# Critical Start and Microsoft, side by side: what each says it will do, and on which data

Side by side, read 10 November 2026

**In short.** Critical Start and Microsoft, read from public pages on 10 November 2026 and set out in the same rows. Actions on the customer's behalf: Critical Start 3 stated, 2 with limits, 1 none found; Microsoft 0 stated, 5 with limits, 1 none found. The page gives no verdict.

## How do Critical Start and Microsoft differ?

On the six actions read, Critical Start has 3 stated, 2 with limits, 1 none found, and Microsoft has 0 stated, 5 with limits, 1 none found. On the seven data sources read, Critical Start has 7 stated, 0 with limits, 0 none found, and Microsoft has 4 stated, 2 with limits, 1 none found.

## Actions on the customer's behalf

| Action | Critical Start | Microsoft |
| --- | --- | --- |
| Isolate a host | Stated Lists endpoint response actions including isolate host. | Stated with limits The provider lists isolate device; experts act only if granted Security Operator access, otherwise actions stay pending. |
| Disable or reset an account | Stated Lists identity actions including disable account and reset password. | Stated with limits The provider lists disable user; experts act only with Security Operator access, otherwise actions stay pending. |
| Block an IP, domain, hash or email | Stated Lists block file hash, block IP and block sender. | None found on 10 November 2026 |
| Kill a process or remove a file | Stated with limits The provider says its customer-approved playbook may include process termination. | Stated with limits The provider lists quarantine file and restrict app execution; experts act only with Security Operator access. |
| Act in a cloud or SaaS console | None found on 10 November 2026 | Stated with limits Experts take action or guide for Defender workloads; guidance only for supported third-party products. |
| Acts without asking first (pre-approved) | Stated with limits Analysts run the customer-approved playbook; customer approves or rejects containment, notice ahead available. | Stated with limits Experts act only if granted Security Operator access; default Security Reader leaves actions pending. |

## Data watched

| Data | Critical Start | Microsoft |
| --- | --- | --- |
| Endpoint | Stated Lists endpoint response actions. | Stated Covers Windows, Linux and macOS devices at High and Medium severity. |
| Identity | Stated Lists identity response actions. | Stated MDR page says it protects identities. |
| Cloud | Stated The provider says cloud is covered by detection; its cloud response is advisory. | Stated with limits Covers cloud apps and workloads; neither plan covers Defender for Cloud storage, containers, databases. |
| Network | Stated Lists network response actions such as block IP. | Stated Microsoft says its MDR page covers network infrastructure across Microsoft and third-party sources. |
| Email | Stated Lists email actions such as quarantine message. | Stated MDR page says it protects email. |
| OT | Stated The provider says OT is read-only monitoring, with Claroty, Nozomi and Defender for IoT integrations. | None found on 10 November 2026 |
| Works on other vendors' tools | Stated Says 100+ integrations and no proprietary agents required. | Stated with limits Plan 2 adds listed third-party sources through Sentinel; Plan 2 requires Sentinel and 1,500 seats. |

