Skip to content
Talk to Us

Blocks.ai publishes this comparison and sells in this market. Every claim about another product links to a public source.

DIY vs ServiceNow AI Control Tower: an in-house build or what ServiceNow calls a central hub for AI

Build it yourself (DIY)

Building agent connections and controls in-house from open protocols, existing infrastructure, and open-source tools

ServiceNow AI Control Tower

What ServiceNow calls a central hub to discover, secure, govern, observe, and measure AI

Short answer

DIY is not a product: you connect and govern agents on openly licensed protocols such as MCP and A2A, which leave authorization logic to you.⁠Source 1, Source 2, Source 3 ServiceNow describes AI Control Tower as a central hub to discover, secure, govern, observe, and measure AI across an enterprise; a kill switch can contain a managed agent.⁠Source 4, Source 5

By Blocks.ai teamLast checked

Where each one sits

Six layers of running AI agents at a company, and what each product’s own public sources say it covers.

These aren’t the same kind of product

Where they overlap
Both can list MCP servers and control access to them: Uber built an MCP registry and proxy gateway; ServiceNow’s community documentation describes AI Control Tower’s AI Gateway.⁠Source 6, Source 7, Source 8, Source 9
Where they differ
ServiceNow’s agents are built in ServiceNow’s separate AI Agent Studio.⁠Source 10 A DIY build can also include agent frameworks like LangGraph and service meshes like Istio.⁠Source 11, Source 12
Running both
ServiceNow’s community documentation says agents on a custom-built platform can connect to MCP servers through AI Gateway, with its Gateway URL and OAuth 2.1.⁠Source 7
Public sources · checked 2 October 2026
  • Offered
  • You build it
  • Not publicly documented

DIY

  • Build agents: You build itOpen-source frameworks like ADK⁠Source 13
  • Host and run agents: You build itSelf-hosted, like LangGraph⁠Source 11
  • Identity and access: You build itYour own identity provider⁠Source 14
  • Registry and governance: You build itYour own agent registry⁠Source 15
  • Traffic between agents, tools, and models: You build itYour gateway and service mesh⁠Source 6
  • Agents across organizations: You build itA2A with API management⁠Source 16

ServiceNow AI Control Tower

  • Build agents: Not publicly documented
  • Host and run agents: Not publicly documented
  • Identity and access: OfferedScoped OAuth tokens (community docs)⁠Source 9
  • Registry and governance: OfferedAI inventory tied to CMDB⁠Source 4
  • Traffic between agents, tools, and models: OfferedAI Gateway (community docs)⁠Source 8
  • Agents across organizations: Not publicly documented

At a glance

TopicDIYServiceNow AI Control Tower
What it isAn in-house build on open protocols such as A2A and MCP, which the A2A specification calls complementary.⁠Source 1, Source 2ServiceNow describes AI Control Tower as a central hub to discover, secure, govern, observe, and measure AI across an enterprise.⁠Source 4 ServiceNow says it runs on the ServiceNow AI Platform.⁠Source 4
Agent inventoryYours to build: A2A prescribes no standard API for curated registries.⁠Source 15 Uber and Pinterest each built their own MCP registry.⁠Source 6, Source 17ServiceNow says it auto-discovers agents, models, MCP servers, and more into one inventory tied to the CMDB.⁠Source 4
Agents built elsewhereA2A is designed for agents built on different frameworks by different companies, running on separate servers.⁠Source 18ServiceNow’s community documentation says it covers ServiceNow and third-party AI.⁠Source 19 Connectors you set up reach external platforms.⁠Source 20, Source 21
Stopping an agentYours to build: each A2A server authorizes requests under its own policies.⁠Source 1A kill switch can contain a managed agent and revoke all of its active credentials across connected systems, such as AWS Bedrock.⁠Source 5, Source 22
PricingThe A2A and MCP specifications are openly licensed, A2A under Apache 2.0.⁠Source 1, Source 2 Build and running costs are not publicly documented.ServiceNow says to contact it for pricing.⁠Source 4 Its tiers page says full management of external AI needs a separate license.⁠Source 23 ServiceNow’s community documentation says usage-based costs may apply.⁠Source 7, Source 24

