London startup Haiket switched on its waitlist this month with a deliberately narrow pitch: voice messages that understand the context of the conversation around them. The company calls it "context-aware" voice messaging, and it launched the app in Europe in November [PRNewswire, Nov 2025]. The product page is spare. One line: "Voice-first messaging, perfected. Join the waitlist" [Haiket].
That is the entire opening hand. A solo founder, a category most people associate with WhatsApp's little microphone button, and a thesis that the next interesting thing in messaging is not another inbox layout but the audio itself.
The bet
Haiket's wedge, as described in the launch materials, is shifting messaging "from screens to speech" [Crunchbase]. The product is consumer-facing, B2C, and built around the idea that voice notes today are dumb pipes: you record, the recipient plays back, end of transaction. Context-aware messaging, in Haiket's framing, means the app understands what the conversation is about and uses that to improve the experience of sending and receiving audio. The company has not published a feature list beyond the launch announcement, and the waitlist gate suggests a staged rollout rather than an open consumer push.
The choice of Europe first is notable. Voice notes are disproportionately popular in European and Latin American messaging behavior compared with the US, where typed chat still dominates. Launching in a region where the underlying user habit is already strong is a reasonable way to test whether "smarter" voice is a feature people will switch apps for, or merely a feature WhatsApp will eventually copy.
Why it could matter
The interesting macro story here is that voice has quietly become the default input layer for a large slice of mobile communication, and the AI stack has finally caught up to do useful things with it in real time. Transcription, translation, summarization, sentiment, and packet-level audio repair are all now cheap enough to run on a consumer app's unit economics. A founder who has worked at the intersection of those problems has a credible reason to think there is product surface here that the incumbents have not bothered to build.
John Donovan, the former AT&T executive who helped launch the iPhone on that carrier, lent his name to the launch announcement, saying Haiket is "leading the next evolution of messaging, making voice the center of digital connection" [PRNewswire, Nov 2025]. Endorsements are not term sheets, but Donovan's willingness to be quoted in the launch suggests the company has at least some operator-level support around it. No institutional investors have been disclosed.
The team
Haiket is the work of Alexander Narest, who is the founder and CEO [Yahoo Finance, 2025]. His background is the most concrete asset on the cap table. Narest worked on voice AI at Google, where he contributed to audio quality improvements in Google Duo using WaveNetEQ, the neural network DeepMind built for packet loss concealment [LinkedIn]. Before that, he worked on VoIP at Fring, one of the early mobile internet calling apps. He also built Holler, a voice messaging app with real-time translation into 65 languages [AppAdvice].
That is a tight resume for someone trying to ship a voice-first consumer app. WaveNetEQ in particular is the kind of infrastructure work, filling in dropped audio packets with synthesized waveforms so a call does not glitch, that gives a founder a real opinion about why most voice notes still sound bad. Holler is the closer precedent: a previous attempt by the same founder to layer AI translation onto voice messaging.
Google Duo (WaveNetEQ work) | 1 | product
Fring (VoIP) | 1 | product
Holler (65-language voice translation) | 1 | product
Haiket (context-aware voice messaging) | 1 | product
What the bears will say, and the bull answer
The most credible pushback is straightforward: voice messaging is a feature inside WhatsApp, iMessage, Telegram, and Instagram, and getting consumers to install a separate app to do something they already do for free inside the app where their friends already are is one of the harder problems in consumer software. Haiket has not disclosed user numbers, funding, or distribution partners, and the launch coverage is press-wire driven [PRNewswire, Nov 2025] [Investors Hangout, 2025]. The competitive set is implicitly every messenger on a European phone.
The bull answer is that incumbents are structurally bad at shipping voice-native experiences because voice is a side feature inside a text-first product, and the AI work needed to do context-aware audio well is not on any messenger's critical path. A small team that treats voice as the product, not a button, can plausibly ship something the incumbents will not match for a year or two. Narest's prior work on the exact subproblems involved (audio quality, real-time translation) is the reason to take the attempt seriously rather than dismiss it.
What to watch
Three things over the next twelve months will tell the story. First, what "context-aware" actually does in the shipped app once the waitlist opens, specifically whether the feature set is something a user can describe to a friend in one sentence. Second, a pre-seed or seed round with named European consumer investors, which Haiket has not yet announced. Third, any signal of organic pull in a single market, since consumer voice apps tend to win city by city before they win countries.
Narest has built voice infrastructure inside one of the largest communication products on earth and has now staked out a consumer brand of his own in a category the incumbents treat as a checkbox. The question for readers: if voice notes do become the dominant messaging modality in Europe over the next few years, is that a feature the platforms quietly absorb, or is there room for a standalone app to own the experience the way Snap once owned disappearing photos?