Palona AI, a Palo Alto startup building conversational agents for restaurants and retailers, secured $10 million in seed funding led by UpHonest Capital, the company announced January 23, 2025. Fusion Fund, NEO Investment Partners, former Yahoo CEO Maynard Webb, and a handful of strategic angels joined the round.
The pitch centers on something the industry has chased for years: voice agents that don't just take orders but actually sell. Palona's bots handle upselling, loyalty enrollment, and payment while plugging into existing point-of-sale systems. What sets the company apart, according to its founders, is a layer of what they call emotional intelligence baked into the models.
"We're enhancing these systems with EQ, humor, current-day texting etiquette, and gentle persuasion," said Steve Liu, Palona's co-founder and chief scientist, who also holds a professorship at McGill University and serves as Co-Director at Samsung AI Center in Montreal. Whether customers actually want their pizza order laced with gentle persuasion remains an open question, but early adopters seem willing to try.
Cloning the Founder's Voice
Palona launched its restaurant product in June 2025 with features that sound pulled from a sci-fi pitch deck: voice cloning that mimics a brand owner's vocal cadence, multilingual support, and a handoff protocol when the bot encounters something it can't handle. The customer roster so far includes regional chains and independent spots like Pizza My Heart, Twenty Five Teishoku House, Kobunga, MOTO Pizza, CaliBBQ, Pleasure Pizza Santa Cruz, and a handful of others scattered across California.
Lee Kindell, who founded MOTO Pizza, offered a testimonial that captures the uncanny appeal: "When I first heard my voice as our ordering AI, I honestly couldn't believe it." The technology works by recording samples of a human voice and training models to reproduce tone, pacing, and inflection with enough fidelity to fool casual listeners.
Palona also fields text-based agents for digital ordering, though the company has disclosed little about adoption rates or revenue. A recent partnership with Goodcall, a phone-answering service for small businesses, bundled Palona's voice engine into a product aimed at restaurant clients. Bob Summers, Goodcall's CEO, described the collaboration as delivering "the most sophisticated, reliable, and brand-authentic voice AI service in the industry." Industry observers might note that nearly every AI vendor claims similar superlatives.
Heavyweight Founding Team
The company's leadership reads like a Silicon Valley reunion. CEO and co-founder Maria Zhang spent years in senior engineering roles at Google and Meta, where she ran AI for products and later oversaw engineering at Tinder. Before that, she built Alike, a recommendation app Yahoo acquired in 2013 for an undisclosed sum. Her co-founder, Tim Howes, co-invented the Lightweight Directory Access Protocol in the 1990s and later co-founded Loudcloud and Opsware, which Hewlett-Packard bought for $1.65 billion in 2007. Most recently, Howes led developer productivity inside Meta's AI infrastructure group.
Palona operates under the legal name Proactive AI Lab, Inc., and the team has grown modestly since launch, though exact headcount figures remain vague. The company has added research talent, including Hai Huang, who previously collaborated with Yann LeCun at Google and now focuses on multimodal agent development at Palona.
Zhang posted on LinkedIn in June 2026 that Palona had processed "10 billion tokens and counting," though the company offered no independent verification of that figure. Token counts have become a favored metric in AI circles, but they tell you little about actual business traction or revenue.
A Crowded Field
Palona enters a market already populated by well-funded competitors. SoundHound AI has deployed drive-thru voice systems at quick-service chains and touts those deals in investor presentations. Presto Automation focuses on similar automation for drive-thru lanes. Smaller players like ConverseNow, Hi Auto, and Kea target phone and drive-thru ordering, according to CB Insights listings.
What remains unclear is whether restaurants will pay enough for these systems to support venture-scale returns. Labor costs are rising, and chains facing worker shortages have shown interest in automation. But voice AI still stumbles over accents, background noise, and the infinite variety of human requests. Palona's bet is that its emotional intelligence layer will smooth over those rough edges.
The company declined to disclose its valuation or total capital raised. Just weeks before the funding announcement, Palona brought on Alison Templin, a DoorDash veteran with eight years at the delivery giant, as vice president of sales to chase larger enterprise contracts. That hire suggests the company is moving beyond pilot deals and aiming for scale.
For now, Palona's technology page describes a "multi-modal, multi-model architecture" that stitches together voice, text, images, and video. How much of that exists in production versus on a roadmap is anyone's guess. The page carries no publication date, a small detail that might matter to customers trying to gauge what's real and what's aspirational.
