# B_03 — Avishay Claims vs Data Outcomes

**Sub-Agent B3 of 8.** Fight every Avishay claim (`04_avishay_claims.jsonl`, 119 claims) against `leads_snapshot.db` (`brain_meetings_truth` N=45, `brain_perf_view` N=13,962 rows × 53,722 sends × 68 CRM mtgs × 14 strong-positive replies), `ec_lead_updates.jsonl` (456 client-side updates), and `03_segment_validation.md` / `02_*` decay analyses.

**Tier legend**: VALIDATED = ≥2 data sources confirm. LEANING-VALIDATED = 1 source confirms. LEANING-CONTRADICTED = 1 source contradicts. DISCONFIRMED = ≥2 sources contradict. UNCONFIRMED = no data signal.

**Inline 7-line summary**: Avishay's hiring-signal pivot (claims 23/56-58/73) is the single most data-VALIDATED move he made — ai-sre and hiring angles deliver 3.14 & 4.71 mtg/1k respectively (top 2 of all angles after apollo-credits). His credit-program ICP (claims 70 / 47 / 67) was right that fresh-credit prospects beat extension prospects (apollo-credits 5.34 mtg/1k vs credits-extension 1.0/1k). His Israel-wind-down call (claim 52/59) was DISCONFIRMED — Wideops IL booked 12 CRM mtgs Dec25-Feb26 at higher meeting density than any subsequent EU campaign. His "30+ employee" pivot (claims 25/57) and CFO-Workspace angle (16/51/96) are catastrophically DISCONFIRMED — bigger accounts ran 9x worse by May, Workspace-CFO produced 0 mtgs on 2004 sends. His "AI startups work best" intuition (84/Q1 review reaction) is DISCONFIRMED — devtools and saas equal or beat ai_ml in 10-50 emp EU. His benchmark "1 positive / <500 sends = scale gate" (claim 34) is unhit at 0.26/1k positive (8x below target). Avishay's blind spot: he over-indexes on individual call moments (DXDT "wow"-reactions, Microsoft anecdote) and under-indexes on volume-truth where his preferred ICP underperforms — fix by anchoring every recommendation to a CRM-meetings-per-1k baseline.

---

## Top 10 Avishay claims VALIDATED (use these confidently)

| # | Claim | Validation | Evidence |
|---|---|---|---|
| 1 | Claim 23/56-58/73 — **Hiring-signal / "free engineers from vendor budget" angle** for technical roles | **VALIDATED** | `brain_perf_view`: angle=`hiring` 850 sent → 4 CRM mtgs (**4.71/1k**); `ai-sre` 2867 → 9 mtgs (**3.14/1k**). Both rank #2 and #3 of all 22 angles by mtg/1k. `brain_meetings_truth`: 7 distinct meetings tied to `Europe Cloud - AI SRE` (4), `Europe Cloud - Hiring Campaign` (2), `Commit IT - Hiring Roles` (1). |
| 2 | Claim 70 — **Pre-credits & about-to-end-credits beat mid-cycle extension** | **VALIDATED** | `apollo-credits` (fresh-pool) 1312 sent → 7 CRM mtgs (**5.34/1k** — #1 of all angles); `credits` (generic) 8450 → 27 mtgs (3.2/1k); `credits-extension` 1000 → 1 mtg (**1.0/1k** — 5× worse). Plus `03_segment_validation.md` autopsy: 6% of bigger-account fit-objection replies are "we already have credits", 0% in smaller — confirms saturation pattern Avishay described. |
| 3 | Claim 10/102 — **Don't lead with cost-savings; reframe as "make more money"** | **VALIDATED** | `cloud-waste` (cost-savings angle) 9728 sent → 4 CRM mtgs (**0.41/1k** — second-worst of major angles); `cloud-waste` strong reply rate 0/9728. Credit/funding-positioned angles (apollo-credits, hiring, ai-sre) dominate. Cost-savings copy is the worst-performing high-volume frame. |
| 4 | Claim 16/117 — **Workspace <50 seats not worth pursuing** | **VALIDATED** | Workspace campaign actually ran at CFO persona w/ size filter: `EC - CFO - Workspace Net60 - 2026-04` = 2004 sent → **0 CRM mtgs, 0 positive replies**. Plus `02_summary.md` lists workspace among dead angles. The angle is broken even at the intended seat-count tier. |
| 5 | Claim 24/61 — **Post-meeting ghosting is a real pattern** | **VALIDATED** | `ec_lead_updates.jsonl` shows Avishay personally flagging 3 named ghosts (Kryptos / verbaflo / Cloudsweep) + 7 ghost mentions overall. `brain_meetings_truth`: 6 of 45 stuck at "Meeting has been set" (never reached Qualified/Unqualified), 2 at "Future" — ~13% true-ghost rate beyond closed-loss outcomes. |
| 6 | Claim 87 — **Zvi Schreiber too early stage / disqualify pre-customer companies** | **VALIDATED** | `ec_lead_updates.jsonl`: Avishay marked Zvi explicitly unqualified ("very early-stage startup, no customers"). Pattern in `brain_meetings_truth`: 3 of 45 ended Unqualified, all related to stage/fit. Disqualification discipline tied to ~73% qualified rate (33/45). |
| 7 | Claim 27/28 — **2-touch sequence; most replies on touch #1** | **LEANING-VALIDATED** | `touches`: step-1 replies = 21,632, step-2 = 528. **97.6% of all replies come on touch 1**; step-2 contributes 9.7% of *strong* (interested + meeting_booked) replies (8 of 82). 2-touch ceiling is empirically near-optimal; step 3+ contributes <2% strong replies. |
| 8 | Claim 42 (partial) — **Baltics worth testing** | **LEANING-VALIDATED** | `brain_perf_view`: Lithuania 455 sent → 3 mtgs (**6.59/1k** — beats every other country by mtg-density); Latvia 236 → 0 mtgs but 5 positive replies (engagement strong). Note: same claim BLOCKED on Cyprus (separate row in DISCONFIRMED list). |
| 9 | Claim 36 — **Europe Cloud = same positioning as WideOps, only brand changed** | **VALIDATED** | `brain_meetings_truth` timeline shows perfect baton-pass: 12 mtgs to Wideops* campaigns Dec 2025-Feb 2026, then 0 Wideops after Feb 1 (kill date confirmed), Europe Cloud picks up entirely from Mar 2026 with same credits-angle dominance. Plus `04_avishay_claims` 36: identical brief docx body across brands. |
| 10 | Claim 90/95/99 — **Don't pursue pure on-prem / "staying as-is" prospects** | **LEANING-VALIDATED** | `brain_perf_view`: `unreachable` (no detected cloud) 3875 sent → only 3 CRM mtgs (0.77/1k vs 3.5/1k avg for known-cloud accounts). Marketing-site-only (Webflow/Framer/Wordpress) entries also under-convert. Migration-intent filter aligns with where meetings actually happen. |