What each one is

Build it yourself (DIY)

DIY means connecting and governing agents without buying a product for it: open protocols such as MCP and A2A wired together, existing infrastructure stretched, an in-house platform, self-hosted open source, or no central approach; Uber and Pinterest each describe building their own MCP registry.⁠Source 6, Source 17

ServiceNow AI Control Tower

ServiceNow describes AI Control Tower as a central hub to discover, secure, govern, observe, and measure AI across an enterprise.⁠Source 4 ServiceNow’s community documentation says AI Stewards approve AI systems and AI Gateway proxies MCP traffic and issues OAuth 2.1 tokens.⁠Source 8, Source 9, Source 19

The differences that matter

  1. Agent inventory

    DIY

    You run your own: A2A prescribes no standard API for curated registries, and Uber and Pinterest each built their own MCP registry.⁠Source 6, Source 15, Source 17

    ServiceNow AI Control Tower

    Connectors you set up discover AI assets on external platforms and sync them into the inventory.⁠Source 20, Source 21

    The official MCP Registry is in preview and does not support private servers; its docs recommend hosting your own private registry for those.⁠Source 25

  2. Access and revocation

    DIY

    A2A leaves authorization logic to each server and MCP makes it optional; Uber’s gateway applies policies from its internal Access Control System.⁠Source 1, Source 6, Source 26

    ServiceNow AI Control Tower

    A kill switch can contain a managed agent and revoke all of its active credentials across connected systems.⁠Source 5, Source 22

    The kill switch covers agents on ServiceNow and four connected platforms.⁠Source 22, Source 27 ServiceNow’s community documentation says AI Gateway issues scoped OAuth 2.1 tokens for MCP.⁠Source 9

  3. Where it runs

    DIY

    Where you put it: Pinterest optimized for MCP servers in its internal cloud, and Uber’s gateway sends requests through its service mesh sidecar.⁠Source 6, Source 17

    ServiceNow AI Control Tower

    ServiceNow says it runs on its AI Platform; data goes to a central ServiceNow environment, maybe in another region or on Azure.⁠Source 4, Source 28

    Separately, a regional data routing option sends LLM and SLM requests to data centers in your region.⁠Source 29 Network requirements, such as ports and egress, are not publicly documented.

For security teams

What a security review asks, answered from each vendor’s public documentation.

TopicDIYServiceNow AI Control Tower
Network exposureMCP servers on Streamable HTTP expose an HTTP endpoint; an A2A agent’s HTTP interface needs an HTTPS URL in production.⁠Source 1, Source 30ServiceNow’s community documentation says agents using AI Gateway call a ServiceNow-hosted URL.⁠Source 8 Ports and egress: not publicly documented.
IdentityIn A2A, identity is established at the HTTP layer, with credentials obtained out of band.⁠Source 1, Source 16 MCP authorization is optional.⁠Source 26ServiceNow’s community documentation says AI Gateway issues scoped, short-lived OAuth 2.1 tokens for MCP connections through it.⁠Source 9
Access changes and revocationYours to build: each A2A server authorizes requests under its own policies, with implementation-specific logic.⁠Source 1A kill switch can contain a managed agent and revoke all of its active credentials across connected systems.⁠Source 5, Source 22
Audit trailA2A’s docs advise auditing task creation, critical state changes, and agent actions; MCP documents trace context propagation.⁠Source 16, Source 31Each kill-switch containment leaves an audit trail.⁠Source 5 ServiceNow’s community documentation says AI Gateway logs every MCP transaction through it.⁠Source 7
ComplianceWith the implementer: A2A’s docs name GDPR, CCPA, and HIPAA; MCP leaves data protections to implementers.⁠Source 3, Source 16ServiceNow says the company has undertaken an annual SOC 2 Type 2 attestation since 2013 and holds ISO/IEC 42001.⁠Source 32

Full comparison

18 criteria in five groups. Every cell links to its source, or says no public source answers it.

