Enterprise voice AI targeting long, multi-goal calls: Voice Memory™ (cross-call context), Emotional Awareness (prosody-based escalation), an AI supervisor over every call, an MCP-first action framework, claimed own voice/speech models and (aspirationally) owned GPUs, 1:20 supervision ratio. Domain aevox.ai (registered 2025-11-18, privacy-proxied, zero Wayback captures, zero press). HQ 952 Atlantic City Blvd, Bayville NJ. Founded 2025 by brothers Dan and Jason Rodd (ex call-center operators, DriveSmart).
Stage: 2 logos live and billing (one is the CEO's own company), ~$500K founder capital, pre-institutional, "all inbound." Entity is a mess: "AeVox Corp" (deck) / "AEVOX AI CORP" (EIN 42-4404871, Dover DE registered-agent address, issued 2026-08-12 — one month before the raise) / "a Subsidiary of Cubix LLC" (brand book) / site footer "a Cubix AI LLC company"; founder emails on cubixai.dev. Delaware ICIS search returned zero "AEVOX" entities (2 attempts, possibly bot-filtered — needs manual re-check). Which entity operates, bills, and signed DebtHunch is unresolved. No cap table in the room at a $30M-post ask.
The honest bull case (from the competition sweep): two operators who own a call center in a compliance-heavy, unglamorous vertical (auto warranty / debt / collections outbound) that the funded players mostly avoid because of TCPA exposure. The bear case: that same TCPA exposure is already visible in the founder's other company's docket.
Missing: certificate of incorporation, cap table, charter, historical financials, patent filings, the executed DebtHunch order form, the drafted ABAX MSA, advisor agreements, round docs. Corporate folder = one EIN letter; financials folder = one projection workbook.
Quality read: the traction doc is unusually candid (it discloses the related party and the real contracted minimums) — but the deck headline, the public site, and the enterprise deck all say more than the traction doc supports. The room's own disclosures are the honest layer; everything above them inflates.
Verified: LinkedIn: CEO AeVox AI; Founder DriveSmart Administration Inc.; Rutgers B.E. Electrical/Electronics 2008–2012; MIT "Professional Education" = one 4-month course (Apr–Aug 2024). Apex United CEO confirmed. DriveSmart is a real multi-state operation (WI insurance registry, DC corp registry, Arkansas call center 2019). Brothers with Jason confirmed.
Claimed vs found: Education inconsistent across his own documents (deck "BSEE Rutgers" vs bio "Rutgers + DeVry Engineering Technology"). "$50M/yr, 200-person operation": Harmonic shows DriveSmart Administration headcount 12. Not evidenced anywhere. Undisclosed: a second warranty brand — Nationwide Vehicle Assurance Corp d/b/a Motor Vehicle Assurance, same Toms River address as Apex United — and a concurrent third venture, Atenda.Care, where his own LinkedIn says he's led AI development "since January 2026" — while raising for AeVox.
Adverse: DriveSmart federal docket trail: 4 TCPA suits (2 OPEN: Connor D.N.J. 3:25-cv-12171; Negrete D.N.J. 3:26-cv-02994, filed 2026-03-23 — the same month AeVox went live at DriveSmart), 2 contract suits (1 OPEN: WalCo Funding, N.D. Ill. 1:24-cv-02117, naming Dan Rodd personally), LINE 5 (FL, on appeal 2026), Marshall v. Nationwide Vehicle Assurance (NJ, employment firm). DriveSmart BBB rating: D, not accredited. Reddit 2026: "They've withheld my refund for over a year."
Clean: no personal criminal / SEC / FINRA / CFPB hits. (The Georgia "Drivesmart Systems" Ch. 11 is an unrelated company.)
Verified: exists (Toms River); music sideline real; "Director of IT at Apex United" (RocketReach) fits the family-business pattern; AutoQuest CRM is real internal software.