## The fifteen questions

| Question | Critical Start | Microsoft |
| --- | --- | --- |
| Who watches the environment around the clock, and who handles a major incident? | Critical Start says a human analyst investigates every alert not auto-resolved; its CIRT, available 24/7, already has environment context. | Microsoft says its experts run MDR; incident response for an active compromise is a separate Defender Experts service. |
| What may the provider do without asking first? | Critical Start says analysts run the customer-approved playbook; the customer approves or rejects containment actions. | Microsoft says experts act on the customer's behalf only if granted Security Operator access; otherwise actions stay pending. |
| Which data sources does the service monitor? | Critical Start says CORR supports 100+ integrations across seven security domains. | Microsoft says MDR protects endpoints, identities, email, cloud apps, cloud workloads and network infrastructure, Microsoft and third-party. |
| Is the provider's own endpoint agent required? | Critical Start says it requires no proprietary agents and connects to existing stack by API. | Plan 1 covers Microsoft Defender workloads; Plan 2 extends to listed third-party telemetry in Sentinel and requires Sentinel. |
| Is threat hunting included, and what output is published? | Critical Start says hunting engagements can be bought with service credits; it also describes continuous hunt-derived detections for all customers. | Microsoft says threat hunting is built in to both plans and customers receive interactive reports on each hunt. |
| Are incident response hours included, or is a retainer needed? | Critical Start shows IR as an add-on to every MDR tier; unused retainer hours convert to credits at 3 hours per credit. | Microsoft presents incident response as a separate Defender Experts service under its own service terms. |
| What response-time commitments are published? | Critical Start publishes a 60-minute mean time to respond in all tiers, credits for misses, and a 15-minute emergency line callback. | Microsoft's overview prints no price or response-time commitment. |
| Is the pricing model published? | Critical Start says MDR pricing is not published and varies by environment, integration scope and engagement level. | Microsoft says the service is sold separately; Plan 2 requires a minimum of 1,500 seats; no price on the overview. |
| Is OT covered? | Critical Start lists OT/ICS as escalation, advisory and read-only monitoring, with Defender for IoT, Claroty xDome and Nozomi integrations. | Microsoft's FAQ excludes IoT devices and does not mention OT. |
| What cloud and identity coverage is stated? | Critical Start lists identity actions (reset password, disable account) and says cloud response is advisory and detection only. | Microsoft says it covers identities and cloud apps; Defender for Cloud storage, containers and databases are not covered. |
| How long does onboarding take? | None found on 10 Nov 2026 | None found on 10 Nov 2026 |
| What reporting does the customer receive? | Critical Start's SLA page publishes service credit remedies; report formats none found on 10 Nov 2026. | Microsoft says customers get live dashboards and reports of the work done on their behalf. |
| Where is customer data kept, and for how long? | None found on 10 Nov 2026 | None found on 10 Nov 2026 |
| Can the customer's analysts see what the provider sees? | None found on 10 Nov 2026 | Microsoft says customers get live dashboards and reports of the work done on their behalf. |
| What are the exit and data-return terms? | None found on 10 Nov 2026 | None found on 10 Nov 2026 |

## What each does well, and where each is limited

| Field | Critical Start | Microsoft |
| --- | --- | --- |
| Does well | Critical Start lists named response actions across endpoint, identity, network and email, and publishes SLA credits. | Microsoft lists named one-click response actions and publishes plan scope, capabilities and boundaries, with interactive hunting reports. |
| Limited | Critical Start lists cloud response as advisory and detection only, OT as read-only monitoring, and says MDR pricing is not published. | Microsoft says MDR does not provide incident response for an active compromise, excludes IoT, iOS, Android and Defender for Cloud workloads, and Plan 2 needs 1,500 seats. |

## What to check in each situation

- A mid-size company with no retainer and no in-house responders: ask both providers "What may the provider do without asking first?" and "Are incident response hours included, or is a retainer needed?" in writing.
- A Global 2000 company with OT, several regions and a board that asks about readiness: ask both providers "Which data sources does the service monitor?" and "Is OT covered?" in writing.
- A company under attack today: ask both providers "Who watches the environment around the clock, and who handles a major incident?" and "What may the provider do without asking first?" in writing.
- A cyber-insured company whose policy names a panel of firms: ask both providers "Are incident response hours included, or is a retainer needed?" and "What response-time commitments are published?" in writing.
- A security team that has tools and alerts but no one awake at 2 a.m.: ask both providers "Is the provider's own endpoint agent required?" and "What may the provider do without asking first?" in writing.

## Related pages

- [Critical Start, all published facts](/services/critical-start/)
- [Microsoft, all published facts](/services/microsoft/)
- [The MDR Service Matrix](/matrix/)
- [All side-by-side pages](/side-by-side/)

## Sources

- Critical Start, "MDR" page (the firm's own page): https://www.criticalstart.com/mdr (read 10 November 2026)
- Microsoft, "Defender experts MDR managed response" page (the firm's own page): https://learn.microsoft.com/en-us/defender-xdr/defender-experts/defender-experts-mdr-managed-response (read 10 November 2026)
- Microsoft, "Defender experts MDR overview" page (the firm's own page): https://learn.microsoft.com/en-us/defender-xdr/defender-experts/defender-experts-mdr-overview (read 10 November 2026)

Last reviewed 10 November 2026. Canonical: https://mdrvsmssp.com/side-by-side/critical-start-and-microsoft/