DIY and ServiceNow AI Control Tower compared on 18 criteria
DIYServiceNow AI Control Tower
What it is
What it is and who it’s forBuilt in-house from parts such as A2A, an open standard for communication between agent systems, and MCP, which A2A’s specification calls complementary.⁠Source 1ServiceNow describes AI Control Tower as a central hub to discover, secure, govern, observe, and measure AI across an enterprise.⁠Source 4
MaturityVaries by part: MCP’s latest revision is 2026-07-28, and the official MCP Registry is in preview.⁠Source 25, Source 31ServiceNow announced general availability on 6 May 2025.⁠Source 33 ServiceNow’s community documentation says AI Gateway became generally available on 10 September 2026.⁠Source 34
Control
Agent registry and discoveryYours to run: A2A prescribes no standard API for curated registries; the official MCP Registry is in preview and does not support private servers.⁠Source 15, Source 25ServiceNow says it auto-discovers agents, models, MCP servers, and datasets into one inventory tied to the CMDB.⁠Source 4 Connectors you set up reach external platforms.⁠Source 20, Source 21
Identity and access controlIn A2A, identity is set at the HTTP layer and each server authorizes requests under its own policies.⁠Source 1, Source 16 MCP authorization is optional.⁠Source 26ServiceNow’s community documentation says AI Gateway verifies agent identity and issues scoped, short-lived OAuth 2.1 tokens for MCP connections through it.⁠Source 9
Ownership, policy, and revocationMCP leaves consent and authorization flows to implementers.⁠Source 3 Pinterest requires an owning team, a registry entry, and review for non-experimental MCP servers.⁠Source 17A kill switch can contain a managed agent and revoke all of its active credentials across connected systems.⁠Source 5 It covers ServiceNow and four connected platforms.⁠Source 22, Source 27
Audit log and observabilityA2A’s docs advise auditing significant events and distributed tracing, for example with OpenTelemetry.⁠Source 16 Pinterest’s MCP servers log inputs and outputs.⁠Source 17Each kill-switch containment leaves an audit trail.⁠Source 5 ServiceNow’s community documentation says AI Gateway logs each MCP transaction through it.⁠Source 7
Connection
How agents connectMCP servers on Streamable HTTP expose an HTTP endpoint; A2A HTTP interfaces need HTTPS in production.⁠Source 1, Source 30 AWS PrivateLink privately connects a VPC to services.⁠Source 35ServiceNow’s community documentation says agents using AI Gateway call a ServiceNow-hosted URL instead of the MCP server, which must be remote.⁠Source 7, Source 8
Agents across organizationsA2A is designed for agents built by different companies on separate servers.⁠Source 18 Each server authorizes requests under its own policies.⁠Source 1Systems like Microsoft Agent 365 can discover publishable agents via an open API.⁠Source 36 Bringing in agents another organization controls: not publicly documented.
Protocol supportMCP defines stdio and Streamable HTTP transports; A2A maps to JSON-RPC, gRPC, and HTTP/REST.⁠Source 1, Source 37 MCP’s latest revision is 2026-07-28.⁠Source 3ServiceNow’s community documentation calls AI Gateway its MCP enforcement layer.⁠Source 9 ServiceNow’s separate AI Agent Studio can connect external agents over A2A.⁠Source 38
