Provenance Record

Well-structured News piece (Overview / What We Know / What We Don't Know / Analysis) at 630 words, within the 400-1200 word News range. Clear, neutral prose with no sensationalism or loaded language. Every bullet in 'What We Know' carries an inline citation.

All 4 source snapshots decompressed and read in full from sources/2026-07/lyzr-lets-its-own-ai-agent-run-a-100-million-series-b-fundraise-as-valuation-nears-500-million/: source-0.html.gz (TechCrunch, HTTP 200) confirms the $100M Series B, ~$500M valuation, agent name 'SivaClaw', 130+ investors, memo drafting, slide-tracking, and the $400M interest figure across Silicon Valley/Middle East/financial-sector investors -- all sourced by TechCrunch to Bloomberg (TC's piece is explicitly a brief 'via Bloomberg'; citing the TechCrunch page itself is accurate since that is the page fetched, but the article never surfaces that the underlying reporting is Bloomberg's). source-1.html.gz (The Next Web, HTTP 200) confirms the round is not yet closed ('is still coming together...on track rather than closed', quoted verbatim), the Series A history (Rocketship.vc-led $8M round, Henry Ford III board seat, Accenture Ventures participation), the verbatim Narayan quotes about the fundraising-cycle compression, and the $250M/$50M valuation trajectory -- but names the current round's agent 'Agent Sam,' not 'SivaClaw' (see Findings). source-2.html.gz (Lyzr's own blog, HTTP 200) confirms verbatim the Siva Surendira quote ('The era of siloed AI copilots is over...'), the Madhu Shalini Iyer/Rocketship.vc quote, the full Series A investor list, the 1B+ simulations / 1M+ agents / 30,000+ developers figures, and a Jersey City, NJ mailing address in the page footer (525 Washington Blvd) that corroborates TechCrunch's 'Jersey City, New Jersey startup' framing despite The Next Web separately calling Lyzr 'Bengaluru-based' -- read as complementary (US corporate address vs. founding/engineering base) rather than contradictory, so not flagged as a finding. source-3.html.gz (PYMNTS, HTTP 200) confirms the $14.5M Series A+ at a $250M valuation (fivefold up from $50M) led by Accenture, and the verbatim Kenneth Saldanha quote with accurate title. No hallucinated quotes found anywhere -- every direct quotation in the article is a verbatim match to its cited snapshot. One finding raised: the article's 'Agent Sam was the predecessor to SivaClaw' framing is the bot's own synthesis, not stated by any source, and is contradicted by The Next Web's own text.

All specific numbers, dates, names, and titles trace correctly to their cited sources (Series A: $8M, Oct 2025, Rocketship.vc-led, 8 named participants; Series A+: $14.5M, March 2026, $250M valuation, Accenture-led; Series B: $100M, ~$500M valuation, in progress, 130+ investors, $400M expressed interest). The one factual/synthesis issue -- treating 'Agent Sam' as a retired predecessor to 'SivaClaw' rather than disclosing the sources' unresolved naming conflict -- is confined to a single body bullet in 'What We Know' and does not affect the headline, summary's core claim (that an agent handled investor outreach for the current round, which is independently true), or the lead's central thesis.

Substantively solid, well-sourced News piece with every quote and headline figure verified verbatim against its cited source. The single recoverable issue is a body-level synthesis choice (treating a genuine cross-source naming conflict as a settled 'old agent vs. new agent' timeline) that a corrections note can honestly resolve for readers without gutting the article's lead or summary, both of which remain accurate to their cited source. Verdict: APPROVE_WITH_CORRECTIONS.