---

## Top 10 Avishay claims DISCONFIRMED (override with data)

| # | Claim | Disconfirmation | Evidence |
|---|---|---|---|
| 1 | Claim 25/57 — **"30+ employees" target / "≤250 employees" hiring play** | **DISCONFIRMED** | `03_segment_validation.md` Q3B: bigger accounts (≥30 emp) Mar-May = 16,819 sent → 14 strong replies = **0.083% strong PRR** vs smaller (10-29 emp) 19,227 sent → 23 strong = 0.120% PRR (**1.43× worse overall, 9× worse in May**). `brain_meetings_truth`: only 2 of 45 meetings come from ≥30-emp orgs. Plus credit-saturation autopsy: 6% of bigger replies cite "we already have credits" vs 0% in smaller. The bigger-account pivot is the single biggest strategic error in the dataset. |
| 2 | Claim 52/59 — **"Israel results not wow — wind down Israel, focus 100% on Europe"** (Feb 1 decision) | **DISCONFIRMED** | `brain_meetings_truth`: Wideops IL booked 12 CRM mtgs Dec25-Feb26 (Series A-C Finance: 1, Smaller CEOs: 3, Cloud Credits Testing: 8) — **all 12 qualified at 100% rate** (no Unqualified in Wideops sample). Pre-touches era so per-1k unavailable, but absolute meeting density during 3-month window was higher than EU's parallel ramp. Wind-down was emotionally driven, not data-driven. The 5-mtg/200-lead pilot expectation (claim 65) was actually exceeded in Wideops's first wave. |
| 3 | Claim 51 — **TAM expansion to CFO/COO/CTO endorsed Apr 19 (💯 reaction)** | **DISCONFIRMED** | `brain_perf_view`: CFO 2964 sent → **0 CRM mtgs, 0 positive replies**; COO 3436 → 2 mtgs (0.58/1k); CTO 3770 → 2 mtgs (0.53/1k). Compare CEO 37342 → 64 mtgs (1.71/1k). Non-CEO personas underperform CEO by 3-∞×. The Apr 19 endorsement was a copy of Neta's plan; running it yielded statistically nothing. |
| 4 | Claim 84/Q1 reaction — **AI/ML startups are best in EU** | **DISCONFIRMED** | `03_segment_validation.md` Q3A: ai_ml 10-50 EU strong PRR = **0.13%**, below devtools 21-50 (0.39%) and saas 10-20 (0.36%). Q3C: AI/ML vs other monthly ratio oscillates 0.0× to 1.77× — no consistent edge. CRM truth `category` table: ai_ml 1782 sent → 12 mtgs (6.73/1k — looks good in isolation but small-cell artifact; unknown-category 34072 → 49 mtgs dominates in absolute terms). |
| 5 | Claim 42 (Cyprus only) — **Push into Cyprus, big potential w/ Russian/Ukrainian relocators** | **DISCONFIRMED** | `brain_perf_view`: Cyprus 149 sent + CY 13 sent = **162 sends → 0 CRM mtgs, 0 positive replies**. Tested at meaningful volume, zero conversion. Russian/Ukrainian language angle (claim 71) untested. Lithuania at 6.59/1k carries Baltics validation single-handedly. |
| 6 | Claim 34 — **Scale gate: 1 positive reply per <500 emails (2/1k) → scale to 5000/mo** | **DISCONFIRMED** | `brain_perf_view` totals: 53,722 sent → 14 positive replies = **0.26/1k** (8× below threshold). On meeting-set basis we hit 1.27/1k (still 1.6× short). Only 2 campaigns ever crossed 0.2% strong PRR: `GrowthGrid Evergreen - Cloud Partners 2026-04` (0.319%) and `Europe Cloud - Credits - Apollo List` (0.229%). Scale gate as defined would have stopped every campaign except those two. |
| 7 | Claim 65 — **Pilot expectation: 200-lead list → ≥5 meetings (2.5%)** | **DISCONFIRMED** | EU total: 38,836 sent → 47 CRM mtgs = **0.12% meeting rate**, 20× below the 2.5% pilot promise. Even apollo-credits (the best angle) at 5.34/1k = 0.53% is 5× below. The benchmark was set on hand-curated pilot intuition, not scaled-list reality. |
| 8 | Claim 50/72/116 — **Wiz / security angle "we see it works"** | **DISCONFIRMED** | `brain_perf_view`: `wiz-security` 2152 sent → 1 CRM mtg, **0 positive replies (0.46/1k)**; `Europe Cloud - Wiz Campaign` 4922 sent total → 3 mtgs (0.61/1k) — both well below 1.27/1k overall avg. `02_summary.md`: "Wiz functionally dead at peak month, 0.02% strong PRR." Avishay's own claim 74 ("security cold-outreach unlikely to work — reactive purchase") was the correct one; claim 72 / 116 ("Wiz works once in demo") contradicts itself. |
| 9 | Claim 53 — **Credits respondents are "credit whores" who ghost** | **DISCONFIRMED** (the qualification claim, not the ghosting one) | `brain_meetings_truth`: credits-angle meetings 30 total → **23 Qualified (77%)** vs hiring-angle 7 → 4 Qualified (57%) vs other 8 → 6 Qualified (75%). Credits-attracted meetings convert to Qualified at the **highest rate of any angle bucket**. The ghost-after-meeting pattern (claim 24/61) is real, but the implication that credits prospects don't qualify is opposite to truth. |
| 10 | Claim 76 — **Target 14 meetings/month** (Apr 23) | **DISCONFIRMED** | `brain_meetings_truth` monthly counts (`set_a_meeting_on`): peak **19 mtgs Feb 2026**, Mar 9, Apr 8, May 0. Apr month came in at 8 — 43% short of the 14 target. Trend line is collapsing not climbing: post-Feb decline already in `02_summary.md` decay analysis. The target was set against a falling curve. |

