Help center Customer portals Control customer-portal access: sign-in, case visibility, knowledge audiences and verification Updated Wednesday, October 28, 2026 · Marcel Gravelle A B2B portal often serves several people at the same customer organization while keeping each customer’s information appropriately separated. Good access configuration starts with explicit decisions: what everyone may read, what requires sign-in, which customer cases a person should see and which knowledge is limited to an audience. Supportbench provides separate controls for these decisions. This guide explains how they work together and how to test the result instead of treating one visibility switch as proof that the whole portal is correctly configured. At a glanceFor: support operations administrators responsible for customer self-service and access. Applies to: Supportbench Core.Start here: Settings → Self-Service → Customer portal → select the intended portal Write the intended access rules before changing settings · Open the portal and review the access controls · Use Require sign-in for the portal-wide sign-in decision · Understand case visibility and “unrestricted” · Keep knowledge navigation separate from article authorization · Choose whether anonymous visitors can create cases · Configure a supported identity route without inventing guarantees · Run an access test that can detect both missing and excessive access · Maintain access through customer and content changes · Troubleshoot without widening access unnecessarily Write the intended access rules before changing settings ReaderExample intended experienceVerification questionAnonymous visitorRead general help and product guidance; submit a request only if your policy allows it.Can this visitor reach any customer-specific case or restricted article?Ordinary customer ContactRead the cases and knowledge explicitly permitted for this identity.Does an unrelated Contact or customer record remain unavailable?Customer account coordinatorReview broader customer context only where the configured relationship permits it.Does the intended company scope stop at the correct organization?Contact with an entitled knowledge audienceRead the applicable restricted instructions.Is the same article unavailable to a reader without that entitlement?Former or suspended ContactLose authenticated customer scope while public help follows the portal’s anonymous rules.Does a fresh request apply the disabled Contact state?Use approved test identities that represent these situations. Do not test only with an agent or a privileged customer administrator: their access can conceal a problem that ordinary customers experience. Record the intended access separately from the current settings so the review has a clear standard. Open the portal and review the access controls Open Settings → Self-Service → Customer portal with permission to administer portals. Select the exact portal by name and address.Review its general access controls, including Require sign-in, Visible divisions, Customer roles, Knowledge base containers, Show all containers, anonymous case creation and case-visibility options. Do not assume every access control is inside the Security panel.Record the existing values and intended change. If several branded portals share customer records, review the effect on each relevant experience.Use Security for the applicable SAML configuration and related settings. Do not replace an operational access review with a successful identity-provider login test.Save the intended changes using the portal editor’s save action, reopen the portal configuration and confirm the stored values. Then perform the reader tests below.For every field and the distinction between saving configuration and publishing portal design, use the customer-portal configuration reference. Use Require sign-in for the portal-wide sign-in decision Require sign-in is the explicit control for requiring an authenticated customer for the portal’s protected experience. Core retains designated recovery, survey and unsubscribe routes where those flows must remain reachable. Requiring sign-in is therefore not a promise that every URL returns an identical login screen.Sign-in verifies an identity. It does not establish that the person should see every customer’s cases or every published article. Review the underlying Contact, company relationship and content restrictions. Do not rely on the portal’s Customer roles selection alone as proof that the entire portal is locked: portal subscriptions and content actions have their own checks, and direct page access needs its own verification.Choose the experience deliberatelyA public help center can reduce unnecessary customer effort for common setup questions, while customer-specific case work remains protected. A private portal can require sign-in for the intended protected experience. The right choice depends on your content and operating model; neither choice automatically establishes regulatory compliance or correct permissions for every feature. Understand case visibility and “unrestricted” Control or relationshipWhat to understandVisible divisionsDefines the portal’s applicable division scope. Check case and knowledge behavior against the configured scope.Show unrestricted casesBypasses the portal division/category restriction for cases. It does not bypass the Contact/company/vendor/secondary-contact authorization that determines which cases the identity can access.Allow secondary contact case visibilitySupports the applicable explicitly related secondary-contact case access. It is not a grant to every case in the company merely because the person is a secondary contact somewhere.Contact and company relationshipsContinue to determine permitted customer context. An identity’s company association needs to be accurate.A direct case linkDoes not replace authorization. Test direct links as well as the case list.For example, a customer may have an authorized case in a service division different from the portal’s knowledge division. Show unrestricted cases can remove that division obstacle while the customer’s identity rules still constrain the case set. It does not make another company’s cases universally accessible. Use the cross-division case visibility procedure for a focused walkthrough. Keep knowledge navigation separate from article authorization Knowledge base containers and Show all containers determine which eligible topics are offered in the portal. A topic placement is navigation, not a