Frameworks, models, and clouds supportedA2A gives a common language to agents built with different frameworks, languages, or vendors.⁠Source 1 Google’s ADK says it is model- and deployment-agnostic.⁠Source 13ServiceNow says it inventories agents, models, and MCP servers from ServiceNow or third parties.⁠Source 4 Connectors reach external platforms.⁠Source 20
Operations
Deployment options and data residencyWhere you choose: Pinterest optimized for MCP servers in its internal cloud, and A2A’s docs leave protecting stored data to your own policies.⁠Source 16, Source 17ServiceNow says it runs on its AI Platform and sends data to a central environment, maybe in another region or on Azure, with controls to turn this off.⁠Source 4, Source 28
Compliance attestationsWith the implementer: A2A’s docs name GDPR, CCPA, and HIPAA; MCP leaves access controls and data protections to implementers.⁠Source 3, Source 16ServiceNow says the company has undertaken an annual SOC 2 Type 2 attestation since 2013.⁠Source 32 It says the company holds ISO/IEC 42001 certification.⁠Source 32
Support and SLADepends on the component: MCP SDKs are tiered partly by maintenance commitments; the official MCP Registry, in preview, gives no uptime guarantees.⁠Source 39, Source 40ServiceNow’s Customer Support Addendum states a 99.8% availability SLA for production instances.⁠Source 41 Its community documentation points to Now Support.⁠Source 19
Time and effort to get runningA curated A2A registry is a service you deploy and maintain.⁠Source 15 Pinterest built a unified deployment pipeline after new MCP servers took too much setup.⁠Source 17Which features you can use depends on your license.⁠Source 42 ServiceNow’s community documentation says many customers implement it with a ServiceNow partner.⁠Source 19
Pricing model and public pricesThe A2A and MCP specifications are openly licensed, A2A under Apache 2.0.⁠Source 1, Source 2 Build and running costs are not publicly documented.ServiceNow says to contact it for pricing.⁠Source 4 It is in all three product tiers.⁠Source 23 ServiceNow’s community documentation says usage-based costs may apply.⁠Source 7, Source 24
Building
Agent building toolsOpen-source frameworks such as LangGraph and Google’s ADK build and deploy agents.⁠Source 11, Source 13 A2A has SDKs in six languages.⁠Source 43ServiceNow agents are built in ServiceNow’s separate AI Agent Studio.⁠Source 10 Creating net-new custom agents is supported only in the Prime tier.⁠Source 23
Model accessChosen by whoever builds the agents: Google’s ADK says it is optimized for Gemini and model-agnostic.⁠Source 13Model provider settings cover ServiceNow-supported providers, such as Now LLM Service and AWS Claude, and ones your organization configures.⁠Source 29
Integrations and ecosystemThe official MCP Registry, in preview, has a REST API for discovering servers.⁠Source 25 Uber says its own gateway hosts over 800 MCP servers.⁠Source 6ServiceNow’s community documentation names discovery connectors for Databricks, Snowflake, and Hugging Face, and says connectors can also be custom-built.⁠Source 19, Source 44

