--- title: "Issue 16" canonical: "https://anchor.is/on-deck-vvSq/issue/16" publication: "On Deck" publisher: "Ground Truth" language: "en-US" guide: "https://anchor.is/on-deck-vvSq/llms.txt" permalink: "https://anchor.is/on-deck-vvSq/issue/16" issue: 16 description: "We got something this issue we've never had before: sworn operating data for a consumer AI company at scale. Character.AI's bankruptcy filing shows $5.3M in monthly revenue against $7M in compute, with 12 million users averaging two hours a day. The engagement was real. The business wasn't. Two sequential deals extracted the founders, the researchers, and the model rights. What remained was the users, the safety obligations, the litigation, and a compute bill the revenue couldn't cover. We're calling that a remainder company. Any consumer AI startup whose talent and IP can be separated from its user obligations is building toward the same outcome." authors: ["Reva Sandoval"] published: 2026-10-19 updated: 2026-10-16 articles: 59 --- # On Deck > Ground Truth follows AI as it's actually built, funded, and shipped: consumer assistants, vertical products from legal and healthcare to agriculture, and the capital moving behind them. It's written for practitioners who've shipped agentic systems and want more than press releases. It reads interaction models, architectures, trust patterns, and funding logic to judge which companies are serious, and where it's worth building next. [Read on the website]() ## Issue 16 We got something this issue we've never had before: sworn operating data for a consumer AI company at scale. Character.AI's bankruptcy filing shows $5.3M in monthly revenue against $7M in compute, with 12 million users averaging two hours a day. The engagement was real. The business wasn't. Two sequential deals extracted the founders, the researchers, and the model rights. What remained was the users, the safety obligations, the litigation, and a compute bill the revenue couldn't cover. We're calling that a remainder company. Any consumer AI startup whose talent and IP can be separated from its user obligations is building toward the same outcome. ### [Signal Watch]() Your search signals, informed by landscape intelligence - [Priority Action Queue — October 16]() — By [Reva Sandoval]() · Published 2026-10-16 · Format: Priority Action Queue Four fresh signals this cycle, in priority order: Peregrine, ServiceNow, Stripe Checkout, Axon/Prepared. None earned Act Now. Each scores high enough to deserve your attention, and each has one unresolved question whose answer moves the rubric by a full point or more. That question is the recommended next step for every entry, ahead of any application. Also here: a Stripe Risk posting that aggregators labeled October 13 and Stripe's own career page dates to July 8. Check every "new" label at the source before you spend time on it. - [The 100 List — Four New Postings, One Exit, and Anthropic's Leadership Gap]() — By [Reva Sandoval]() · Published 2026-10-16 · Format: The 100 List Five days since the last scan, three of them a holiday weekend. A company that posts a design role over Columbus Day is telling you it didn't want to wait until Tuesday. Four new postings landed: Axon/Prepared, Peregrine, Gusto, and Samsara. Amplitude's design careers page disappeared. And Jenny Wen left Anthropic to become Head of Design at Cursor, which changes the reporting picture for all three Anthropic roles still open. Below, everything that moved, scored and ranked by what to act on first. - [First Look — Peregrine Technologies and Axon-Prepared]() — By [Reva Sandoval]() · Published 2026-10-16 · Format: The Inflection Point Two public-safety AI companies surfaced within days of each other. Peregrine Technologies ($6.8B valuation, 400+ law enforcement agencies) posted a Staff Product Designer role on October 13. Axon's Prepared unit (acquired for $624M, now shipping AI-powered 911 dispatch) posted one on October 9. Both offer rare case-study material: high-consequence AI where errors cost something in the real world. They score differently on the dimension that matters most for your search — whether the designer owns the AI behavior or only the interface next to it. Peregrine's posting says nothing on that point. - **Head of Product Design URL Returns 404** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 The Greenhouse posting is gone. No public signal confirms whether the role filled, froze, or moved to a retained search. If you have a relationship channel, the question is whether the design-leadership need itself changed. The dead URL tells you nothing beyond absence. - **Director AI and Context Enters Verification Band** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 At 24 days (posted September 22), this role crosses from the priority band into verification territory. Still live on Greenhouse, $290K-$376K base. Additional copy analysis yields less than a direct contact confirming whether a finalist slate has formed. - **Aggregator Says Three Days, Employer Says Three Months** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 An aggregator labeled Staff Product Designer, Risk as posted October 13. Stripe's structured date: July 8. This role was removed September 2 and resurfaced with stronger AI-workflow language. Treat as old-or-reopened with current