Guava

Cover Block

Publicly reported

Field	Detail
Name	Guava
Tagline	Enterprise voice-AI platform providing a proprietary voice stack as an API for high-stakes call environments. [Business Wire/National Law Review, November 2026]
Headquarters	Los Angeles, CA [LinkedIn/Omri Dahan, retrieved 2026]
Founded	2012 [getmacha.com]
Stage	Seed [Business Wire/National Law Review, November 2026]
Business Model	API / Developer Platform [Guava FAQ]
Industry	Deeptech
Technology	AI / Machine Learning
Growth Profile	Venture Scale [Business Wire/National Law Review, November 2026]
Founding Team	Omri Dahan, Anthony Scodary, Nico Benitez, Mike Piechna [Business Wire/National Law Review, November 2026]
Funding Label	Seed
Total Disclosed	Approximately $12.5 million [Business Wire/National Law Review, November 2026]

Links

PUBLIC

• Website: https://goguava.ai

Summary and Signal

PUBLIC Guava builds a voice AI API for regulated and high-stakes call environments, and it merits attention now because it surfaced with a $12.5 million seed round in November 2026 at a moment when enterprise buyers are starting to separate demo-quality voice agents from systems built for production reliability [Business Wire/National Law Review, November 2026] [Axios Pro, November 2026]. Public materials place the company in Los Angeles and trace its founding to 2012, with the current financing announcement framing Guava as a more visible commercial push rather than a brand-new formation [Business Wire/National Law Review, November 2026] [getmacha.com, 2026].

The product pitch is straightforward: one proprietary voice stack, delivered as an API, that combines speech recognition, language processing, and text-to-speech rather than asking customers to assemble separate layers themselves [Guava FAQ] [goguava.ai, 2026]. That positioning is aimed at healthcare, banking, insurance, government, and BPO workflows where latency, auditability, and call resolution matter more than novelty, and Guava also publicly claims SOC 2 Type II, HITRUST i1, and PCI DSS Level 1 compliance, with BAAs available for healthcare use cases [Yahoo Finance] [Scommerce, 2026] [goguava.ai, retrieved 2026].

The team reads as commercially and technically relevant to that thesis. Co-founder and CEO Omri Dahan previously served as chief revenue officer at Marqeta until early 2021, while co-founder Anthony Scodary's background includes engineering work at NASA's Jet Propulsion Laboratory, and co-founder Nico Benitez previously worked at cloud gaming company OnLive [Bloomberg] [Business Insider, 2021] [Crunchbase/Anthony Scodary] [speechtek.com, 2019] [resources.gridspace.com] [theorg.com].

On funding, Guava has one publicly confirmed round: a $12.5 million seed led by Acrew Capital with participation from Flourish Ventures, WillowTree Ventures, Clocktower Ventures, and Commerce Ventures [Business Wire/National Law Review, November 2026] [Axios Pro, November 2026]. The business model appears to be an enterprise API platform sold into large organizations and BPOs, but the key operating question for the next 12 to 18 months is whether public claims of more than a dozen enterprise customers handling tens of thousands of daily calls convert into referenceable deployments, durable compliance-led expansion, and evidence that the integrated stack wins on reliability rather than only on model quality [Citybiz, November 2026] [Coval Benchmarks].

One source, partially checked -- Core financing and founder details are corroborated by multiple public sources, while product, compliance, and traction claims rely partly on company materials and announcement coverage.

Taxonomy Snapshot

Axis	Value
Stage	Seed
Business Model	API / Developer Platform
Industry / Vertical	Deeptech
Technology Type	AI / Machine Learning
Growth Profile	Venture Scale
Founding Team	Co-Founders (3+)
Funding	Seed, total disclosed about $12,500,000

Company Overview

PUBLIC

Guava is a Los Angeles based enterprise voice AI company founded in 2012, according to the company website and founder profile materials surfaced in public records [goguava.ai, 2026] [LinkedIn/Omri Dahan, retrieved 2026]. In plain terms, the company sells a voice stack as an API for enterprises that run sensitive phone interactions, with public materials pointing to healthcare, banking, insurance, government, and business process outsourcers as the core target set [goguava.ai, 2026].