Claimed vs found: only findable LinkedIn is a dead shell — headline "ceo at jroddcorp," 1 connection, zero history. Nothing externally verifies the NJIT BS CS 2018, the DriveSmart IT directorship 2018–23, or claimed patent authorship. "10+ years building Asterisk dialers" is arithmetically strained against a 2018 degree. "Independently created the first natural language-SQL query tool" is false as stated — NL-to-SQL long predates him; no product, repo, or press exists. Per the people agent: "the weakest verification profile I have ever pulled for a CTO of a company claiming proprietary voice models."
Adverse: Wells Fargo Financial Leasing v. AutoQuest Tech, Ocean County NJ Superior Court, OCN-L-000243-26 (2026) — his company sued by Wells Fargo's leasing arm.
Clean: no federal-court, SEC, FINRA, or CFPB hits under his name.
Verified: Harvard BA 2013 (History). Publicly lists the AeVox role. Horizon Space Acquisition I Corp board seat confirmed in EDGAR (2022-12-09 filings; Comp Committee chair + Audit Committee). Genuine paid-media / fractional-CMO career (Assembly Global, BMG, White Glove Elite).
Claimed vs found: bio's "NASDAQ: HSPOU" is stale and misleading — Horizon Space is a blank-check SPAC seeking a "Chinese Technology" combination that found no deal in 3.5+ years and voluntarily delisted from Nasdaq (announced Dec 3, 2025; now OTC as HSPUF). Key link: Singh worked at NY Capital Management Group — Angel Colon's firm. Singh and Colon are a pre-existing pair (NYCMG + the same SPAC board).
Adverse: none. Clean across all swept databases.
Verified: LinkedIn matches the bio exactly (Digital Marketing → BD Director → COO, DriveSmart Administration, current); Montclair State BA 2019; ~30.
Issue: no inflation — the problem is structural. He is simultaneously COO of the related-party "customer" anchoring AeVox's traction, and AeVox doesn't appear on his LinkedIn at all. Adverse: none. Clean.
Verified: Software Developer @ DriveSmart Auto Care May 2023–Aug 2024, BayMaster 6 mos, WGU BS Software Engineering, GitHub CydoEntis. Resume checks out; junior (~2 years). AeVox not on his profile. Adverse: none.
Verified (FINRA BrokerCheck CRD 2924711): Entoro Securities 8/8/2019–8/29/2025; TriPoint Global Equities 12/2013–7/2015 (the BANQ era); prior Bonwick, Tribal, Cuttone, Neuberger Berman; Series 7, 6, 63, 65, 24, SIE; zero disclosures — a real positive. EDGAR confirms directorships: Horizon Space (Forms 3/4), Sentage Holdings audit chair since 2021, NetClass Technology, Albert Origin Acquisition (2026 S-1/A).
Claimed vs found: "Co-founder of BANQ" is inflated — BANQ was TriPoint's platform (founded by Elenowitz/Boswell); Colon was a VP there for 19 months. "25+ years" spans an 11-year unregistered gap (2002–2013). Board portfolio is China-facing microcaps. His LinkedIn shows he advised Opiria/PDATA — Christian Lange's ICO-era venture, which explains Lange's presence in this room.
Risk flag: unregistered since 8/29/2025. If he solicits or places this $5M raise for compensation, that is unregistered-broker exposure (rescission rights against AeVox). No document states his role; the silence is the problem.
Verified: TUM Dipl.-Ing 2004, Dr.-Ing 2007; founder Ergoneers GmbH (real eye-tracking firm); VanillaWorks; Head of R&D Wacker Neuson, Ingolstadt (current). Claimed vs found: "2 exits" unverified — no acquisition of Ergoneers or VanillaWorks found. Route into the room is the Colon/Opiria connection, not voice AI. Adverse: none.
Verified: real Barbadian gospel artist (Barbados Today 2026-07-15). His LinkedIn shows what the data room hides: "Business Development Partner @ AeVox AI (5 mos)" — an actual role. Also "Promoter @ Equiturn" and "Business Consultant @ WealthVP" — he works startup capital-introduction circles. Not the inexplicable curiosity he first appears; a promoter node in the raise apparatus. Adverse: none found.
