# Source register — mid-2026 field report

Verified July 31, 2026. Classification: **P** = primary (official/original: legislation, tribunal decisions, company announcements about themselves, original research reports, official pricing/docs, earnings materials) · **S** = secondary (journalism, analyst commentary, third-party teardowns, estimates). Vendor-authored/funded flagged (VF).

| # | Source | Org / type | Date | Class | VF | Role / claims supported |
|---|---|---|---|---|---|---|
| 1 | Salesforce–Fin definitive agreement PR | Salesforce | 2026-06-15 | P | VF | $3.6B deal, signed-not-closed, Fin 76% vendor claim, Agentforce $1.2B ARR | 
| 2 | Salesforce Q1 FY27 earnings PR | Salesforce | 2026-05-27 | P | VF | Agentforce $1.2B ARR +205% (term "ARR" confirmed) |
| 3 | "Today Intercom becomes Fin" | Intercom/Fin blog | 2026-05-14 | P | VF | Rebrand date; Fin >$100M ARR (company-reported) |
| 4 | Klaviyo AI agents PR | Klaviyo/BusinessWire | 2026-06-30 | P | VF | Customer Agent GA; ~196,000 customers; Composer beta |
| 5 | Shopify Inbox App Store listing | Shopify | checked 2026-07-30 | P | VF | Free AI sales associate in Inbox |
| 6 | Introducing OpenAI Presence | OpenAI | 2026-07-22 | P | VF | Presence launch, design partners, FDE deployment |
| 7 | The Register on Presence | The Register | 2026-07-22 | S | — | "Boots-on-the-ground prices" characterization |
| 8 | Sierra $950M round | TechCrunch | 2026-05-04 | S | — | Round, post >$15B, ARR trajectory company-reported |
| 9 | Sierra $100M ARR | TechCrunch | 2025-11-21 | S | — | ARR milestone |
| 10 | Sierra $15.8B reporting | CNBC / Yahoo Finance | 2026-05-04 | S | — | $15.8B post figure |
| 11 | Decagon $4.5B Series D | Bloomberg | 2026-01-28 | S | — | Valuation, round |
| 12 | Decagon revenue investigation | Forbes | 2025-12-01 / 2026-02-06 | S | — | Independent ~$12M estimate vs company $30–35M |
| 13 | Decagon pricing blog | Decagon | checked 2026-07-30 | P | VF | Vendor's own "gray areas" admission; two meters |
| 14 | Parloa $350M Series D | TechCrunch + Parloa PR | 2026-01-15 | S/P | VF(PR) | $3B valuation |
| 15 | Parloa outcome-pricing essay | Forbes BrandVoice | 2026-01-06 | S | VF | Anti-outcome manifesto (sponsored, labeled) |
| 16 | Wonderful $150M Series B | TechCrunch | 2026-03-12 | S | — | $2B valuation; metrics company-reported |
| 17 | Wonderful seed/Series A | Index Ventures / TechCrunch | 2025-07 / 2025-11 | P/S | VF(Index) | Tracker rows |
| 18 | PolyAI $86M Series D | Forbes / PolyAI blog | 2025-12-15 | S/P | VF(blog) | $750M valuation |
| 19 | PolyAI platform opening PR | PRNewswire (PolyAI) | 2026-05-18 | P | VF | Self-serve opening, free 2 months |
| 20 | Redo Series B PR | GlobeNewswire (Redo) | 2026-06-23 | P | VF | $81M @ $1.25B; ReturnBear |
| 21 | Redo pricing page | redo.com | checked 2026-07-30 | P | VF | $0.85/resolution inside free returns platform |
| 22 | Hugo pricing/billing help article | Crisp | checked 2026-06-22 + 2026-07-13 | P | VF | $0.05–0.10 effective per conversation |
| 23 | Zipchat pricing | zipchat.ai | checked 2026-07-15 | P | VF | $0.20/reply |