The chronology that can be established from public sources is narrow but useful. The business traces its founding to 2012 [goguava.ai, 2026], and in November 2026 it publicly announced a $12.5 million seed round led by Acrew Capital, with participation from Flourish Ventures, WillowTree Ventures, Clocktower Ventures, and Commerce Ventures [Business Wire/National Law Review, November 2026]. The same announcement identified Omri Dahan, Anthony Scodary, Nico Benitez, and Mike Piechna as co-founders [Business Wire/National Law Review, November 2026].

What stands out is the age of the company relative to the financing stage. Guava's homepage says it has been in regulated production for 13 years, a claim that is directionally consistent with a 2012 founding date, though the public record in this packet does not fully map the intervening operating history year by year [getmacha.com, 2026]. That leaves investors with a company that appears mature in product history but newly visible in venture financing terms.

One source, partially checked -- Founded date and Los Angeles footprint are supported by company and profile sources; the November 2026 financing is confirmed by the company announcement, but legal entity details and a full milestone history are not established in the cited public record.

The Product and the Stack

MIXED

The product claim is straightforward, even if the supporting evidence is still mostly company-originated. Guava describes itself as a proprietary voice stack delivered as an API for high-stakes call environments such as healthcare, banking, insurance, government, and business-process outsourcers, with automatic speech recognition, language processing, and text-to-speech built as one system rather than assembled from separate vendors [Guava FAQ, 2026] [goguava.ai, 2026] [Business Wire/National Law Review, November 2026]. That matters because the company is positioning reliability, compliance, and call resolution as the core product surface, not a general-purpose voice demo, and its public materials also reference WebRTC and SIP connectivity plus sub-1 second latency for enterprise deployments [goguava.ai, 2026].

The public proof points lean more heavily toward operating posture than customer-level verification. Guava says it can go live in 14 days, signs business associate agreements when protected health information is involved, and holds SOC 2 Type II, HITRUST i1, and PCI DSS Level 1 credentials, claims that are repeated across company materials and third-party pickup rather than independently audited in the sourced record here [Guava | Voice Agent API, 2026] [goguava.ai, 2026] [Yahoo Finance, 2026] [Scommerce, 2026]. On model quality, Coval's published benchmark shows Guava's Daytona TTS ranked first for accuracy among 30 text-to-speech models, with substitutions of 0.8 percent, deletions of 0.5 percent, and insertions of 0.0 percent, which is the clearest public technical datapoint in the set [Coval Benchmarks, 2026].

The technical narrative is promising, but the evidence base is uneven. There is enough public material to establish what Guava says it has built and where it intends to deploy it, but less independent documentation on how the integrated stack performs across real production workloads beyond benchmark results and company-distributed claims [Citybiz, November 2026] [Axios Pro, November 2026].

One source, partially checked -- Product scope is corroborated across company materials and funding coverage, but several important claims, including deployment speed, latency, and compliance posture, remain primarily company-sourced.

The Market They Are Entering

Publicly reported The market matters now because enterprise buyers have moved from experimenting with voice agents to testing whether they can run regulated, customer-facing call flows without adding compliance or resolution risk, and Guava is positioned against that narrower problem set rather than the broad consumer voice market [Axios Pro, November 2026] [Business Wire/National Law Review, November 2026].

The evidence base for formal market sizing is thin in the materials at hand, so this section stays conservative and uses adjacent public markets rather than inventing a TAM. Public reporting confirms Guava sells an API-based voice stack for healthcare, banking, insurance, government, and business-process outsourcers, with the product framed for calls "that have to be right" and for deployments where compliance is a buying criterion [Guava FAQ] [Yahoo Finance] [Scommerce, 2026]. That points to an addressable spend pool sitting at the overlap of contact center software, enterprise AI infrastructure, and vertical compliance tooling, but no cited third-party report in this record quantifies Guava's specific serviceable market.

Demand drivers are easier to document than market size. The November 2026 financing coverage and company materials point to the same practical customer need: voice AI that can move from demo quality to production use in sectors where errors are costly, latency matters, and procurement requires security review [Axios Pro, November 2026] [Business Wire/National Law Review, November 2026] [goguava.ai, 2026]. Guava's public positioning also emphasizes SOC 2 Type II, HITRUST i1, PCI DSS Level 1, and healthcare business associate agreements, which suggests the sales motion is tied not only to automation budgets but also to governance requirements that generic developer tools may struggle to satisfy in enterprise workflows [Yahoo Finance] [Scommerce, 2026] [goguava.ai, retrieved 2026].