Which to choose

Choose DIY if

  • You want MCP servers in your own cloud, under your own routing and security logic, as Pinterest chose.⁠Source 17
  • You want agents at different companies to authorize each other’s calls themselves: under A2A, each server applies its own policies.⁠Source 1, Source 18
  • You already run a service mesh or access-control system and want it to check MCP calls, as Uber and Pinterest do.⁠Source 6, Source 17
  • You want no agent platform contract: the A2A and MCP specifications are openly licensed, A2A under Apache 2.0.⁠Source 1, Source 2

Choose ServiceNow AI Control Tower if

  • You run ServiceNow on a Foundation, Advanced, or Prime tier, each of which includes AI Control Tower.⁠Source 23
  • You want connectors you set up to discover AI on external platforms and sync it into one inventory.⁠Source 20, Source 21
  • You want a kill switch that can contain a managed agent and revoke all of its active credentials, with each containment audited.⁠Source 5
  • You want MCP traffic governed in one place, as ServiceNow’s community documentation describes AI Gateway: approved servers only, scoped tokens, and logs.⁠Source 7, Source 9

Questions buyers ask

Is ServiceNow AI Control Tower an alternative to building it yourself?

For governance, in part: ServiceNow says it inventories agents, models, and MCP servers, and a kill switch can contain a managed agent.⁠Source 4, Source 5 Agents are built in ServiceNow’s separate AI Agent Studio.⁠Source 10

Can ServiceNow AI Control Tower govern agents built outside ServiceNow?

Yes, with limits: ServiceNow’s community documentation says AI Gateway governs MCP connections from any agent platform, and discovery covers AWS Bedrock, Azure AI Foundry, GCP Vertex AI, and more.⁠Source 7, Source 19 ServiceNow’s docs say its connector for Amazon also discovers Amazon Bedrock AgentCore agents.⁠Source 46

Does ServiceNow AI Control Tower support MCP and A2A?

MCP, yes: ServiceNow’s community documentation calls AI Gateway its MCP enforcement layer.⁠Source 9 For A2A, ServiceNow’s separate AI Agent Studio can connect external agents over A2A.⁠Source 38

How do the costs compare?

ServiceNow says to contact it for AI Control Tower pricing.⁠Source 4 It is included in the Foundation, Advanced, and Prime tiers.⁠Source 23 ServiceNow’s community documentation says usage-based costs may apply.⁠Source 7, Source 24 DIY’s A2A and MCP specifications are openly licensed; build and running costs are not publicly documented.⁠Source 1, Source 2

How we compare

Read the full method

Every claim on this page links to a public source. Where none answers a question, the page says so.

We re-check every fact at least every 90 days. This page was last checked .

Something wrong or out of date? Tell us and we’ll correct it.

Sources

51 public sources, each with the date we checked it. Every one opens in a new tab.

  1. Source 1: Agent2Agent (A2A) Protocol Specification a2a-protocol.org · checked Back:abcdefghijklmnopqr

  2. Source 2: modelcontextprotocol/modelcontextprotocol LICENSE (GitHub) github.com · checked Back:abcdef

  3. Source 3: Specification - Model Context Protocol 2026-07-28 modelcontextprotocol.io · checked Back:abcde

  4. Source 4: AI Control Tower - ServiceNow (product page) ServiceNow · checked Back:abcdefghijklmno

  5. Source 5: AI agent containment using kill switch protocol manually (ServiceNow product documentation, Australia release) ServiceNow · checked Back:abcdefghi

  6. Source 6: Designing MCP Gateway Uber's MCP Management Platform - Uber Blog uber.com · checked Back:abcdefghi

  7. Source 7: AI Gateway FAQ (ServiceNow Community, AI Control Tower articles) ServiceNow · checked Back:abcdefghij

  8. Source 8: AI Gateway Implementation Guide (ServiceNow Community, AI Control Tower articles) ServiceNow · checked Back:abcde

  9. Source 9: What's new in AI Gateway v3.4 - September 2026 release (ServiceNow Community, AI Control Tower articles) ServiceNow · checked Back:abcdefghi

  10. Source 10: AI Agent Studio (ServiceNow product documentation, Australia release) ServiceNow · checked Back:abc

  11. Source 11: langchain-ai/langgraph README (GitHub) github.com · checked Back:abc

  12. Source 12: Security - Istio concepts istio.io · checked Back to text

  13. Source 13: google/adk-python README (GitHub) github.com · checked Back:abcd

  14. Source 14: Integrating with Model Context Protocol (MCP) - Keycloak keycloak.org · checked Back to text

  15. Source 15: Agent Discovery - A2A Protocol a2a-protocol.org · checked Back:abcde

  16. Source 16: Enterprise Features - A2A Protocol a2a-protocol.org · checked Back:abcdefgh

  17. Source 17: Building an MCP Ecosystem at Pinterest - Pinterest Engineering Blog medium.com · checked Back:abcdefghij

  18. Source 18: a2aproject/A2A README (GitHub) github.com · checked Back:abc

  19. Source 19: AI Control Tower Welcome Guide (ServiceNow Community, AI Control Tower articles) ServiceNow · checked Back:abcdefg