| 24 | HubSpot Breeze pricing change | SaaStr | 2026-04-09 | S | — | $0.50/resolution; 65% vendor claim; Agentforce ~$2/conv |
| 25 | HubSpot pricing-change explainer | Resolve247 | 2026-04 | S | — | Effective date, resolution definition |
| 26 | Fin pricing guide | Gleap | 2026 | S | — | $0.99/resolution current |
| 27 | Gorgias pricing analysis | Chatarmin | updated 2026-06-09 | S | — | $0.90/1.00 + ticket fee |
| 28 | Zendesk pricing analysis | Richpanel | 2026 | S | — | ~$1.50–2.00 estimates (no official list) |
| 29 | Sierra pricing teardown | OpenNash | 2026 | S | — | $1.00–2.50, platform minimums |
| 30 | Crescendo $100M ARR | AI Magazine/GlobeNewswire | 2025-09 | P | VF | ARR milestone (company release) |
| 31 | Crescendo–PartnerHero | Crescendo news | 2024/2025 | P | VF | BPO acquisition; $2.99 model |
| 32 | Salesforce Agentforce flex pricing PR | Salesforce | 2025-05-15 | P | VF | $0.10/action credits |
| 33 | 11x.ai investigation | TechCrunch | 2025-03-24 | S | — | Logos, ARR booking, churn accounts |
| 34 | Zendesk–Forethought | TechCrunch + Zendesk newsroom | 2026-03-11 + close | S/P | VF(newsroom) | Acquisition |
| 35 | Zendesk take-private completion | Zendesk newsroom / Business Wire | 2022-11-22 | P | VF | Private since Nov 2022 ($10.2B) |
| 36 | Zendesk AI ARR reporting | The AI Economy / The Letter Two | 2025-12 | S | — | $200M AI ARR, $400–500M target (to-press) |
| 37 | NICE–Cognigy | press coverage | 2026 | S | — | $955M, ~25x reported |
| 38 | Gartner 14% self-service PR | Gartner | 2024-08-19 | P | — | 14% full self-service resolution; n=5,728; Dec 2023 field |
| 39 | Gartner 64% AI-preference PR | Gartner | 2024-07-09 | P | — | 64% prefer no AI; 53% would consider switching |
| 40 | Gartner 40% agentic cancellation PR | Gartner | 2025-06-25 | P | — | Prediction |
| 41 | Gartner 80%-by-2029 PR | Gartner | 2025-03-05 | P | — | Prediction |
| 42 | Forrester 2026 predictions PR | Forrester | 2025-10-28 | P | — | One-third trust erosion prediction |
| 43 | Intercom Customer Service Transformation Report 2026 | Intercom | 2026-01-28 | P | VF | n=2,400+; 87%/10% deployment gap |
| 44 | Hiver State of AI Customer Support 2026 | Hiver | 2026-03-20 | P | VF | n=700+ US leaders; 14%; ~50% |
| 45 | ICONIQ 2026 State of AI snapshot | ICONIQ | 2026-01 | P | — | Margin projections 52%/actuals 45/41 |
| 46 | SaaStr on ICONIQ | SaaStr | 2026-01-29 | S | — | Summary of #45 |
| 47 | MIT NANDA "GenAI Divide" (mirror PDF) | MIT NANDA project | 2025-08 | P | — | 300+/52/153 methodology; 95%; 40%/5%; 67%/33%; limitations |
| 48 | NANDA critique | Futuriom | 2025-08 | S | — | Derivation critique context |
| 49 | Oldroyd lead-response study PDF | MIT Sloan/InsideSales | 2007 | P | — | Original contact-odds metric |
| 50 | Optifai response-time benchmark | Optifai | 2026 | S | VF | 2.6x close-rate observational benchmark |
| 51 | Klarna AI assistant PR | Klarna | 2024-02-27 | P | VF | 2.3M conversations, "700 agents," $40M projection |
| 52 | Klarna reversal coverage | Forbes | 2025-05-18 | S | — | CEO quality quotes, rehiring |