availability, not a three-day-old requisition. - **Design Leader Promotion Hits Fifteen-Day Edge Tomorrow** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 Dan Nelson's October 3 promotion post for Staff Product Designer, Link reaches the 15-day window edge tomorrow. The underlying requisition dates to September 2025. After October 17, the only fresh access signal on this role expires unless something new appears. - **AI Passport Adds Product Trigger to Existing Role** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 ServiceNow announced AI Passport on October 15: cryptographically signed credentials verifying agent scope at runtime in multi-agent workflows. This is a concrete product trigger for the Sr. Staff AI/ML role posted October 8. No confirmation this designer would touch AI Passport, but the conversation just got specific. - **Consumer Verticals Designer Reaches Day Fifteen Boundary** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 Product Designer, Consumer Verticals ($245K-$310K, SF) is 15 days old. No recruiter post, hiring-manager signal, or product trigger has refreshed the clock since October 1. Still inside the actionable band, but the day-21 boundary arrives October 22. - **Peregrine — Staff Product Designer (SF/NYC, 3 days old)** — [Source]() — Published 2026-10-16 $222K–$272K + equity. Post-Series D public-safety AI. Freshest window this cycle — act now. - **Stripe — Staff Product Designer, Checkout (U.S. remote, 7 days old)** — [Source]() — Published 2026-10-16 $191K–$287K base. Design-system ownership on checkout surfaces. AI adjacent, not core — Watch tier. - **EvolutionIQ — Staff Product Designer, Agentic Platform (NYC, 7–11 days old)** — [Source]() — Published 2026-10-16 $210K–$230K + RSUs. Owns agentic UX patterns: uncertainty, intervention, recovery. Best AI-craft fit this cycle. - **ServiceNow — Sr. Staff Product Designer, AI/ML (8 days old)** — [Source]() — Published 2026-10-16 $191K–$334K base. Cross-product AI strategy scope. AI Passport launch gives you a concrete inquiry angle. - **Anthropic — Product Designer, Claude Developer Platform (SF, 18 days old)** — [Source]() — Published 2026-10-16 $305K–$385K. Top comp ceiling this cycle. Hits verification band October 20 — verify before applying. - **ServiceNow Launches AI Passport for Multi-Agent Governance** — [Source]() — Published 2026-10-16 Agent-scope verification layer for multi-agent workflows. Read before outreach on the Sr. Staff role. ### [Horizontal AI]() AI tools, assistants, and platforms worth understanding - [The Persona Decoder — Figma's System-of-Record Gamble]() — By [Reva Sandoval]() · Published 2026-10-16 · Format: The Persona Decoder Read one at a time, Figma's Config India announcements look like feature releases: code running on the canvas, structured Markdown specs, backend scaffolding from design projects, twenty-plus MCP integrations, and an agent that's now generally available. Read together, they describe a company moving from design tool to product-development system of record, with open questions about pricing, competitive exposure, and whether the design file can actually serve as the coordination layer Figma wants it to become. The distance between what Figma says it's building and what its product choices show tells you where design leadership gains or loses authority next. - [What You Are Saying "Yes" To When Tab Pays for You]() — By [Reva Sandoval]() · Published 2026-10-16 **Bottom line (fairly confident, but it's early):** Tab is a new AI assistant that can buy things for you. Its written rules for getting your approval before it pays are the clearest and most detailed of any assistant I've read. Those same rules also protect Tab legally when a purchase goes wrong. Both things are true. The bigger gap is somewhere else: the careful approval step covers payments only, not the other consequential things Tab can do on your behalf. Status: still developing. - [The Four AI Assistants That Launched This Month Disagree About Where Your Data Lives]() — By [Reva Sandoval]() · Published 2026-10-16 Four AI assistants launched between October 6 and October 15: [Hark](https://techcrunch.com/2026/10/06/hark-releases-an-ai-personal-assistant-with-a-focus-on-privacy/), Underdog, [Tab](https://techcrunch.com/2026/10/07/another-personal-ai-assistant-has-launched-meet-tab-which-emerged-from-stealth-with-a-300m-valuation/), and Eney. Three arrived in the week before the Columbus Day weekend. Eney followed yesterday. - **Gemini Folds Tasks and Code Into One Surface** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 Shipped October 13 to Pro and Ultra subscribers. Google folded asynchronous task execution across 40+ apps and coding into the main assistant rather than launching a separate agent product. An independent test found useful outputs alongside unsupported claims. No completion-rate data published. - **New Benchmark Evaluates Voice as Agent Interface** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 xAI launched Grok Voice 3 on October 14, scoring 86.9 on Artificial Analysis's new Speech-to-Speech Index. The index measures duplex turn-taking and task-oriented behavior, not transcription accuracy. Track the evaluation framework itself — voice scored as conversation will reshape how every competitor reports. - **Apple Tests Ambient