  20. Source 20: Discovering AI assets through connectors (ServiceNow product documentation, Australia release) ServiceNow · checked Back:abcde

  21. Source 21: Configuring connectors (ServiceNow product documentation, Australia release) ServiceNow · checked Back:abcd

  22. Source 22: Configure AI agent containment using kill switch protocol manually (ServiceNow product documentation, Australia release) ServiceNow · checked Back:abcde

  23. Source 23: ServiceNow product tiers (ServiceNow product documentation, Australia release) ServiceNow · checked Back:abcde

  24. Source 24: AI Control Tower Observability & Monitoring FAQ (ServiceNow Community, AI Control Tower articles) ServiceNow · checked Back:abc

  25. Source 25: The MCP Registry - Model Context Protocol modelcontextprotocol.io · checked Back:abcd

  26. Source 26: Authorization - Model Context Protocol specification 2026-07-28 modelcontextprotocol.io · checked Back:abc

  27. Source 27: Control enforcement points (ServiceNow product documentation, Australia release) ServiceNow · checked Back:ab

  28. Source 28: AI Control Tower (ServiceNow product documentation, Australia release) ServiceNow · checked Back:ab

  29. Source 29: AI model providers (ServiceNow product documentation, Australia release) ServiceNow · checked Back:ab

  30. Source 30: Streamable HTTP - Model Context Protocol specification 2026-07-28 modelcontextprotocol.io · checked Back:ab

  31. Source 31: Key Changes - Model Context Protocol specification 2026-07-28 modelcontextprotocol.io · checked Back:ab

  32. Source 32: Compliance - ServiceNow Trust ServiceNow · checked Back:abc

  33. Source 33: ServiceNow Launches AI Control Tower, a Centralized Command Center to Govern, Manage, Secure, and Realize Value From Any AI Agent, Model, and Workflow (ServiceNow news release, investor relations PDF) ServiceNow · checked Back to text

  34. Source 34: What's new in AI Control Tower for August & September 2026 (ServiceNow Community, AI Control Tower articles) ServiceNow · checked Back to text

  35. Source 35: What is AWS PrivateLink? - Amazon Virtual Private Cloud docs.aws.amazon.com · checked Back to text

  36. Source 36: External Registries (ServiceNow product documentation, Australia release) ServiceNow · checked Back:ab

  37. Source 37: Transports - Model Context Protocol specification 2026-07-28 modelcontextprotocol.io · checked Back to text

  38. Source 38: Integrating external AI agents (ServiceNow product documentation, Australia release) ServiceNow · checked Back:ab

  39. Source 39: SDK Tiers - Model Context Protocol modelcontextprotocol.io · checked Back to text

  40. Source 40: MCP Registry Aggregators - Model Context Protocol modelcontextprotocol.io · checked Back to text

  41. Source 41: Customer Support Addendum (ServiceNow legal schedules, Version 12MAR2025) ServiceNow · checked Back to text

  42. Source 42: Activating AI Control Tower (ServiceNow product documentation, Australia release) ServiceNow · checked Back to text

  43. Source 43: A2A protocol roadmap a2a-protocol.org · checked Back to text

  44. Source 44: AI Control Tower: What's new in the June 2026 release (ServiceNow Community, AI Control Tower articles) ServiceNow · checked Back to text

  45. Source 45: Activate evaluation scoring for external AI systems (ServiceNow product documentation, Australia release) ServiceNow · checked Back to text

  46. Source 46: AI Service Graph Connector for Amazon release notes (ServiceNow Store version history) ServiceNow · checked Back to text

  47. Source 47: Your company's private network Blocks.ai · checked Back to text

  48. Source 48: Network requirements Blocks.ai · checked Back to text

  49. Source 49: Solutions: Agent sprawl Blocks.ai · checked Back to text

  50. Source 50: Solutions: Partner networks Blocks.ai · checked Back to text

  51. Source 51: Pricing Blocks.ai · checked Back to text