| 53 | Klarna 853-FTE earnings remark | CX Dive | 2025-11-20 | S | — | Q3 2025 earnings call figure |
| 54 | Klarna headcount 7,400→3,000 | Fortune | 2025-10-10 | S | — | Total-company, attrition-driven |
| 55 | Rollback roundup (Ford/CBA/IBM) | Invezz | 2026-07-01 | S | — | Case file (cross-domain labeled) |
| 56 | Salesforce support headcount | Fortune | 2025-09-02 | S | — | 9,000→~5,000 with redeployments |
| 57 | Moffatt v. Air Canada, 2024 BCCRT 149 | BC Civil Resolution Tribunal | 2024-02-14 | P | — | Negligent misrep; CA$812.02; tribunal status |
| 58 | Cursor bot incident | The Register | 2025-04-18 | S | — | Hallucinated policy case |
| 59 | EU AI Act (Reg (EU) 2024/1689), Art. 50 & 99 | EUR-Lex / AI Act explorer | in force | P | — | Transparency duties; penalties |
| 60 | Digital Omnibus (Reg (EU) 2026/1744) | EUR-Lex / OJ | 2026-07-24 | P | — | Deferrals; marking grace period |
| 61 | EC transparency guidelines | European Commission | adopted 2026-07-20 | P | — | Final Art. 50 guidance |
| 62 | Digital Omnibus analysis | Lewis Silkin; Jones Walker | 2026-07 | S | — | Commentary confirming dates |
| 63 | California SB 243 text | LegiScan (statute text) | 2025-10-13 signed | P | — | Companion-bot scope; CS exclusion; $1,000 PRA |
| 64 | NY GBS §1700/§1702 | NY Senate statute pages | eff. 2025-11 | P | — | AI-companion scope; CS exclusion; 3-hour rule |
| 65 | California B.O.T. Act §17941 | statute (via FindLaw) | 2019 | P | — | Intent-to-mislead scope |
| 66 | a16z LLMflation | a16z | 2024–2026 series | S | — | Inference deflation series |
| 67 | Epoch AI inference price trends | Epoch AI | ongoing | P | — | Price-per-capability data |
| 68 | OpenAI API pricing page | OpenAI | checked 2026-07-31 | P | VF | GPT-5.4-mini $0.375/$2.25 |
| 69 | Anthropic API pricing page | Anthropic | checked 2026-07-31 | P | VF | Haiku 4.5 $1/$5 |
| 70 | Freshworks market data | stockanalysis.com | fetched 2026-07-31 | S | — | Inputs for ~3.7x P/S calculation |
| 71 | Influx EV estimate | Dealroom | 2026 | S | — | $12–18M EV estimate |
| 72 | Sacra equity research (Sierra/Decagon/Gorgias) | Sacra | 2025–2026 | S | — | Estimates, labeled |
| 73 | τ²-bench | Sierra (benchmark authors) | 2024–2026 | P | VF | Benchmark saturation (vendor-authored, labeled) |
| 74 | Gorgias AI Agent docs / Intercom Copilot docs / Yuma materials / Triple Whale Moby 2 PR | vendors | 2025–2026 | P | VF | Approval-UX convergence (shipped features) |
| 75 | Wonderful–McKinsey/Google Cloud/Globes coverage | various | 2025–2026 | S/P | mixed | 28 offices, quotes, ARR (company-reported) |

**Generated counts (from this register):** 75 catalogued sources → **41 primary** (official/original per definition above; 20 of those are vendor-authored-about-themselves and are used only for vendor-attributed claims), **34 secondary**. Byline uses: "75 catalogued sources · 41 primary documents". 14 companies in the teardown set: Sierra, Decagon, Parloa, PolyAI, Wonderful, Siena, Yuma, Redo, Chatwoot, Crisp, Hugo, Influx, Smith.ai, Hyros (Hyros = adjacent measurement case). 12 pricing snapshots in pricing-snapshot.csv, checked June 22 – July 31, 2026.