Know a company with years of records? Refer it and earn up to $100K
Operations records

Change Advisory Board (CAB): Members, Meetings and Approvals

What an IT change advisory board (CAB) does, who sits on it, which changes it reviews, how a CAB meeting runs and what research says about approval boards.

Short answerChecked

In IT, a change advisory board (CAB) is the group that reviews risky changes to systems and services before they go ahead, such as a data center move or a firewall overhaul. It weighs risk, impact and timing, then approves, rejects or asks for more work. Routine low-risk changes are usually pre-approved, and emergency changes are reviewed after the fact. It is not a customer advisory board.

Change types in common use
3Standard, normal and emergency (Atlassian)
Steps to set up an effective CAB (Atlassian)
6From gap assessment to iterating
Other name in NIST usage
CCBConfiguration or change control board
A server rack with blue patch cables and a laptop showing a ticket queue
Key points
  • This page is about the IT change advisory board, not a customer or product advisory board.
  • Standard changes are pre-approved; normal changes may need the CAB; emergency changes move first.
  • DORA research found no evidence that formal external approval lowers change failure rates.
  • Every change should leave a record with its reason, diff, approver and result, CAB or not.

What is a change advisory board (CAB) in IT?

A change advisory board is an IT governance group, not a panel of customers. Atlassian describes it as "a group of people from different facets of a business who provide oversight and guidance to change management teams" and as "a decision-making body responsible for evaluating and endorsing modifications within a company's IT infrastructure" (Atlassian). A change, in this sense, means "adding, modifying, or removing anything that could have a direct or indirect effect on services" (Atlassian).

Government and defense teams often use another name. NIST's glossary defines a configuration control board (CCB) as "a group of qualified people with responsibility for the process of regulating and approving changes to hardware, firmware, software, and documentation throughout the development and operational life cycle of an information system," and notes it "may also be referred to as a change control board" (NIST CSRC). The job is the same: no risky change goes in without someone weighing it first.

Who sits on a change advisory board?

There is no fixed membership. Atlassian says "no set rules govern CAB composition" and lists the people a working CAB usually includes (Atlassian):

MemberWhat they bring
Change managerLeads the meetings and has the final say
Service desk analystWhat users will notice and what calls to expect
Operations managerDay-to-day running of the systems affected
Application manager or engineerHow the application works and what could break
Information security officerThreats and weaknesses the change might open
Senior network engineerNetwork and cloud dependencies
Business relationship managerCustomer and partner impact

The author of the change usually attends for their item only, to answer questions and explain the rollback plan.

Which changes need CAB approval?

Not all of them. Most teams sort changes into three types (Atlassian):

TypeWhat it isGoes to the CAB?Example
StandardLow risk, repeated often, follows a pre-approved processNo, approved once as a typeAdding storage; replacing a failed router with an identical one
NormalNot an emergency and no pre-approved processHigh-risk ones yes; low-risk ones by a change authority or peer reviewMigrating to a new data center
EmergencyFixes an unexpected error or threat nowReviewed quickly or after the factApplying a security patch; dealing with a server outage

Atlassian recommends moving more work into the standard bucket over time: as teams learn from past incidents, "it should be possible to designate a higher share of changes as standard and pre-approve them" (Atlassian).

What happens in a CAB meeting?

Atlassian's setup steps include creating "a standard CAB agenda" and deciding "the meeting cadence" (Atlassian). Weekly is common for busy teams. A typical agenda:

  • Review changes made since the last meeting, especially any that failed or caused an incident.
  • Go through each new request for change: reason, risk, impact, test results, rollback plan and timing.
  • Decide: approve, approve with conditions, send back for more work, or reject.
  • Check the change calendar for clashes, freezes and busy business periods.
  • Look at emergency changes made since the last meeting and confirm they were recorded.
  • Record every decision and its reason in the change record, not only in the minutes.

Does a change advisory board make changes safer?

Not by itself, according to DORA (DevOps Research and Assessment), a long-running research program on software delivery. Its guidance says heavyweight approval by "a change advisory board (CAB) or a senior manager" has "a negative impact on software delivery performance," and that "no evidence was found to support the hypothesis that a more formal, external review process was associated with lower change fail rates" (DORA). DORA recommends peer review captured in the development platform, plus automated testing and monitoring.

That does not make the CAB useless. DORA lists a "vital role" for it: coordinating between teams, helping with process improvement, and weighing in on business decisions such as "the decision between time-to-market and business risk" (DORA). In practice the CAB works best on the few changes that are truly high risk, while pull request reviews and automated checks handle the rest. The config drift guide covers the fields every change record should keep either way.