Verified: Co-founder & President-West, Mechanicus Insurance Services LLC (registered 2025-05-12 — 16 months old); COO of R H Mason Insurance 21+ yrs; EVP Sterling Risk Managers; CA/NV/SC licenses live. Claimed vs found: "CEO Spectrum Advisory Group and WestRidge Insurance Services" — neither on LinkedIn, unverified. "$100M in premiums saved" unverifiable. UCSB 1997–2003, no degree shown. Adverse: none.
Verified: ~40-year reinsurance career real — Guy Carpenter (decades), JLT Re, Towers Perrin (10 yrs), Sedgwick, JH Minet, Scan Re; now President & CEO of Mechanicus. Claimed vs found: "ex-partner Towers Perrin" vs LinkedIn "Vice President" — mild title inflation. Mechanicus is new, not an established platform. Adverse: none.
Verified: CRO Flagship Credit Acceptance ~5 yrs; Chief Risk & Credit Officer Carvant; ~18 years at Ally in residuals/valuations; GE; RVI. Real automotive-risk heavyweight. Claimed vs found: "Trained for his Ph.D. in economics at Boston College" = PhD candidate 1998–2000, left without the degree — phrasing engineered to imply more. Material omission: joined Black Book as VP/Head of Residual Value Strategy ~July 2026 — a full-time executive job taken weeks before this raise. His AeVox involvement is a name on a page, not bandwidth. Adverse: none.
Verified (partial): profile restricted, but Google's index of his LinkedIn reads "CRO at SoundWaveIQ"; Baker University, Overland Park KS; EMC chapter confirmed on an older profile. GE Medical 13 yrs and "two startup exits" unverified.
Conflict finding: SoundWaveIQ Inc. is one of AeVox's nine traction-register accounts (the stale demo lead). Its Chief Revenue Officer is on AeVox's team page. An "inbound lead" whose CRO sits on your own external team is not arm's-length pipeline — and the traction doc doesn't disclose the affiliation.
DebtHunch — the sole arm's-length signed customer — is InboundProspect, Inc. (Newport Beach CA), a Texas-regulated lender hit with TX OCCC Order L23-00091 ordering it to cease violating the Finance Code after failing to file its 2022 annual report; its BBB profile sits in Whitefish MT, matching the odd Yireh Business Trust (Montana) contracting entity. CFPB complaints: zero (clean). Its CEO Scott Couto is NOT the Columbia Threadneedle/Star Mountain Scott Couto — different person.
The people check kills the arm's-length illusion the traction page depends on. The operator core is real but is one family warranty complex: Dan, Jason, Lubas, and Stine all trace to DriveSmart — D BBB rating, five-plus federal suits including an open TCPA case and an open contract case naming Dan Rodd personally, a second undisclosed warranty brand at the same address, and tracked headcount (~12) that makes "200-person, $50M/yr" unsafe to accept. The CTO is unverifiable to a degree remarkable for a technical founder claiming proprietary models and four patents. The credentialed externals are real people with clean records, but none has a documented AeVox commitment, Fontaine just took a full-time job elsewhere, and the two externals who quietly DO have roles are undisclosed conflicts (Cobham = BD partner/promoter; Bokath = CRO of a listed pipeline account). The Singh–Colon–Cobham–Lange cluster is a pre-existing capital-markets network (NYCMG, the delisted Horizon Space SPAC, Opiria/PDATA), anchored by a securities professional whose FINRA registration ended a year before this raise. Nothing found rises to provable fraud; everything found says related-party traction, edge-inflated bios, live litigation in the core business, and a fundraising apparatus larger than the company itself.