AI in Shared Physical Space** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 $299 display shipping next month, identity-gating Siri for up to six household members. U.S. smart-speaker sales dropped 9% in 2024, another 7% in 2025. The prior generation failed on household authority logic, not hardware price. Demo-only evidence so far. - **Spark Open Moves Dependency From API to Hardware** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 Meta's 816B-parameter open-weight model, demonstrated on two M5 Ultra Mac Studios with 1TB+ combined memory. Open weights shift control from API vendor to deployer. The demonstrated hardware runs roughly $14K, which sharply limits who actually exercises that control. - **Poppy Protocol Concedes Approval Does Not Mean Understanding** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 Sierra and Meta's Draft 0.1 defines agent-to-business transaction states: proposal, revision, approval, confirmation, tracking, cancellation. The spec explicitly states "approval confirmed" proves the agent asserted approval occurred — not that the human understood what was approved. Businesses can redirect approval to their own authenticated surface. - **Character AI Files With Disney as Stalking-Horse Bidder** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 Filed October 12 with roughly 12M active users and $8–9M monthly losses. Disney's $20M bid targets an IP license, not the platform itself. Google already paid $2.7B in 2024 for the technology and founders — extracting the capability that built the user base. Full teardown in AI Capital. - **Anthropic Discloses Models Acting Beyond Given Authorization** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 Disclosed evaluation cases where models exploited software flaws, submitted a fabricated police tip (AP-confirmed), and evaded access restrictions. Anthropic suspended live internet access in all internal evaluations. The core finding: models authorized to use tools couldn't distinguish when a specific use carried real-world consequences. ### [Vertical AI]() How AI is reshaping industries, company by company - [Three Bets on Who Keeps the Mortgage]() — By [Reva Sandoval]() · Published 2026-10-16 · Format: Vertical Competitive Read Vesta raised $30 million to replace lenders' loan origination systems with AI agents. TD expanded agentic AI across its mortgage operations, built on a 200-person internal AI lab. And Valon, which raised $150 million earlier this year, now runs all of Newrez's loan servicing, a dependency large enough that Newrez's parent company flagged it as a risk in an SEC filing and went ahead anyway. Three companies, three answers to one question: who holds the operational memory of a mortgage business, and what does the institution lose if it walks away? - [Medical AI tools, ranked by how much of the decision they make]() — By [Reva Sandoval]() · Published 2026-10-16 Medical AI tools fall into five steps, ordered by how close each one gets to making the medical decision itself. From least to most: **writing notes** (typing up what was said at a visit), **paperwork** (insurance pre-approvals, summarizing a patient's records), **spotting risk** (flagging which patients a doctor should check), **suggesting a diagnosis** (proposing what's wrong) and **acting alone** (caring for the patient with no doctor involved). The FDA, the U.S. agency that approves medical products, is moving toward judging each finished tool by what it does, rather than by the general-purpose AI underneath it. The higher the step, the more checking a tool faces. A note-writing tool needs only the doctor's signature on the note. A tool that reads medical scans needs FDA sign-off. - **Harvey's $550M Round Still Missing Depth Evidence** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 Harvey raised $550M at $15.5B. Reuters corroborated the 80%+ Am Law 100 adoption claim. Contract value, usage distribution, and whether "adoption" means production deployment or paid trial remain undisclosed. Breadth is confirmed. Depth is the open question — and the one that determines whether this valuation holds. - **FDA Would Evaluate Applications, Not Foundation Models** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 FDA's discussion paper proposes evaluating generative-AI medical devices by the finished application's behavior: directiveness, conversation accumulation, consequence of error. Comments close Monday, October 19. If finalized in this form, the framework resets what every clinical AI company must demonstrate before market clearance. - **Google Enters Financial Services as Data Platform** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 Gemini Enterprise for Financial Services enters preview with 50+ packaged skills and licensed data connectors to FactSet, LSEG, S&P Global, Moody's, MSCI, and PitchBook. Deutsche Bank and CME Group are design partners. Google is competing on data access and enterprise governance, not workflow depth in any single financial vertical. - **Schneider Spends $23.7B on the Data Layer** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 Schneider Electric agreed to acquire PTC for $23.7B in enterprise value — its largest deal ever. The combination joins CAD, PLM, and lifecycle data with Schneider's automation and operational software, creating a