The adjacent markets are clear even if boundaries between them remain unsettled. One substitute path is the multi-vendor stack, where enterprises assemble speech recognition, language processing, text-to-speech, telephony, and orchestration from separate providers; Guava's stated differentiation is that it bundles those layers into one proprietary API [Guava FAQ] [goguava.ai, 2026]. Another adjacent market is traditional contact center software, where incumbents already control call routing and agent workflows but may not own the full voice model layer. A third is the broader enterprise automation and BPO tooling market, especially where outsourced operators want a resellable or embeddable intelligence layer rather than a standalone application [goguava.ai, retrieved 2026].

Macro and regulatory forces cut both ways. On the positive side, sectors such as healthcare, banking, insurance, and government have large call volumes and a recurring need to document, route, and resolve conversations consistently, which can support adoption if model quality is good enough [Business Wire/National Law Review, November 2026] [Citybiz, November 2026]. The constraint is that these same sectors tend to impose longer security reviews, vendor diligence, and liability sensitivity, which raises the bar for claims around accuracy, latency, and deployment speed. Guava's public materials lean directly into that tension, including a claim that customers can go live in 14 days and a benchmark result for its Daytona TTS model, but investors should still treat those as product-positioning signals until broader third-party adoption data is visible [Guava | Voice Agent API] [Coval Benchmarks].

Market lens	What the cited evidence supports	Sources
Core market	Enterprise voice AI for regulated and high-stakes call environments	[Axios Pro, November 2026], [Business Wire/National Law Review, November 2026], [Guava FAQ]
Buyer segments	Healthcare, banking, insurance, government, and BPOs	[Guava FAQ], [Business Wire/National Law Review, November 2026], [Citybiz, November 2026]
Substitute market	Multi-vendor voice stack assembled from separate ASR, NLP, TTS, and telephony vendors	[Guava FAQ], [goguava.ai, 2026]
Adjacent market	Contact center and workflow software with AI layers added on top	[Business Wire/National Law Review, November 2026], [goguava.ai, retrieved 2026]

The useful read-through is not a headline TAM figure, because the source set does not support one. The more defensible conclusion is that Guava is aiming at a higher-friction, potentially higher-value slice of enterprise voice spend where compliance, latency, and call resolution shape procurement more than raw model novelty.

One source, partially checked -- Section conclusions rely on public financing coverage and company materials, with limited independent third-party market sizing in the cited record.

The Competitive Field

Positioning

MIXED Guava appears to be positioning itself above generic voice tooling and closer to a regulated-enterprise infrastructure layer, where the selling point is not simply voice generation quality but an integrated stack built for calls that carry compliance, latency, and resolution requirements [Guava FAQ, 2026] [goguava.ai, 2026].

Company	Positioning	Stage / Funding	Notable Differentiator	Source
Guava	Enterprise voice-AI API for healthcare, banking, insurance, government, and BPO call environments	Seed, $12.5M disclosed	Positions around one proprietary voice stack, API delivery, compliance posture, and high-stakes production calls	[Business Wire/National Law Review, November 2026] [Guava FAQ, 2026] [goguava.ai, 2026]
Vapi	Developer platform for voice agents	Not established in the provided research	Named competitor in the structured facts	[Structured Facts]
Retell AI	Voice AI platform	Not established in the provided research	Named competitor in the structured facts	[Structured Facts]
Synthflow	Voice AI platform	Not established in the provided research	Named competitor in the structured facts	[Structured Facts]
Bland	Voice AI platform	Not established in the provided research	Named competitor in the structured facts	[Structured Facts]
OpenAI	Adjacent model and platform provider	Not established in the provided research	Named competitor in the structured facts as an upstream model-layer substitute for some buyers	[Structured Facts]

The table reads as sparse because the public record supplied here is much denser on Guava than on the alternatives. That asymmetry matters on its own: what can be said with confidence is that Guava is trying to compete in a narrower lane than broad voice-agent builders, and that lane is enterprise infrastructure for sensitive call flows rather than horizontal experimentation [Guava FAQ, 2026] [goguava.ai, 2026].

The competitive map breaks into three layers. Third are incumbent internal paths, meaning enterprises and BPOs that stitch together automatic speech recognition, language processing, and text-to-speech vendors on their own, which is precisely the workflow Guava argues it can replace with a single integrated API [Guava FAQ, 2026] [goguava.ai, 2026].

