When a Zulay Kitchen shopper has a problem with a garlic press, the company would rather watch them explain it on camera than read a one-to-five star rating. That preference, scaled across an ecommerce support queue, is the wedge Hark is selling.
The New York startup, founded in 2022 by Fran Brzyski and Matt Ring, runs what it calls an AI-powered Voice of Customer platform built around video and audio submissions rather than traditional written surveys [Hark Website]. The pitch to direct-to-consumer brands is straightforward: a 30-second clip of a real customer showing a defect, a confusion, or a delight signal carries more usable context than a Net Promoter score, and modern transcription and clustering models can now turn those clips into something a CX or product team can actually act on.
The bet
Hark sits inside the support stack rather than trying to replace it. The product integrates with help-desk systems including Zendesk, and its Zendesk Marketplace app is bundled into all Hark subscriptions [Hark Website]. Pricing is structured without lock-in contracts or per-seat fees [Hark Pricing], a deliberate contrast with the seat-based economics that dominate enterprise CX software. For a 20-person ecommerce ops team running on Shopify and Zendesk, that pricing posture lowers the trial barrier considerably.
The core product loop, according to the company, is an AI feedback system that ingests customer-submitted video and audio, extracts themes, and routes the resulting insights to CX, product, and marketing teams [Hark Website]. One published case study, with kitchenware brand Zulay Kitchen, claims efficiency gains in support workflows after deploying Hark [Hark Case Study]. That is one named customer with a named outcome, which is more than most seed-stage CX startups put on the record.
The opportunity
The Voice of Customer category has been dominated for a decade by survey-first incumbents whose data outputs are largely structured numbers and free-text comments. The bet Hark is making is that generative AI changes the unit economics of unstructured feedback. Transcribing, tagging, and summarizing a thousand customer videos used to require either a research team or a willingness to ignore the data. If a model can do it overnight for the cost of API calls, video stops being the expensive feedback channel and starts being the rich one.
The cap table reflects investors who like that kind of thesis trade. Oceans Ventures led the $3.5 million seed round that closed in June 2024 [Yahoo Finance][Founder Lodge, 2024], with participation from Converge VC, Atman Capital, Alumni Ventures, BDMI Fund, Tenzing Capital, Gaingels, Hatchet Ventures, and Plug and Play Tech Center. An earlier pre-seed round brought in roughly $1.5 million [Hark Blog], putting total disclosed funding around $5 million.
Pre-seed | 1.5 | $M
Seed (June 2024) | 3.5 | $M
Total disclosed | 5.0 | $M
That is a deliberately modest war chest for a category with deep-pocketed competitors, and it suggests a capital-efficient go-to-market focused on mid-market DTC brands rather than a head-on enterprise sales motion.
The team and traction
Brzyski, Hark's CEO and co-founder, has spent significant airtime articulating the company's view that customer-generated video content is an underused asset for brands, including appearances on Churn FM, CXChronicles, Venture Everywhere, and The OP Show. The Fund profiled him in a founder spotlight that traces the origin of the video-feedback thesis [The Fund]. Co-founder Matt Ring rounds out the founding team. The company describes itself on its careers page as a venture-backed startup based in New York City.
Beyond the Zulay Kitchen case study, Hark has rolled out a product line called Hark Insights, which packages the captured feedback into what the company describes as actionable customer stories [Startup Weekly]. The Zendesk Marketplace listing gives the company a distribution channel into the installed base of one of the largest help-desk vendors, which matters more for a seed-stage company than a press release usually conveys.
The honest counterfactual
The credible bear case is competitive gravity. Zendesk itself, along with the broader CX suite incumbents, is shipping AI features into existing workflows that customers already pay for, and a Voice of Customer point solution has to demonstrate that its video-native approach generates insights the suite tools cannot. Hark's answer, visible in its product positioning [Hark Website], is that the suite vendors are optimizing structured ticket data and survey responses while Hark is building the data layer for unstructured customer-submitted media, a different ingestion problem with a different model stack. Whether that distinction holds up as the suites add multimodal features is the question that will define the next funding round. The Zendesk integration is a hedge: Hark wants to be the video layer on top of the queue, not a replacement for it.
What to watch
The next twelve months should answer three questions. First, can Hark publish a second and third named customer case study at Zulay Kitchen's level of specificity, ideally in different verticals than kitchenware. Second, does the no-seat pricing model translate into expansion revenue inside accounts, or does it cap deal sizes in a way that makes a Series A harder to raise. Third, with $5 million in disclosed funding and a seed round that closed in mid-2024, the clock on a Series A conversation is ticking; watch for either a priced round in 2025 or a strategic announcement around the Zendesk channel.
The broader question for readers: when AI makes unstructured feedback cheap to process at scale, does the survey itself become the legacy format, and which of the current crop of video-first VoC startups gets to define what replaces it?