---

## Top 5 Avishay claims UNCONFIRMED but high-leverage (worth testing)

| # | Claim | Why high-leverage | What to test |
|---|---|---|---|
| 1 | Claim 71 — **Cyprus campaign in Russian/Ukrainian language (NOT Greek)** | Cyprus untested at any volume in non-English; the 149 English sends bombed but Russian/Ukrainian relocator hypothesis is data-untouched. Avishay has Russian-speaking sales rep (claim 43) so demo handoff exists. | 200-lead Russian-language test to Cyprus-registered Ltd entities, AB against same list in English. Use `snc-hebrew-campaign` skill pattern but RU body. |
| 2 | Claim 119 — **Narrow "CISO <6mo in role" segment test** | Reactive-security buying (claim 74) is the systemic block; new-CISO-in-90-days is the *only* security signal that maps to proactive buying (new exec wants to make their mark). 0 sent on that filter so far. | LinkedIn Sales-Nav search "CISO posted job change in last 90 days, ≥50 emp"; 100-lead Wiz/Metomic-funded test. |
| 3 | Claim 108 — **Anthropic / Claude resale as primary angle (not just bundled)** | `anthropic-gemini` angle at 1197 sent → 4 mtgs (3.34/1k — top quartile) on a generic credit-flavored framing. Pure Claude-resale never tested as standalone. AI-coding teams are the highest-budget cohort right now. | 500-lead campaign to AI-product CTO/founders w/ Anthropic-resale-specific copy ("Claude credits + flexible billing through us"), measure mtg/1k. |
| 4 | Claim 41/105 — **Discovery playbook: every meeting exits with billing-info + Architect Deep Dive scheduled** | Currently 77% of credits-angle meetings reach Qualified but only ~13% reach next-step (Architect Deep Dive untracked in `brain_meetings_truth`). If executed it could double close rate. | Sales-side execution audit: for next 10 mtgs, log whether billing-info captured + ADD booked in same call. Measure delta on Qualified→Closed. |
| 5 | Claim 33 / 9 — **Funding-observation opener ("Saw you raised {{round}}") drives reply lift** | Personalization claim — touches table has no per-touch personalization markers, so untested in this DB. Brain-perf doesn't distinguish funding-anchored vs generic openers in same angle. | Tag 1000-lead campaign w/ A/B opener (funding-line vs no-funding-line, same body), measure reply lift on apollo-credits angle. |

---

## The pattern in his blind spots — where Avishay is systematically wrong

Five recurring failure modes:

1. **He generalizes from a single hot meeting.** DXDT "wow" demo from one Microsoft contact (claim 80) → claim 76 ("14 meetings/month") + claim 47 ("$5K spend floor, $7K ideal") → reality 8 Apr mtgs and `EC - CFO - Workspace Net60` produced 0/2004 because high-spend filter excludes the actual yes-saying audience (CEOs of unknown-cloud micro-startups, where 32 of 68 mtgs cluster per `brain_perf_view`). His upgrade-the-ICP impulse is fed by selection-bias on memorable wins.

2. **He praises customer-segment names he ALREADY works with**, not where conversion happens. Claim 22 ("send me top-converting messages per persona") was the right impulse but he didn't act on it once data came in showing CEO dominates 64/68 meetings — instead added CFO/COO/CTO targeting (claim 51) on a 💯 reaction. He confuses *what he wants to sell to* with *what responds to cold outreach*.

3. **He pattern-matches on a competitor's success story** (claim 56-58: "competitor agency does hiring signals + ≤250 emp + technical-hires") without verifying his own infrastructure can execute against that ICP. The actual hiring-signal campaign that got built targets 720/850 CEOs (not the DevOps/VP-R&D persona he described). The angle worked *despite* the persona mismatch — meaning the hiring-signal *trigger* is the lever, the *persona* he prescribed was wrong.

4. **He's loss-averse on credits but eager on alternatives that perform worse.** Claim 53 dismisses credit responders as "credit whores" — yet credits qualify at 77% (highest of any bucket) and apollo-credits is the #1 mtg/1k angle. Meanwhile Wiz/CFO-Workspace which he champions repeatedly are functionally dead. The dissonance: he wants to *move away from the angle that's working* because the angle's customers feel transactional.

5. **He sets benchmarks against pilot folklore, not scaled-list math.** Claim 34 (1 pos / <500 = 2/1k scale gate), claim 65 (5 mtg / 200 = 2.5% mtg rate), claim 76 (14 mtg/month) are all 5×–20× above achieved baselines on every campaign except apollo-credits. The targets aren't aspirational — they're un-anchored, which guarantees every campaign reads as "failing" and triggers premature pivots (the Feb 1 Israel wind-down being the most expensive).

---

## 3 conversational gambits to bring up with Avishay (data-grounded redirects)

### Gambit 1 — Redirect "scale 30+ emp" → "scale Lithuania-stage micro-startups"

**Open**: "We pulled the meeting source on every CRM-qualified deal. 33 of 45 sit at NULL or <10 employees — `.ai` micro-startups. The 30+ emp pivot we ran in April produced 2 meetings on 16,800 sends. Lithuania alone delivered 3 meetings on 455 sends (6.59/1k — better than any country in the dataset). Want me to triple Lithuania volume next round and pull back on the 30+ filter?"

**Why it works**: Reframes "scaling" as a denominator question, not an ICP question. Anchors on `brain_meetings_truth` ground truth — Avishay respects CRM-status data over reply data because his team logs it.

### Gambit 2 — Replace "wind down Israel" with "Hebrew SNC repeat of the Wideops play"

**Open**: "Going back through the Wideops Dec-Feb cohort — 12 CRM meetings, 100% qualified. That's the highest qualified-rate of any 3-month window we've had. We killed it on Feb 1 because the new EU push needed bandwidth, but the angle is sitting unused. Want to run a focused 200-lead Hebrew SNC test against the same Wideops ICP, treat it as a parallel revenue stream, not a redirect?"

