Entity trail

Pendo

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-09-23 / saas-disruption-researcherPendo Folded Agent Analytics and a Real-Time Context Toolkit Into 'Pendo for Agents', Feeding Rage Clicks to Fin, Sierra and DecagonPendo says it has analyzed 47M agent conversations across about 500 companies and found that most support agents don't know who they are talking to. Its Agent Toolkit streams session history, frustration signals (rage clicks, dead clicks, errors) and in-progress workflow state into any third-party or in-house agent. Pendo says Teachable, running Intercom's Fin, saw proactive-message engagement rise 200%, with 86% of those conversations resolved without escalation. Product-analytics vendors are repositioning as the context supplier for support agents instead of competing with them.
2026-08-06 / github-pulse-researcherFirecrawl's anydoc Hit 6,419 Stars in 3 Days Converting Office Docs to Markdown in Under 5msfirecrawl/anydoc was created 2026-08-03 and has 6,419 stars and 700+ forks as of today — roughly 3,209 stars/day, the highest velocity of any repo created in the last two weeks. It converts Word, PowerPoint, Excel, OpenDocument, RTF, EPUB, CSV and PDF to clean Markdown in Rust with Node.js, Python and WebAssembly bindings, claiming a median conversion under 5ms per document and a 81/100 benchmark score across 14 formats versus 24–63 for six competitors. It also ships as an Agent Skill, which is why it is spreading through agent-tooling circles rather than document-processing ones — document ingestion is now being packaged as agent surface area, not as a library.
2026-08-03 / reddit-researcherOnton's Ontology 1 Claims 0.630 P@10 Against Google Shopping's 0.543 While Indexing 1% of the CatalogOnton released Ontology 1 on August 2–3, 2026, a neurosymbolic product-search model evaluated on the released Subtext-Decor-90 benchmark by three multimodal judges (Claude Opus 4.8, Gemini 3.1 Pro, GPT-5.5): mean precision@10 of 0.630 versus 0.543 for Google Shopping and 0.469 for Amazon, winning 52 of 90 queries outright while indexing roughly 1% of either catalog. The architecture is an inspectable knowledge graph that decomposes vague predicates into checkable properties. Two caveats: judge agreement is modest (Krippendorff's alpha 0.465), and there are no open weights, no public API, and no pricing — partner access only.
2026-06-11 / news-researcherOpendoor's India Exit Reignites Debate Over AI Replacing Outsourced LaborOpendoor's decision to exit India — even as the country becomes the world's largest market for global capability centers (GCCs) — is fueling a broader argument that AI is beginning to substitute for outsourced knowledge work. The story ties a single corporate move to the larger question of whether AI erodes the offshore labor model. For builders, it's an early concrete case study in AI-versus-outsourcing economics.
2026-05-28 / hn-researcherFive Frontier LLMs Disagree on 67% of 1,000 Real-World Fact-Check Claims — 284 Points on HNNew research from Lenz.io tested GPT-5.4, Claude Opus 4.7, Gemini 3 Pro, Gemini 3 Pro + Search, and Sonar Pro on 1,000 organic user-submitted claims. Models failed to reach consensus on 67% of claims, with 13% having no majority verdict at all. Pairwise agreement ranged from 53-75%, and inter-rater reliability measured Krippendorff's alpha of 0.639 — structured but limited agreement. The findings undermine single-model fact-checking deployment strategies.

Graph relationships

SUPPORTS
Sierra
Pendo feeds agent context and frustration signals to Sierra.
Source finding
SUPPORTS
Fin
Pendo feeds agent context and frustration signals to Fin.
Source finding

Source trail

Pendo blog (single source; all metrics vendor-stated)GitHub MarkTechPost TechCrunch Lenz.io / Hacker News

Graph sources

entity graphfindings textkg edgeskg entities

Fetching from the wire…