Change request example: a firewall rule overhaul

This request for change (RFC) is an illustration written for this page, not a real company's record. It shows what a CAB needs to see and how the decision joins to the code change and the result.

FieldValue
Change IDCHG-2045 (normal change, high risk)
SummaryReplace 140 legacy rules on the edge firewall with 38 grouped rules
ReasonAudit found 61 unused rules and two that allowed any source to the admin range
Code changePull request #311 in the firewall config repository, reviewed by two network engineers
Test evidenceRules applied to the staging firewall for 7 days with no blocked legitimate traffic
RollbackRe-apply the previous tagged rule set (fw-2026-03-01), about 5 minutes
WindowSunday 06:00 to 08:00 UTC, outside the month-end freeze
CAB decisionApproved with condition: security officer on the bridge during the window
ResultApplied 06:12 UTC; one partner VPN blocked; fixed by PR #314 at 06:40; closed as successful with issue
Linked recordsIncident INC-5120 (partner VPN), PR #311, PR #314, audit ticket AUD-77

The last two rows matter most later: they tie the approval to the diff, the incident it caused and the follow-up fix. For how the config side is versioned, see infrastructure as code examples.

Common mistakes with change advisory boards

Most CAB problems come from treating every change the same way.

  • Sending every change to the CAB, so the queue grows and teams batch changes into larger, riskier releases.
  • Approving changes from a one-line summary with no test evidence or rollback plan.
  • Recording the decision in meeting minutes but not on the change record itself.
  • No link from the change record to the pull request or config diff, so nobody can see what actually changed.
  • Emergency changes never reviewed afterwards.
  • Adding more approval steps after every outage. DORA lists "responding to problems by adding more process" as a common pitfall (DORA).

Are change records and CAB approvals also a data asset?

They can be, when each one is joined to the change itself and to what happened next. AI data buyers have asked Origin for the records of companies that run and defend their own infrastructure: incidents joined to the change that fixed them (the ticket, the chat thread, the config diff or pull request, the outcome), configuration and infrastructure-as-code history with commits and reviews going back years, and the runbooks, on-call records and postmortems from the same period. A change record with its CAB decision, its pull request and any incident it caused or fixed is that join, written down at the time. The records must cover the company's own infrastructure (a client's only where that client has agreed), and the company must own them. See config drift, WAF rules and the hub page for infrastructure and incident records.

Origin is an independent data partner. Your company decides what is shared and signs any license itself, and a buyer may pay Origin when a deal closes.

  • Secrets come out first: config diffs often hold keys, passwords and certificates. Scan and rotate before anything leaves (remove secrets from git history).
  • Approver and requester names become consistent pseudonyms; roles can stay.
  • Customer names, tenant IDs and public IP addresses of customers are masked.
  • Changes made in a client's environment stay out unless that client has agreed.

Does your company hold records like these? See what your data is worth.

Common questions

Is a change advisory board the same as a customer advisory board?

No. A change advisory board is an internal IT group that reviews risky changes to systems and services before they go live. A customer advisory board is a group of customers who give feedback on a product and its direction. Both can be called a CAB, but they have nothing else in common.

What is the difference between a change advisory board and a change control board?

Mostly the setting and the name. IT service management teams usually say change advisory board. Government, defense and engineering programs often say configuration or change control board, which NIST defines as the group that regulates and approves changes to hardware, software and documentation over a system's life.

Who chairs a change advisory board?

Usually the change manager. Atlassian describes the change manager as the person who leads CAB meetings and has the final say in decisions. Other members bring operations, security, network, application and business views, and the author of each change normally attends to present it and answer questions.

How often should a change advisory board meet?

Weekly is common for teams with a steady flow of changes; smaller teams may meet every two weeks or only when a high-risk change is waiting. Many teams also review high-risk changes asynchronously in the change tool between meetings, so a change does not wait days for a slot.

What is a standard change in ITIL terms?

A standard change is a low-risk change that is done often and follows a documented, pre-approved process, such as adding storage or replacing a failed device with an identical one. It is approved once as a type, so each instance does not go to the change advisory board.

Is a CAB still needed with continuous delivery?

In a smaller role. DORA research found heavyweight external approval slowed delivery without lowering failure rates, and recommends peer review plus automated testing. It still sees a role for the CAB in coordinating teams, improving the process and weighing business trade-offs on the riskiest changes.

What should a request for change include?

A clear summary, the reason, the systems affected, the risk level, test evidence, the rollback plan, the proposed window and a link to the actual change, such as a pull request or config diff. After the change, add the result and any incident it caused, so the record shows the full story.

Sources and further reading

What is your company’s data worth?

AI companies pay $20K to $5M for company records. Six questions, no files.