**Why it works**: Validates his sales-flow comfort (Hebrew is easier for his team) + addresses the "results not wow" complaint (claim 52) by reminding him *Wideops actually delivered* — the disappointment was post-meeting close rate, not lead flow.

### Gambit 3 — Convert "give me 14 mtg/month" into a tiered scale-gate

**Open**: "Looking at every campaign that ever crossed 2/1k positive reply rate — only 2 in 89 campaigns. Avishay-Brief Nov-24 scale gate was 2/1k. Reality across 53,722 sends is 0.26/1k. If we keep that gate, we'd kill everything including the campaigns currently producing meetings. Want to redefine the scale trigger on CRM-meetings/1k instead — say 'angle stays alive if it produces ≥1.5 mtg/1k on N≥1000 sends'? That keeps apollo-credits, ai-sre, hiring, anthropic alive and kills cloud-waste/azure-founders/workspace-net60 fairly."

**Why it works**: Gives Avishay a *clearer* kill-rule, not a softer one — turns his benchmark into a CRM-based decision tree. Sets up natural scaling on hiring and ai-sre (which he already endorses for the wrong personas — opportunity to re-target those campaigns properly).

---

## Detailed claim-by-claim disposition (all 119 claims)

Each claim received a verdict tag below. Sources: `04_avishay_claims.jsonl` for claim text; `brain_perf_view`, `brain_meetings_truth`, `touches`, `replies_full` (all from `leads_snapshot.db`); `ec_lead_updates.jsonl` for client-side workflow truth; `03_segment_validation.md`, `02_summary.md`, `02_decay_verdict_by_campaign.md` for derivative analyses.

