuration chosen to test the Fortell aid, he said, “is a configuration that’s particularly good to show the advantages of this aid,” adding that different configurations might show different results. “For example, what if the noise sources were at +90 or -90 degrees? My educated guess is that the Fortell aid would still have an advantage over its competition, but this educated guess would have to be confirmed with actual behavioral testing of hearing aid users.”
Abram Bailey, AuD, of Hearing Tracker, was also enthusiastic, with a caveat. He said the independent HearAdvisor acoustic lab, of which he’s a cofounder, tested a prototype of the Fortell hearing aids, which revealed better speech in noise performance than any the lab has tested to date. When I asked him if he could comment on the magnitude of the difference, he said he would want to do further testing before commenting on how much better Fortell is versus the competition.
The second test, the Preference study, was conducted in July 2025. It was more subjective, asking test participants to compare the Fortell hearing aid to five top hearing aids from major manufacturers. This study, also randomized-controlled, used recordings of the Fortell and of five top-ranked hearing aids from major manufacturers. Each participant ranked paired recordings (one from the Fortell, one from one of the five controls). The test used an A/B format: Which is clearer, A or B? The test is similar to an optician testing your vision by adding and removing lenses: Which is better, this one or that one? The Preference study question was “In which sample is it easier to understand what the target speaker is saying?” Each participant answered 100 A/B questions.
After participants chose A or B, they responded to a multiple-choice question about how much easier one was over the other. 1) A much easier. 2) A slightly easier. 3) Both samples similar. 4) B slightly easier. 5) B much easier. The results overwhelmingly favored the Fortell Spatial AI, by 95 percent. Here is a link to the study itself to see more about how the test was administered.
On the street
In early testing, described by Levy in the Wired article, Fortell staffers would follow up the testing with a trip down to the street in noisy Soho, where the offices were then located. In addition to street noise, two Fortell employees stood behind the person trying out the hearing aid, having a loud conversation of their own.
Wired had fun in describing what happened next: “Despite the din, the testers clearly make out what the person in front of them is saying. The clouds lift. Angels croon. ‘This was so incredible that I burst into tears,’ says Ashley Tudor, one of the beta testers.” She wasn’t the only person so moved: “Even with construction behind me, I could hear Matt clearly,” another person recalled, “and [it] literally brought me to tears.”
I can’t test the Fortell because my hearing loss is too severe, even in the ear with a hearing aid (the other has a cochlear implant). But although I probably wouldn’t cry, I would be impressed to be able to hear under those circumstances. I can’t even hear at my own dinner table if there are more than a couple of people speaking at the same time.
Who can use them
Can you get the Fortell hearing aids? Yes, with restrictions. The first has to do with your own hearing loss, which must meet Fortell guidelines. The Fortell is intended for bilateral loss, but it can be of any severity that a conventional high-end aid would work for. In the beta testing that meant primarily people with age-related hearing loss, or presbycusis. If the Fortell audiologists don’t think the hearing aid could help, they may suggest a cochlear implant evaluation. Asymmetrical loss requires medical clearance, under New York State law. I asked whether they might eventually treat single-sided loss. Maybe, he said. How about when the buyer, like me, already has a cochlear implant in one ear? Could I buy a pair and use just one? Theoretically yes, he said: One would go in the ear and one would stay in the box. It would not sync with the cochlear implant, as my current hearing aid does.
The hearing aid looks like most hearing aids, bigger than some, smaller than some. The algorithm software, which initially took a computer-sized processor to run, requires a large chip. The aids come with two rechargeable batteries, which lasted for 18 hours of continuous use in Fortell tests.
The hearing aid programs can be changed on the aid (programs on the right aid, volume on the left) or on an iPhone app. The Fortell has three basic programs, or modes. The first is “All Day AI,” which itself has two settings: All Voices (omnidirectional, usually the default mode), which auto-optimizes in response to the hearing environment. All Day AI also includes the Front Voices mode, which is accessed manually within the All-Day program.
The second is Conventional, similar to a traditional high-end hearing aid that de-activates the novel AI but still uses state-of-the-art noise suppression and other techniques. As de Jonge said, “It’s for those moments in life where you don’t want AI deciding what you want to hear.” The third program is Live Music, for concerts and performances, which optimizes sound in various ways for a richer listening experience. Each mode is tuned by a Fortell audiologist based on the user’s needs and preferences. The user manually changes modes.