Guava's clearest edge, based on public evidence, is not scale but fit for a specific procurement motion. The company states that it has built proprietary ASR, TTS, and language models as one stack, offers WebRTC and SIP support, targets sub-1s latency, and maintains SOC 2 Type II, HITRUST i1, and PCI DSS Level 1 credentials, with BAAs available for healthcare workloads [goguava.ai, 2026] [Yahoo Finance] [Scommerce, 2026]. If those claims hold in customer diligence, the advantage is that regulated buyers can source fewer vendors and shift less integration risk in-house. The durability is less certain. Compliance artifacts can create real friction for new entrants, but model quality and API packaging are perishable edges if larger platforms close the gap or if competitors earn the same certifications over the next 12 to 18 months [Yahoo Finance] [Scommerce, 2026].

The company is most exposed where breadth and ecosystem matter more than specialization. There is also a narrower category exposure: Guava's messaging is strongest in high-stakes call environments, which can be an asset in banking or healthcare but may leave it less relevant in lower-risk, high-volume use cases where buyers optimize first for price, experimentation speed, or general-purpose tooling [Guava FAQ, 2026] [goguava.ai, 2026].

The most plausible 18-month scenario is a split market rather than a single winner. Guava looks best positioned if enterprise buyers keep moving away from stitched-together voice stacks and begin treating compliance, latency, and deployment accountability as first-order selection criteria. In that case, Guava is a plausible winner if regulated call centers standardize on specialist infrastructure over modular assembly, particularly given its stated traction with more than a dozen enterprise customers handling tens of thousands of calls daily as of November 2026 [Citybiz, November 2026]. The more vulnerable outcome sits with undifferentiated horizontal voice tooling.

Opportunity

Upside Path

PUBLIC The prize here is not a better call bot, it is a chance to become core infrastructure for regulated voice interactions, a layer enterprises standardize on when the cost of a bad call is high and the buyer cares as much about compliance and latency as model quality [Business Wire/National Law Review, November 2026] [goguava.ai, 2026] [Yahoo Finance] [Scommerce, 2026].

The headline opportunity is straightforward. If Guava can prove that an integrated voice stack performs more reliably in production than stitched-together point solutions, it has a plausible path to becoming the default API for healthcare, banking, insurance, government, and BPO voice deployments, where call quality failures can create regulatory exposure, lost revenue, or both [Guava FAQ] [goguava.ai, 2026]. That outcome is reachable, rather than purely aspirational, because the company has at least some early signals in the right places: a disclosed $12.5 million seed round led by Acrew Capital, investor participation from firms that have backed regulated-infrastructure businesses, product positioning centered on certifications and BAAs, and a public claim that more than a dozen enterprise customers were already handling tens of thousands of calls daily on the platform as of November 2026 [Business Wire/National Law Review, November 2026] [Axios Pro, November 2026] [Citybiz, November 2026] [goguava.ai, retrieved 2026].

The upside cases branch from that same starting point: one product surface, sold into environments where replacement cost is high once workflows are embedded.

Scenario	What happens	Catalyst	Why it's plausible
Regulated voice system of record	Guava becomes the preferred voice API for healthcare, insurance, banking, and government teams that need compliance and production reliability, not just low-cost inference	Certifications and contracting readiness, including SOC 2 Type II, HITRUST i1, PCI DSS Level 1, and BAAs for healthcare buyers [Yahoo Finance] [Scommerce, 2026] [goguava.ai, retrieved 2026]	The product is already framed around high-stakes sectors, and enterprise buyers in those categories often consolidate on vendors that clear security and compliance review early [Guava FAQ] [Business Wire/National Law Review, November 2026]
BPO distribution engine	Guava wins outsourced call operators and becomes the embedded platform behind many end-enterprise deployments at once	Expansion of go-to-market for enterprises and BPOs using the new seed financing [Business Wire/National Law Review, November 2026]	BPOs can serve as force multipliers because one platform decision can propagate across multiple client programs, and Guava explicitly names BPOs as a target segment [Guava FAQ] [Business Wire/National Law Review, November 2026]
Performance-led platform adoption	Guava uses model performance and low latency to win developers first, then climbs into enterprise standardization	Independent benchmark recognition for Daytona TTS and public claims of sub-1s latency with WebRTC and SIP support [Coval Benchmarks] [goguava.ai, retrieved 2026]	In infrastructure markets, measurable quality advantages can pull in initial use cases before procurement broadens the relationship, particularly when deployment can start through an API rather than a full stack rip-and-replace [Guava