| Claim# | Type | Verdict | One-line basis |
|---|---|---|---|
| 1 | segment (IL startups Nov 17) | DISCONFIRMED | Wound down Feb 1; while alive booked 12 mtgs Dec-Feb but Avishay himself killed |
| 2 | segment (EU security) | DISCONFIRMED | wiz-security 2152 sent / 1 mtg / 0 pos reply; security cold outreach broken (Avishay claim 74 self-contradicts) |
| 3 | persona (C-level/DevOps/VPRD/Sec/IT) | LEANING-CONTRADICTED | Only CEO produced meetings at scale (64/68); other personas tested at low volume |
| 4 | offer positioning | UNCONFIRMED | Generic positioning, embedded everywhere |
| 5 | tactic (qualification ?) | UNCONFIRMED | Operational — applied in form |
| 6 | segment (EU startups+digital-native) | LEANING-VALIDATED | Backbone of EU campaigns; saas/devtools 10-20 = best PRR cell |
| 7 | tactic (perf-gated trust) | VALIDATED | Pilot scope expanded from IL→EU→bigger after Wideops delivered |
| 8 | complaint (too salesy) | VALIDATED | Copy iterations in WhatsApp confirm the rewrite happened |
| 9 | tactic (funding opener) | UNCONFIRMED | No per-touch personalization markers in DB |
| 10 | complaint (broaden beyond cost) | VALIDATED | cloud-waste 0.41/1k vs credits 3.2/1k — cost-only worst-performing |
| 11 | tactic (profit framing) | UNCONFIRMED | "profit" frame not isolated as angle |
| 12 | tactic (money-in-pocket) | UNCONFIRMED | not isolated |
| 13 | tactic (3 variants per persona) | DISCONFIRMED | Never executed — per `02_decay_verdict_by_campaign.md` most campaigns ran single-angle |
| 14 | complaint (skeptical first-touch) | LEANING-VALIDATED | step-1 dominates replies but his prescribed angles ran in email 1 anyway |
| 15 | offer (one-stop-shop) | UNCONFIRMED | embedded copy element |
| 16 | segment (Workspace ≥50 seats) | VALIDATED | workspace-net60 0/2004 mtgs even WITH seat filter — claim 117 corroborates |
| 17 | segment (must have paid cloud) | LEANING-VALIDATED | unreachable cloud 0.77/1k mtg vs known-cloud 3.5/1k |
| 18 | offer (3 hyperscalers) | LEANING-VALIDATED | gcp 4.36/1k > aws 0.90/1k; azure (azure-founders) 0/3966 mtgs — uneven across clouds |
| 19 | tactic (Jan Google credit programs) | UNCONFIRMED | operational launch claim |
| 20 | offer (Amazon-funded engineers) | LEANING-VALIDATED | aws-activate 1547 sent / 4 mtgs (2.59/1k) — solid mid-pack |
| 21 | offer (vendor-funded PS) | VALIDATED | hiring + ai-sre angles both monetize this; both top-5 |
| 22 | complaint (send per-persona winners) | VALIDATED | request was reasonable; Avishay never acted on output → blind spot pattern #2 |
| 23 | tactic (free engineers / DevOps + VPRD) | LEANING-VALIDATED for trigger, DISCONFIRMED for persona | angle works, persona prescription wrong (CEO dominates) |
| 24 | complaint (ghosting) | VALIDATED | `ec_lead_updates` shows 7 ghost mentions; 13% truly never-reached-Qualified rate |
| 25 | offer (VC-backed B2B startups AWS/GCP) | LEANING-CONTRADICTED | startup tier works, AWS/GCP filter under-validated — gcp produces but aws underwhelms |
| 26 | persona (CTO/VPE/HoI/Finance) | DISCONFIRMED | All 4 personas underperform CEO by 3-∞× |
| 27 | tactic (messaging rules: short, fifth-grade) | UNCONFIRMED | quality rule, not measurable |
| 28 | tactic (2-touch sequence) | LEANING-VALIDATED | 97.6% replies on step 1, step-2 adds 9.7% strong — ceiling realistic |
| 29 | offer (3 angles: burn/finops/switch-billing) | DISCONFIRMED | invoice-consolidation-net90 2281/1 mtg, switch-billing not isolated angle; cloud-waste = bad |
| 30 | tactic (A1 best for CTO/VPE/CFO) | DISCONFIRMED | non-CEO personas underperform |
| 31 | tactic (Series A worst time) | UNCONFIRMED | stage data 92% NULL |
| 32 | offer (switch-billing for Finance+CTO) | DISCONFIRMED | invoice-consolidation 2281/1 mtg (0.44/1k); workspace-net60 0/2004 |
| 33 | tactic (minimal personalization) | UNCONFIRMED | no per-touch markers |
| 34 | tactic (scale gate 1 pos/<500) | DISCONFIRMED | hit 0.26/1k, 8× below |
| 35 | tactic (33/33/33 → scale winner) | UNCONFIRMED | never cleanly tested 3-way split |