| # | Room claim | External evidence | Verdict |
|---|---|---|---|
| 1 | "4 patents / 82 claims," patent-pending | Google Patents + USPTO assignment API: zero AeVox/Cubix filings; inventor "Jason Rodd" = 0. Their own live site footer says "Utility patent filed on Voice Memory™" — singular | CONTRADICTED by their own site. At most 1 unpublished filing; 4 granted patents is not plausible |
| 2 | SOC 2 Type II: window Q1 2027, report ~May 2027 (room disclosure) | Public site says "SOC 2 Type II audit underway · Q3 2026 target" | CONTRADICTED — site claims ~3 quarters ahead of what investors are told. The room version is the honest one; the site version is what customers see |
| 3 | Latency sub-400ms (brand/bios) vs <200ms (enterprise deck) | No independent benchmark exists | UNVERIFIED + internally inconsistent |
| 4 | DriveSmart live customer; "removed ~90% of $600–700K/mo call-center cost"; "$50M/yr, 200-person" | Harmonic: DriveSmart Administration Inc., founded 2016, headcount 12, $0 equity funding. No AeVox mention on dsacoverage.com. Savings claim also drifts 75–80% vs ~90% across docs | PARTIALLY CONTRADICTED — company real, "200-person" irreconcilable, savings unsupported. Related party regardless |
| 5 | Secure Car Care LIVE arm's-length logo (~$61K/yr modeled) | Real Studio City CA auto-warranty seller; zero AeVox or AI mention on their site; same vertical as DriveSmart | PARTIAL — company real; deployment unverifiable outside the room |
| 6 | DebtHunch signed 9/8, countersigned 9/9/26 ($0.20/min, $500/mo min, 90-day trial; modeled at 10× its ~2,000-call baseline = $192K) | debthunch.com live, no AeVox mention. Contracting entity Yireh Business Trust (Montana) — nothing findable. Signed one day before the room shipped | UNVERIFIED — real website, contract unverifiable, odd counterparty stands |
| 7 | Pipeline: ABAX ($7,500 committed, $279K modeled, MSA UNEXECUTED), Horizon Financial, BenMedix, CRS Credit, SoundWaveIQ, Se7en (Jamaican construction, off-vertical) | All 9 domains resolve to real sites; none mentions AeVox. SoundWaveIQ's CRO is on AeVox's team page | PARTIAL — companies exist; relationships unverified; one is a hidden affiliate |
| 8 | "Sourced entirely inbound," podcast presence | The two "podcast" links are AeVox's own YouTube @AeVoxAI: joined May 21 2026, 15 videos, 3 subscribers, 637 total views | CONTRADICTED in spirit — self-published to a dead channel; no third-party media exists |
| 9 | Trained on "400K hrs of proprietary DriveSmart calls" | Unverifiable. DriveSmart's calls are the subject of multiple TCPA consent suits — the training corpus itself may embody consent violations; data-rights/consent between related parties unaddressed | UNVERIFIED + legal-risk flag |
| 10 | Founded 2025, live since Mar 2026 | Domain created 2025-11-18; Harmonic first indexed 2026-05-22; zero Wayback snapshots; no press | PARTIAL — leaves ~4 months to a claimed production deployment of an entire proprietary voice stack |
| 11 | Traction headline: "$986K identified annual value" / $592K stage-weighted | Room's own register: ~$10.5K in actual contracted minimums. Measured billing evidence in total: 42 ABAX calls = $19.53 ($0.209/min); 361 production calls over 24 days = $252.90, 86.1% gross margin | ~94× inflation of contracted reality (the register itself discloses this — the deck headline doesn't) |
| 12 | Pro forma: $734K FY2026 → $55.9M FY2030, $72M exit ARR, EBITDA+ FY2028, 6→140 headcount, one $4M raise | FY2026 revenue modeled retroactively from Jan 2026 with nothing to reconcile against; raise mismatch vs deck ($5M @ $30M post); scenario tab's prose ($82.5M upside) contradicts its own table ($101.1M); "own GPUs" contradicted by pro forma's rented GPUs at $1.60/hr | INTERNALLY INCONSISTENT |
None named anywhere in the room. ~$500K founder money only. Harmonic (urn:harmonic:company:76082528, initialized 2026-05-22): $0 raised, 0 rounds, no investors, NONE_ANNOUNCED — consistent with the room; no hidden prior rounds. No SEC filings mention Aevox (EDGAR full-text, 2026-09-10).