The compounding mechanism, if it emerges, would come from three layers reinforcing one another. First, every production deployment in a regulated or high-volume call flow can generate harder edge cases around turn-taking, synthesis, context management, and workflow compliance, which should be more valuable to an integrated stack than to a vendor assembled from separate ASR, language, and TTS components [goguava.ai, 2026] [Guava FAQ]. Second, certifications such as SOC 2 Type II, HITRUST i1, and PCI DSS Level 1 can shorten the path into adjacent buyers once one department or one outsourcing partner has already approved the platform [Yahoo Finance] [Scommerce, 2026]. Third, if the company can actually deliver fast implementation, its claim of going live in 14 days suggests a land-and-expand motion where pilots convert into broader call types without long rebuild cycles [Guava | Voice Agent API]. The public evidence does not yet confirm that this flywheel is operating at scale, but the ingredients are visible.

The size of the win is easiest to frame by reference to enterprise infrastructure outcomes rather than current revenue, which is not disclosed. A scenario where Guava becomes one of a small number of standard platforms for regulated voice automation could support a multibillion-dollar outcome, particularly because public markets and strategic buyers have historically assigned substantial value to infrastructure companies that sit inside critical transaction or communication flows. Omri Dahan's prior operating role at Marqeta is relevant mainly as a pattern: Marqeta built category-level value by becoming embedded infrastructure rather than an end-user brand [Bloomberg] [Business Insider, 2021-05] [Reuters]. If Guava were to follow an analogous path inside voice, the upside could be venture-scale in the strongest sense, with a valuation measured in billions of dollars (scenario, not a forecast). That remains conditional on proving that the integrated architecture translates into durable enterprise adoption, but the public record is enough to say the target outcome is large and intelligible, not theoretical.

One source, partially checked -- Funding is corroborated by Axios Pro and Business Wire/National Law Review, while most product, compliance, deployment, and performance claims rely on company materials or a single benchmark source.

Sources

Publicly reported

1. [Business Wire/National Law Review, November 2026] Guava Lands $12.5 Million to Tackle Voice AI’s Real-world Growing Pains | https://natlawreview.com/press-releases/guava-lands-125-million-tackle-voice-ais-real-world-growing-pains

2. [LinkedIn/Omri Dahan, retrieved 2026] Omri Dahan | LinkedIn | https://www.linkedin.com/in/omridahan/

3. [Axios Pro, November 2026] Exclusive: AI voice startup Guava raises $12.5M | https://www.axios.com/pro/enterprise-software-deals/2026/11/03/voice-agents-guava-gridspace

4. [Yahoo Finance, retrieved 2026] Guava Lands $12.5 Million to Tackle Voice AI’s Real-world Growing Pains | https://finance.yahoo.com/news/guava-lands-12-5-million-130000000.html

5. [Scommerce, 2026] Guava raises $12.5M seed round led by Acrew Capital to scale enterprise voice AI | https://scommerce.com/guava-raises-12-5m-seed-round-led-by-acrew-capital-to-scale-enterprise-voice-ai/

6. [goguava.ai, retrieved 2026] Guava FAQ | https://goguava.ai/faq

7. [Bloomberg, retrieved 2026] Omri Dahan, Marqeta Inc: Profile and Biography | https://www.bloomberg.com/profile/person/17766983

8. [Business Insider, 2021] Goldman-backed fintech Marqeta nabbed an AWS exec to serve as chief revenue officer as it prepares for an IPO | https://www.businessinsider.com/marqeta-nabs-aws-exec-losing-hr-leader-ahead-ipo-fintech-2021-5

9. [Citybiz, November 2026] Guava Raises $12.5M Seed Round Led by Acrew Capital to Scale Enterprise Voice AI | https://www.citybiz.co/article/928613/guava-raises-12-5m-seed-round-led-by-acrew-capital-to-scale-enterprise-voice-ai/

10. [Business Insider, 2021-05] Goldman-backed fintech Marqeta nabbed an AWS exec to serve as chief revenue officer as it prepares for an IPO | https://www.businessinsider.com/marqeta-nabs-aws-exec-losing-hr-leader-ahead-ipo-fintech-2021-5

11. [Reuters, retrieved 2026] Mastercard backs card issuing start-up Marqeta | https://www.reuters.com/article/us-mastercard-marqeta/mastercard-backs-card-issuing-start-up-marqeta-idUSKBN26T1WN/