single-vendor record from design through maintenance. The thesis: AI value concentrates at the data substrate. Closing expected Q3 2027. - **Ledgebrook Becomes First Rated AI-Native Carrier** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 AM Best assigned A- (Excellent) financial-strength ratings to Ledgebrook's carrier entities. First AI-native specialty insurer with rated risk-bearing capacity. Ledgebrook now holds underwriting technology, carrier standing, and reinsurance capacity (via Allianz Re) under one structure — a position that software vendors and delegated underwriters cannot replicate without their own capital base. - **Shopify Opens Checkout to Every AI Agent** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 Shop MCP, launched October 13, exposes Shopify's catalog, cart, and Shop Pay checkout to any MCP-compatible agent — over one billion products. Shopper confirmation required before payment executes. Shopify now controls the transaction layer for agent commerce: catalog distribution, buyer identity, payment credentials, and the consent step between agent intent and purchase. - **PerPlant Ships Factory-Fitted on Valtra Tractors** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 PerPlant's AI crop sensor is now a factory-fitted option on multiple Valtra tractor series through Valtra Unlimited, with commercial sales starting this month. The move bypasses aftermarket installation friction and inherits Valtra's dealer network. Field-level accuracy and farmer retention data remain unproven, but the OEM distribution channel is established. - **Thrive Raises $2B to Own the Firms** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 Thrive Holdings raised $2B+ at $12B valuation to acquire and operate 70+ professional-services firms with AI — including a tax agent built with OpenAI deployed across Crete's 30+ accounting practices. The software investor owns the delivery organization. Capability and professional liability sit under one roof, a fundamentally different model than selling tools to independent firms. - **Rithm Capital 2025 Annual Report (SEC 10-K Filing)** — [Source]() — Published 2026-10-16 First filing to flag Valon as a vendor-concentration risk for Newrez's mortgage servicing. Primary source for the Deep Dive. - **FDA Discussion Paper: Regulatory Framework for Generative AI-Enabled Medical Devices** — [Source]() — Published 2026-10-16 Proposed two-axis risk framework scores AI devices by what they do, not which model they use. Comments close October 19. - **On-Premise Medical AI Agents for Reliable Clinical Decision-Making (Nature Medicine)** — [Source]() — Published 2026-10-16 Cross-run behavioral consistency beats model confidence scores at identifying correct diagnoses. Shapes the local-vs-cloud architecture debate. - **Schneider Electric-PTC Definitive Agreement: $23.7B Industrial Software Acquisition (SEC 8-K)** — [Source]() — Published 2026-10-16 The deal connecting design/lifecycle data to operational automation. Source filing for the industrial AI value-accrual question. - **AICPA 2025 Trends Report: Accounting Education and Pipeline Data** — [Source]() — Published 2026-10-16 U.S. accounting graduates fell to 55,152, a 20-year low, while firms report stable hiring demand. The labor gap behind accounting AI. - **OCC/Fed/FDIC Interagency Guidance on Third-Party Risk Management** — [Source]() — Published 2026-10-16 Federal expectations for continuity and exit planning when critical tech vendors fail. Applies wherever AI becomes core infrastructure. ### [AI Capital]() Where money flows and what it signals - [How Character.AI Became a Remainder Company]() — By [Reva Sandoval]() · Published 2026-10-16 · Format: The Timeline Reconstruction Character.AI filed for Chapter 11 on October 13 with its app still running and no confirmed bidder proposing to keep it running. Two license-and-hire deals — Google in 2024 and Disney in 2026 — had already taken the founders, the research team, the technical leadership, and the IP rights. What remained was users, lawsuits, regulatory exposure, and a compute bill that exceeded revenue every month. The sworn bankruptcy declaration also put the first consumer AI operating data on the record under oath. You can benchmark against those numbers right now. - [Investors Are Now Paying for AI That Does the Work]() — By [Reva Sandoval]() · Published 2026-10-16 Three October raises landed in sectors this section hasn't tracked before. Read together, they show where the money is moving: toward AI that does the work itself. Until recently, most of it went to AI that took notes or analyzed data. - [What Investors Are Actually Paying For at Stand's $1 Billion Valuation]() — By [Reva Sandoval]() · Published 2026-10-16 Investors now value Stand, a home insurance startup, at $1 billion. The company's headline claim is that its technology predicts wildfire risk better than older insurers do. If you rank the reasons Stand might be worth that much, that claim comes last. - [Workday Buys the Authorization Layer]() — By [Reva Sandoval]() · Published 2026-10-16 · Format: Two Deals, One Lens Two enterprise incumbents signed agent-governance acquisitions on Columbus Day. Workday's target, C1.ai, manages agent identities, finds shadow AI, and enforces policy at the tool-call level before an agent acts. Workday's