The $5M-at-$30M-post ask vs the pro forma's $4M Nov-2026 close is an internal contradiction with no external referent. The presence of a broker-dealer veteran (Colon, unregistered since 8/29/2025), a SPAC audit-committee pair (Colon + Singh), an insurance MGA duo (Mechanicus), and a capital-intro promoter (Cobham) around a 1-month-old entity with no signed roles looks like a placement/channel scaffold assembled for the raise, not an operating team. Aside: Harmonic shows DriveSmart with a $2M EDA grant (Mar 2024) as its only "funding" — likely a data quirk, but worth one question.
| Case | Court | Type | Status |
|---|---|---|---|
| Brown v. DriveSmart, 1:20-cv-05744 | S.D.N.Y. (2020) | TCPA | closed |
| Advanced Marketing & Processing v. DriveSmart, 8:21-cv-02477 | M.D. Fla. (2021) | Contract | closed |
| Callier v. DriveSmart, 3:23-cv-00162 | W.D. Tex. (2023) | TCPA | closed |
| WalCo Funding v. DriveSmart + Daniel J. Rodd personally + Nationwide Vehicle Assurance, 1:24-cv-02117 | N.D. Ill. (2024) | Contract | OPEN |
| Connor v. DriveSmart, 1:24-cv-08413 | N.D. Ill. (2024) | TCPA class action | terminated 1/2025 |
| Connor v. DriveSmart, 3:25-cv-12171 (refiled) | D.N.J. (6/2025) | TCPA | OPEN |
| Negrete v. DriveSmart, 3:26-cv-02994 | D.N.J. (filed 2026-03-23) | TCPA | OPEN — filed the same month AeVox's "autonomous outbound sales" went live at DriveSmart; worth pulling the complaint to see if AI-placed calls are at issue |
| Company | Raised | Stage / note |
|---|---|---|
| Sierra | $1.70B | Series E 5/2026, headcount 1,028 |
| Parloa | $567M | Series D |
| Decagon | $481M | Series D |
| PolyAI | $204M | Series D |
| Replicant | $113M | |
| Bland | $112M | Series C 6/2026, hc 126 |
| Regal | $82M | |
| Vapi | $78M | Series B 5/2026, hc 202 |
| Crescendo | $50M | |
| Retell | $4.6M | Seed, YC, hc 19 |
Aevox's claimed differentiators (cross-call memory, emotion detection, human supervision, in-call actions) are shipping features at Sierra, Parloa, Decagon and Bland — none is a moat — and "own models + owned GPUs" is contradicted by their own pro forma (rented GPUs at $1.60/hr). The narrow real edge: operators who own a call center in the TCPA-heavy vertical the funded players avoid. But a $30M post for ~$10.5K of contracted arm's-length minimums against competitors with 100–1,000× the capital is not a price, it's a hope. On this field, Retell raised $4.6M at seed with YC and 19 people; Aevox is asking 6× that on a related-party logo and a one-day-old contract.
PASS. Company 5.0/10 (Base 3.00 + EV 1.00 + Geo 1.00 + Chain 0). FOR DAXOS 2.5/10 — hard pass at $5M / $30M post.
There is a real kernel: two call-center operators automating their own compliance-heavy vertical, a live product with measured (tiny) billing at 86.1% gross margin, and a traction document that is more honest than its own deck. But the investable story dissolves on contact: ~$10.5K of contracted arm's-length minimums against a $986K headline and a $30M-post ask; every load-bearing element (customer, data, staff) tracing to a D-rated, serially-sued related party; claims contradicted by the company's own website; an unverifiable CTO; no corporate hygiene; and a raise apparatus with unregistered-broker rescission risk. Nothing found rises to provable fraud — and nothing found supports the price.
What would reopen it (all, not some): clean single entity with cap table and IP assignment from DriveSmart/Cubix; the actual patent filing(s); 3+ arm's-length logos billing at contract for 2 consecutive quarters; written confirmation no compensated unregistered finder touches the raise; Negrete/Connor complaints pulled and shown not to implicate AI-placed calls; disclosure of the Bokath and Cobham affiliations; and a cap in the single digits. Even then this is a micro-check relationship bet on the Rodds' distribution, not a $30M-post company.