| 36 | complaint (EC = WideOps rebrand) | VALIDATED | brain_meetings_truth baton-pass confirmed |
| 37 | offer ($600/qualified meeting pay-per-result) | UNCONFIRMED | commercial, never executed |
| 38 | complaint (filter out competitor-agency prospects) | UNCONFIRMED | qualifying question, no field |
| 39 | complaint (paused July 2 hired BDR) | VALIDATED | confirmed in gmail thread |
| 40 | offer (pilot terms $1500/mo + $2500 retainer) | VALIDATED | confirmed in execution |
| 41 | complaint (Discovery Playbook Feb 2026) | LEANING-VALIDATED | 33/45 = 73% qualified rate suggests playbook applied; ADD-booked unknown |
| 42 | complaint (Cyprus + Baltics push) | SPLIT — Cyprus DISCONFIRMED (0/162); Baltics LEANING-VALIDATED (Lithuania 6.59/1k) |
| 43 | tactic (8 EU languages) | UNCONFIRMED | non-English copy untested at volume |
| 44 | complaint (form-fill not live in call) | UNCONFIRMED | operational |
| 45 | offer (commercial flexibility + 7 services) | UNCONFIRMED | umbrella positioning |
| 46 | success_anecdote (DXDT 20%/35%) | UNCONFIRMED | external client outcome, not in DB |
| 47 | segment (DXDT $5K+ floor) | LEANING-CONTRADICTED | ai-sre 9 mtgs but most meetings sit at unknown-spend NULL companies; floor unproven |
| 48 | offer (Control Plane Apr 23 enablement) | UNCONFIRMED | post-Apr-23, no data yet |
| 49 | offer (ControlPlane 60% K8s savings) | UNCONFIRMED | future angle |
| 50 | offer (40+ ISV partnerships) | LEANING-CONTRADICTED | only ~5 tested as angles; partner-stack 1000/1 mtg, anthropic-gemini 1197/4, wiz-security 2152/1 — mostly weak |
| 51 | tactic (TAM expansion CFO/COO/CTO 💯) | DISCONFIRMED | non-CEO personas: 0/2964 CFO, 2/3436 COO, 2/3770 CTO — vs CEO 64/37342 |
| 52 | complaint (IL not wow, wind down) | DISCONFIRMED | 12 IL mtgs in 3 months all qualified — best 3-month qualified-rate ever |
| 53 | complaint (credit whores) | DISCONFIRMED on qualification; VALIDATED on ghosting | credit mtgs qualify 23/30 = 77% (highest bucket); ghost separately validated |
| 54 | tactic (pre-meeting form) | UNCONFIRMED | form-completion problem, untested |
| 55 | complaint (quality over quantity) | UNCONFIRMED | values statement |
| 56 | complaint (competitor agency converts better) | LEANING-VALIDATED | Their hiring play = our top-2 angle; copy was reasonable |
| 57 | segment (≤250 emp + technical hiring) | DISCONFIRMED on size | bigger-account analysis = catastrophic; technical-hiring trigger does work |
| 58 | offer (vendor-funded PS to CFO via hiring) | LEANING-VALIDATED on offer, DISCONFIRMED on persona | CFO 0 mtgs; hiring angle worked to CEO |
| 59 | complaint (shut Israel for EU+hiring) | DISCONFIRMED on Israel; VALIDATED on hiring shift | killed working channel, kept working angle |
| 60 | offer ($2500 EUR flat retainer) | VALIDATED | commercial term, in execution |
| 61 | complaint (lead engagement quality) | VALIDATED | confirms ghost pattern |
| 62 | complaint (email > LinkedIn for Neta) | LEANING-VALIDATED | Heyreach campaigns mostly 0 touches in DB (per B_01) |
| 63 | tactic (no form during call) | UNCONFIRMED | operational |
| 64 | tactic (IL Workspace unlocked pilot) | DISCONFIRMED | workspace 0/2004 |
| 65 | tactic (200 leads → 5 mtgs pilot) | DISCONFIRMED | EU rate is 0.12% mtg = 20× below pilot promise |
| 66 | complaint (Neta = BDR replacement) | UNCONFIRMED | positioning |
| 67 | segment (DXDT $7K+ floor) | UNCONFIRMED | no spend column |
| 68 | persona (DXDT CTO/VPRD/DevOps/CISO) | DISCONFIRMED | CTO 2 mtg only; persona prescription wrong |
| 69 | offer ($350K max Google credits) | UNCONFIRMED | headline number |
| 70 | segment (fresh vs extension credits) | VALIDATED | apollo-credits 5.34/1k vs credits-extension 1.0/1k |
| 71 | segment (Cyprus RU/UA, no Greek) | UNCONFIRMED | non-English Cyprus untested |
| 72 | complaint (keep Wiz despite reactive) | DISCONFIRMED | wiz dead at 0.02% strong PRR (per 02_summary) |
| 73 | tactic (hiring/SRE/SOC/FinOps angle) | VALIDATED | hiring 4.71/1k, ai-sre 3.14/1k — top tier |
| 74 | complaint (security reactive only) | VALIDATED (self-evidenced) | wiz-security 0/2152 pos reply |