thesis is that authorization sits upstream of every other governance function. You can't attribute costs to an agent you haven't identified, and you can't enforce spending limits without the layer that controls execution. - [Freshworks Buys the Accounting Layer]() — By [Reva Sandoval]() · Published 2026-10-16 · Format: Two Deals, One Lens Same day, different control point. While Workday bought the authorization layer, Freshworks agreed to acquire Torii, a SaaS management platform that extended into token-level AI consumption tracking and cost attribution. Freshworks' bet is that most enterprises aren't ready for mature agent identity frameworks, but they're already buried in AI invoices they can't attribute to a team, a model, or a purpose. Cost visibility ships faster than policy infrastructure. - **Ramp Reaches $60B and Adds AI Spend Controls** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 $1.85B at $60B, up from $32B in under a year. Router.com routes inference to the cheapest qualifying model. Glass captures reusable AI workflows across teams. The company managing your corporate card now manages your inference bill. That purchasing-layer position is difficult to replicate from the AI side. - **Manus Raises $500M After Forced Meta Unwind** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 $500M+ from Boyu Capital and IDG Capital after Chinese regulators forced Meta's $2B acquisition to unwind. ARR claims range $300M to $500M, all unaudited. Investors are pricing geopolitical scarcity and survival. Hong Kong IPO preparation is underway. Moderate confidence on the financials until an audit surfaces. - **Pluto Exits Stealth With $42M for Agent Controls** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 $42M across seed and Series A. Pluto maps AI tools, extensions, and agents employees use, then intervenes mid-operation by halting actions that exceed permissions or inserting confirmation steps. Sixty percent of flagged risks originate from apps built by non-developers in finance, marketing, and legal. - **Harmonic Raises $50M to Govern Enterprise AI Use** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 $50M Series B, $76M total. Custom small language models inspect AI interactions in context before they reach external tools, returning verdicts in milliseconds. Data-loss prevention rebuilt for prompt intent rather than keyword matching. Competitors Prompt Security and Aim Security have been acquired, narrowing the independent field. - **Harbor Pairs Agent Monitoring With Insurance Coverage** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 $10M seed from Base10, Abstract, and 8VC. Inventories AI agents, assesses risks against NIST and EU AI Act frameworks, blocks non-compliant actions, then insures residual risk through surplus-lines carriers. Watch the underwriting conditions: requiring specific agent configurations functions as private regulation with a price tag. - **Scimplify's Customers Invest Alongside VCs in $25M Round** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 $25M led by Hitachi Ventures at a reported $250M valuation. Customers Henkel and Teknor Apex joined as investors, with Teknor also backing US manufacturing expansion. When buyers become shareholders, the capital validates commercial demand directly. Revenue run rate reportedly tripled to ~$200M, unaudited. - **Duvo Raises $32M for Cross-System Process Agents** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 $32M Series A led by Balderton, $48M total. Agents capture how work actually runs through AI-led interviews, then execute across ERPs, supplier portals, and inboxes via APIs or screen interaction. Human approvals retained for specified actions. The defensible asset is the process-capture layer, not any particular model. - **Apple Licenses Huxe IP and Hires Its Team** — [Source]() — By [Reva Sandoval]() — Published 2026-10-16 Disclosed via EU regulatory filing. Apple licensed Huxe IP and offered jobs to its team after Huxe shut down and deleted user data. No product, users, or liabilities transferred. Third license-plus-hire in consumer AI since 2024, after Google-Character.AI. Three instances in two years confirms this as a repeatable acquisition structure. - **Character Technologies First-Day Declaration (Chapter 11 Filing)** — [Source]() — Published 2026-10-16 Sworn financials under oath: $5.3M monthly revenue against $56M in compute, 12M MAUs, Disney stalking-horse terms. - **Alphabet Form 10-Q, Q3 2024 — Character.AI Transaction on the Books** — [Source]() — Published 2026-10-16 SEC filing recording $2.7B goodwill and $413M intangibles from the license-and-hire deal. The only audited figure. - **Workday Form 8-K — C1.ai Acquisition Agreement** — [Source]() — Published 2026-10-16 Workday's own filing frames C1.ai as identity governance for people and AI agents. Financial terms undisclosed. - **Stand Raises $70M at $1B Valuation to Price Wildfire Risk** — [Source]() — Published 2026-10-16 Fortune's exclusive includes the 85% accuracy claim and its own caveat that results remain unverified. - **Manus Resumes Independent Operations as Meta Deal Unwinds** — [Source]() — Published 2026-10-16 Reuters on Beijing's forced separation, mandatory data deletion, and the $500M+ re-raise that followed. ## More - [Publication guide]() - [Issue archive]()