substitute for the article’s publication workflow and audience. An article can be externally eligible while a version is not yet published, or it can require a particular customer audience.Confirm that the intended article version is published and externally eligible.Check applicable article roles, restricted tags, assets or other configured audience restrictions.Confirm the topic is public/available in the intended portal and that the article is placed or included as intended.Test portal search, the topic list and the direct article link with the allowed reader.Repeat with a reader who should not have access. A hidden navigation link alone is not evidence that the direct article is protected.If an article is not discovered, determine whether the cause is publication, topic selection, search or authorization before broadening its audience. Making a troubleshooting article public just to fix a missing menu entry can expose information to a wider audience than intended. Choose whether anonymous visitors can create cases The anonymous-case-creation option controls an incoming support path. It does not mean the anonymous person receives access to existing customer case history. Decide whether your support model should accept requests before sign-in and what identifying information the case workflow requires.Test a new request from an anonymous browser context using approved sample data.Inspect the resulting Contact/company association, division, queue and activity visibility.Check the customer’s acknowledgement and subsequent reply/access path.Review upstream abuse controls and operational monitoring appropriate to the channel. Do not assume one portal checkbox prevents every spam submission.For an installed-equipment customer, accepting a first request may be useful when the engineer does not yet have a portal account. Your team still needs to verify identity and entitlement before sharing account-specific instructions or service history. Configure a supported identity route without inventing guarantees Use the current SAML guide when connecting an identity provider. Verify the intended customer mapping, certificate, sign-in return path and the provider’s actual policy. A forced redirect does not independently prove that MFA is enforced, that every session has ended or that the reader has the correct customer permissions.The retained SAML Username field setting does not control the current Core matching implementation; do not configure it as a generic LDAP-to-portal mapping mechanism. LDAP sign-in is not available on the native Core portal path, so the previous article’s LDAP setup walkthrough is not current Core guidance. Use a verified supported sign-in route and the customer sign-in procedure for the user experience.Microsoft and Google customer sign-in availability depends on the portal and configured provider connection. Customer identity can require a linking or verification step and is separate from agent enrollment. Confirm the actual configured route instead of promising every portal has the same buttons. Run an access test that can detect both missing and excessive access List the portal address, setting change, allowed identities and specific case/article examples that should be allowed or denied. Use approved test records where possible.Start anonymously. Open the home page, search, a public article, a restricted article’s direct link and the case entry point. Compare each result with the intended policy.Sign in as an ordinary authorized customer Contact. Verify the expected case list and article, then test an unrelated customer’s approved test case/link as a negative check.If using cross-division visibility, test an authorized case outside the visible division and a case that fails customer ownership. The first may be allowed by the setting; the second must not become allowed merely because division restrictions are lifted.If using secondary-contact visibility, test the explicitly related case and another company case without that relationship.Test the relevant identity-provider return and recovery paths. Confirm the correct Contact is admitted, not just that the provider accepted credentials.Save the observed results and correct any unexpected access before describing the configuration as verified. Re-run the affected checks after changing portal scope, customer relationships or article restrictions. Maintain access through customer and content changes Review access when a customer employee leaves, moves organizations or takes on a broader account role. Preserve a deliberate distinction between disabling portal access, changing a relationship and deleting a Contact. Keep an accountable owner for portal settings and one for article audiences.When retiring an article, verify its replacement and audience before redirecting old links. A redirect should lead to relevant guidance that the intended reader can actually open. When adding new public knowledge, inspect screenshots and examples for customer information rather than assuming the topic’s name determines what is safe to publish. Troubleshoot without widening access unnecessarily SymptomInspectCustomer signs in but sees no expected casesActual Contact/company scope, visible divisions, explicit secondary relationship and the case’s current state.A cross-division case is missingShow unrestricted cases plus the continuing customer authorization; the option only removes the division/category gate.An article appears for an agent but not a customerPublished version, external eligibility, article audience, portal/topic availability and the customer’s identity.A disabled Contact still sees general helpAnonymous public-content rules versus authenticated customer access.A role selection seems to have no whole-portal effectUse Require sign-in for the portal-wide gate and test direct content/actions. Do not assume role selection alone protects every route.Sign-in loops or returns the wrong identityProvider configuration, intended tenant/account, customer binding and portal return path.A setting change saved but the experience is wrongReopen the stored configuration and test the actual intended portal/identity. A save acknowledgement is not an access test. Continue learning Configure the customer portalShow permitted cases across portal divisionsHelp customers sign in and recover accessConfigure SAML customer sign-inManage Contact credentials and portal accessPublish knowledge for the correct audience Was this article helpful? Yes No Thank you for your feedback.