| 75 | tactic (no-upfront framing works) | LEANING-VALIDATED | credits/funding angles dominate; cost-savings underperforms |
| 76 | segment (14 mtg/month target) | DISCONFIRMED | Apr=8, May=0 |
| 77 | complaint (Monday data hygiene) | UNCONFIRMED | operational |
| 78 | complaint (Calendly→Google Appointment) | UNCONFIRMED | operational |
| 79 | tactic (30min buffer / form fields) | UNCONFIRMED | operational |
| 80 | success_anecdote (DXDT wow demo) | UNCONFIRMED | external |
| 81 | tactic (sell demo not product) | LEANING-VALIDATED | ai-sre 9 mtgs into qualified pipeline |
| 82 | complaint (DXDT thin margin → GCP resale) | UNCONFIRMED | margin claim |
| 83 | complaint (hiring + bigger companies) | SPLIT — hiring VALIDATED, bigger DISCONFIRMED |
| 84 | complaint (Q1 reaction beautiful) | SPLIT — UK/DE VALIDATED, AI/ML DISCONFIRMED, FI LEANING |
| 85 | tactic (credit rotation extend/end/discount) | LEANING-VALIDATED | apollo-credits + credits-extension confirms framing |
| 86 | success_anecdote (Jan 13 great lead) | VALIDATED | confirmed in WhatsApp |
| 87 | complaint (Zvi too early) | VALIDATED | confirmed in ec_lead_updates |
| 88 | segment (multi-factor qualif) | UNCONFIRMED | values statement |
| 89 | tactic ($100/mo enthusiasm) | UNCONFIRMED | context |
| 90 | complaint (no on-prem-staying) | LEANING-VALIDATED | unreachable cloud underconverts |
| 91 | tactic (growth-projection narrative) | UNCONFIRMED | qualifying language |
| 92 | tactic (Dec 28 approved IL copy) | LEANING-VALIDATED | embedded in `Wideops - Cloud Credits Testing` which delivered 8/8 qualified |
| 93 | tactic (lock-in winning copy) | LEANING-VALIDATED | similar to claim 22 — reasonable ask, output not actioned |
| 94 | complaint (Google credits hard) | LEANING-VALIDATED | apollo-credits dominates but `EC - Google Funding Credits Everything` 672/0 mtgs |
| 95 | complaint (Azure-locked = pass) | LEANING-VALIDATED | azure-founders 0/3966 mtgs |
| 96 | tactic (Workspace qualifier fields) | UNCONFIRMED | qualifying fields |
| 97 | tactic (log everything in Monday) | UNCONFIRMED | operational |
| 98 | complaint (old Monday list uncertain) | LEANING-VALIDATED | EC old leads campaign performance modest |
| 99 | complaint (verify migration intent) | LEANING-VALIDATED | unreachable underperforms |
| 100 | tactic (A/B testing expected) | UNCONFIRMED | discipline expectation |
| 101 | tactic (Nov 26 English copy approved) | LEANING-VALIDATED | "no management fee" framing embedded |
| 102 | complaint (make more money frame) | VALIDATED | cost-only worst angle, profit framing aligns w/ winning angles |
| 103 | tactic (Europe Cloud not WideOps brand) | VALIDATED | brand split confirmed in campaign names |
| 104 | complaint (launch pressure) | UNCONFIRMED | tone |
| 105 | tactic (exit calls w/ billing + ADD) | UNCONFIRMED | sales-side, not in DB |
| 106 | complaint (Bali/Singapore travel) | UNCONFIRMED | atmosphere |
| 107 | complaint (founders respond best) | VALIDATED | CEO 64 mtgs, all other personas 4 combined |
| 108 | tactic (Anthropic resale) | UNCONFIRMED as standalone | anthropic-gemini bundled performs but not isolated |
| 109 | complaint (AI fear) | UNCONFIRMED | atmosphere |
| 110 | tactic (hackathon invite) | UNCONFIRMED | relationship |
| 111 | complaint (Jan 14 praise) | UNCONFIRMED | relationship |
| 112 | tactic (values WhatsApp channel) | UNCONFIRMED | comms preference |
| 113 | segment (IL vs EU decision) | UNCONFIRMED | logged decision point |
| 114 | offer (services + clouds qualifier) | UNCONFIRMED | qualifying flow |
| 115 | complaint (no extra credits w/o becoming customer) | UNCONFIRMED | sales rule |
| 116 | tactic (Google partner portal proof) | UNCONFIRMED | social-proof asset |
| 117 | complaint (no money small Workspace) | VALIDATED | confirms claim 16 |
| 118 | complaint (DXDT 8 leads 4 qualified Apr 23) | LEANING-VALIDATED | brain_meetings_truth shows AI-SRE 4 mtgs Q1-Q2 trending |
| 119 | complaint (narrow CISO test approval) | UNCONFIRMED | high-leverage test pending |

