AI Dispatch — Daily Dispatch, August 22, 2026

## Sponsor disclosure

No sponsored placement in this issue. `listings` carries zero rows with `tier='featured'` and `status='active'`, checked directly against the database at draft time. When a slot is sold, it will be labeled sponsored, here and on the site.

## Published today

No new entity entered the index in the last 24 hours. Three wire pieces published today, all about vendors already in the index, each built from a single vendor's own site with no unbacked claim:

- **Munich Re** extended its aiSure AI-error performance-warranty product to two new distribution partners — Mosaic Insurance ("Mosaic Insurance has partnered with Munich Re to harness the power of aiSure™") and iTOO in South Africa ("iTOO and Munich Re have partnered to bring aiSure to South Africa"). Munich Re's own page describes the product as indemnifying clients "for their financial losses or legal liabilities directly related to AI errors" but names no coverage limit, premium, or claims process for either arrangement. Source: munichre.com. - **Agent Shield**, an autonomous-agent security tester sold into HIPAA-regulated medical practices under the brand "Tough Love Security," states "We carry professional liability insurance" but names no insurer, limit, or claims process, and caps its own liability at "the engagement fee paid." Source: agentshield.win. - **Mem0**, an AI-agent memory infrastructure product the vendor says is used by "90,000+" developers, claims SOC 2 Type 1, HIPAA, and GDPR compliance and logs "every read and write" — but publishes nothing on insurance, liability caps, or dispute resolution. Source: mem0.ai.

The pattern across all three, and across the index generally: compliance and security claims are common; insurance and liability terms are not.

## What changed in the index this week

Real counts, queried against `invsbcblyrjsvmygulcz` at draft time (August 22, 2026, 16:24 UTC).

- **Published entities: 1,696 of 2,988 tracked.** - **33 new entities entered the index in the trailing 7 days; 45 scorecard rows were written; 17 wire articles published**, 3 of them today. - **10 entities were promoted to published** in the trailing 7 days: Cortx, Skima AI, SuperNinja (Aug 19); iCustomer, Upwex (Aug 18); Keywords AI, Arcade (Aug 17); CampaignStack, OpenHands, Sumora (Aug 16). - **14 previously published entities were returned to candidate status** after audits found their stored evidence described a different company than the one the record was filed under — including a record on a postman.com URL whose capture was actually a competitor's page, and a Twilio-branded record describing a different Twilio product. This was traced to a specific defect: entity names carried from a competitor directory's listing titles while the stored URL points elsewhere. None were deleted; each is re-publishable once correctly re-sourced. - **2 entities were archived**: `happyhorse-ai` (previously published) and `phidata` (a candidate).

We report the demotions the same way we report an absence: plainly, with the count. A reader deciding whether to trust this index should see it catch its own mistakes.

## The assurance gap, in real numbers

How many scored entities land at 0–1 out of 3 — a trust gap under the published rubric — on the three dimensions hardest to fake:

| Dimension | 0–1 of scored | % | Flat 0 | |---|---|---|---| | `insurance_indemnity` | 77 of 79 | 97.5% | 67 | | `audit_trail` | 458 of 487 | 94.0% | 49 | | `outcome_settlement` | 180 of 217 | 82.9% | 69 |

A score of 0 means we read the entity's own material and found nothing on that dimension — not that we didn't look. Full rubric: `/methodology`.

The granular fields behind `insurance_indemnity` are starker than the dimension score alone shows. Each has now been checked, present or absent, on 1,693 to 1,695 of our 1,696 published entities:

| Field | Present | Checked | |---|---|---| | `coverage_limits` | 0 | 1,695 | | `insurance_carriers` | 2 | 1,694 | | `indemnification` | 5 | 1,693 | | `insurance_available` | 6 | 1,695 | | `liability_cap` | 8 | 1,693 |

Not one published entity discloses a policy limit. Zero of 1,695 checked.

## One assurance/insurance development

Today's own wire coverage doubles as this week's development: Munich Re's aiSure product — "AI insurance, pioneered by Munich Re in 2018, is triggered by unexpected errors in the ongoing performance of AI models in daily business use" — now reaches customers through two additional distribution partners rather than through Munich Re directly. Its own page states the warranties "enable you to indemnify your clients for their financial losses or legal liabilities directly related to AI errors," but neither the Mosaic nor the iTOO arrangement discloses a coverage limit, premium structure, or claims process. Source: munichre.com, retrieved 2026-07-31.

## Methodology

Full rubric and sourcing standard: `/methodology`. Every figure above was queried directly against the live database at draft time (August 22, 2026, 16:24 UTC) rather than carried forward from a prior issue.

---

*AI Dispatch is not a growth-hacked directory. Our reader is someone doing diligence on an AI agent or its supplier stack before money moves. We publish absence as a finding and cite everything. If we ever stop doing that, stop trusting the newsletter.*