Telecoil, Auracast
It does not have a telecoil – the feature that lets the user connect to a hearing loop like those in some public venues — for space reasons. The AI programming that originally needed a computer is now being inserted into a small hearing aid.
The hearing aid is Auracast capable, but if you live in New York City, there are not many Auracast-equipped venues. (Auracast is a new Bluetooth LE audio-sharing technology that works for large groups, much like hearing loops.) In addition, the software would have to be updated to use Auracast. If you use hearing loops or Auracast a lot, this is not the hearing aid for you. It connects by Bluetooth to the iPhone but not, so far, to Android phones. The price includes five years of unlimited audiological care, a three-year warranty and insurance for loss or damage. There is a 100-day full refund window.
The Fortell is a very good hearing aid, but the reason to buy it is for the speech in noise program, the turbo program, Front Voices. Basically, the Fortell works like any other high-end hearing aid until you switch to Front Voices. That’s when the beta testers started crying. Front Voices can only be accessed manually. De Jonge describes it as “very very aggressive intervention,” and one that should be under the user’s control: “We don’t ever want to do that [activate Front Voices] at the wrong time for you, because it literally eliminates people. It brings the voices of people you’re not looking at to zero.” The default mode, “conventional,” turns off the AI and turns on omnidirectional microphones. “It sounds like a conventional state-of-the-art hearing aid,” de Jonge said.
Cost and availability
The cost is $6,800 per pair, which is high but not excessive. Most people I know, with moderate to severe or profound loss, pay $3,000-$4,000 per aid. (Many people with that level of loss have a cochlear implant in one ear and a hearing aid in the other.) Overall, according to a new report on hearing-aid prices from Hearing Tracker, prices on high-end hearing premium brand aids typically range from $3,500 to $5,200 per pair, depending on the degree of loss being treated, the features, where they are bought, and whether insurance or the buyer is paying. Medicare does not cover hearing aids, and most other plans cover only a portion of the cost, or they require that you buy their brand of hearing aid. Costco is the largest retailer of hearing aids in the U.S. second only to the Veterans Administration. Their prices for brand-name hearing aids, both brand-name prescription and over the counter, are considerably lower than you would pay at a private audiologist.
At the moment, the hearing aids are sold only at Fortell headquarters in Manhattan. You can call or apply online. Here’s a link to the website. The first available appointment for a new customer is April. Those who have already bought the hearing aid are seen much more quickly for adjustments and programming. Potential buyers apply online or by phone. The criteria are explained and they are asked to submit a recent audiogram. If they don’t have one, Fortell’s clinicians will administer the test.
Fortell has four audiologists on staff, and will expand to six in February. Fitting is done at Fortell’s office at Park Avenue and 61st Street. Buyers need to return at least a few times for fitting and adjustments, so they need to be able to get to New York.
Eventually, de Jonge said, the Fortell might be dispensed by audiologists outside of Fortell, but the choice of audiologists would be very selective. Also eventually, the Spatial AI program and the chip might be made available to other hearing aid companies. But it’s a highly competitive field, so don’t expect to see this soon.
Hearing aids improve all the time, developing new ways to improve the hearing experience, especially in noise, and new features. What used to be available only at the high end is now available at Costco or even over the counter. The Fortell technology seems like a genuine substantive advance, one that other companies will no doubt be working to match. The Fortell is not right for everyone, but it does seem likely to push manufacturers to come up with their own new algorithms and chips and approaches.
Hearing aids have been getting better and better, but progress has been incremental in the years since the technology made a major change from analog to digital. The improvements thus far have been significantly better but not revolutionary. Fortell’s technology could turn out to be the first in a whole new approach to making hearing aids as effective as the human ear when trying to hear speech in a noisy place. Cocktail party syndrome may be on the way out, just as the cocktail party itself has faded away.
Katherine Bouton is the author of the memoir Shouting Won’t Help, as well as Smart Hearing, a guide to living with hearing loss. Her blog Smart Hearing covers all aspects of hearing loss. She is a former editor at The New York Times. .