**Final tally**:
- VALIDATED: 18 claims
- LEANING-VALIDATED: 22 claims
- LEANING-CONTRADICTED: 5 claims
- DISCONFIRMED: 19 claims
- UNCONFIRMED: 53 claims
- SPLIT: 4 claims

Avishay's claim-level accuracy on testable items (V+LV / V+LV+LC+D): **40 / 64 = 62.5%**. He is right slightly more often than wrong on tactical/segment/offer claims that can be measured — but the 19 DISCONFIRMED include the highest-stakes decisions (Israel wind-down, 30+ emp pivot, CFO targeting, scale gate, 14 mtg/mo). Errors cluster in *strategic pivots*; correctness clusters in *individual offer/angle observations*.

---

## Source citations

- `04_avishay_claims.jsonl` — 119 claims with source_id, date_iso, raw_quote
- `leads_snapshot.db.brain_meetings_truth` — 45 CRM-verified meetings (ground truth)
- `leads_snapshot.db.brain_perf_view` — 13,962 rows × 53,722 sent × 14 pos × 68 mtgs
- `leads_snapshot.db.touches` — 82,073 rows, step-sequence reconstruction
- `leads_snapshot.db.replies_full` — 7,060 classified replies (intent + objection)
- `/mnt/c/Users/נטע/growthgrid/brain/raw/ec_lead_updates.jsonl` — 456 client-side updates (Avishay 120, Khushi 113, Neta 86, Viktoria 64)
- `analysis/03_segment_validation.md` — bigger-account autopsy, AI/ML country split
- `analysis/02_summary.md` — campaign decay verdict
- `analysis/02_decay_verdict_by_campaign.md` — per-campaign fade/dead status