Skip to content

QryptoCon 2026 · Dec 9–10 · Expert sessions on PQC, PKI, & Certificate Management

Save Your Seat →

CLM Implementation Guide: Your First 90 Days with a CLM Platform

Certificate Lifecycle Management

Buying a certificate lifecycle management (CLM) platform does not prevent a single outage on its own. Outages stop when the platform knows about every certificate, every certificate has a named owner, and renewals happen without anyone having to remember them. That work happens in the first 90 days after purchase, and right now the clock is tighter than usual. The CA/Browser Forum’s 100-day maximum for publicly trusted TLS certificates takes effect on March 15, 2027. A team that starts this playbook in mid-November 2026 finishes its first 90 days in mid-February 2027, with roughly four weeks to spare.

This guide lays out those 90 days as a phased playbook. Each phase has one accountable owner, a short list of deliverables, and exit criteria you can actually check off. It is written for the people who carry the rollout: the project sponsor, the PKI team, and the infrastructure and application owners whose systems hold the certificates.

Quick Answer: What Does a 90-Day CLM Implementation Look Like?

A 90-day CLM implementation runs in five phases after a short readiness step. Teams deploy the platform and connect their certificate authorities (Days 1-15), discover and baseline every certificate (Days 16-35), set policy, alerting, and workflows (Days 36-55), pilot automated renewal on a few high-impact systems (Days 56-75), and then scale with runbooks, KPIs, and a wave plan (Days 76-90). Each phase ends at a gate with a single accountable owner.

Key Takeaways

  • Readiness work before Day 1 (service accounts, firewall rules, CA credentials, change approvals) is the most common reason a CLM schedule slips. Start it the week the contract is signed.
  • Discovery is a phase with its own owner and deadline. The first full scan almost always finds certificates that were never on the spreadsheet, and each one needs an owner before it can be automated safely.
  • Ownership comes before automation. A renewal agent can replace a certificate in seconds, but only the application owner knows whether replacing it will break a pinned client or a forgotten integration.
  • Automate first where an outage costs the most, such as customer-facing load balancers and web tiers, rather than where certificates are most numerous.
  • With the 100-day TLS cap arriving March 15, 2027 and the 47-day cap on March 15, 2029, a good 90-day plan ends with zero-touch renewal proven in production and a dated wave plan for everything else.

Why the First 90 Days Decide Whether CLM Works

The first 90 days decide whether a CLM platform becomes the system of record for certificates or a dashboard one PKI administrator checks on Monday mornings. CLM projects rarely fail on technology. They stall on access requests, unclear ownership, and change approvals that nobody chased. A phased plan with named owners keeps those blockers visible while there is still executive attention on the purchase.

The cost of stalling is well documented. Keyfactor’s Digital Trust Digest: The Automation Edition (September 2025, 450 PKI and certificate management practitioners) found that 86% of organizations had at least one outage from expired or mismanaged certificates in the past year, and 10% reported one every week. The DigiCert Trust Pulse Survey (July 2025) found that 45% of enterprises had certificate-related downtime in the prior year, 37.5% traced outages specifically to expired certificates, and more than half of those affected endured 5 to 24 hours of downtime.

The Validity Schedule Your Rollout Is Racing

CA/B Forum Ballot SC-081v3, adopted in April 2025 with unanimous support from Apple, Google, Microsoft, and Mozilla, reduces the maximum validity of publicly trusted TLS certificates and the period for reusing domain validation in three steps:

Effective dateMaximum TLS certificate validityMaximum domain validation reuseStatus (November 2026)
March 15, 2026200 days200 daysIn effect
March 15, 2027100 days100 daysAbout 4 months away
March 15, 202947 days10 daysUpcoming

At 100-day validity, a certificate renewed at about two-thirds of its lifetime is replaced roughly every 67 days, or five to six times a year. At 47 days, that rises to about 11 or 12 renewals per certificate per year. Privately trusted certificates issued by an internal CA are not bound by SC-081v3, but many organizations are choosing to shorten internal lifetimes as well. Our blog on TLS certificate validity reduction to 47 days covers the schedule in more detail.

Who Owns What: The CLM Implementation Team

A CLM rollout needs one accountable person per role, not one accountable team. “The network team” cannot approve a firewall change; a named network engineer can. The table below lists the roles that appear in every phase of this playbook. The time commitments are planning estimates, so adjust them to the size of your certificate estate.

RoleTypical personAccountable forTime in the first 90 days
Executive sponsorCISO, VP of Infrastructure, or Head of Security EngineeringUnblocking access and change approvals; signing off each phase gateAbout 1 hour every two weeks
CLM program ownerPKI lead or security engineering managerThe plan, schedule, exit criteria, and weekly status50-100%
PKI / CA administratorMicrosoft AD CS administrator, private CA owner, public CA account administratorCA connectors, certificate templates and profiles, enrollment protocols30-50%, mostly Phases 1 to 3
Infrastructure ownersServer, network and load balancer, cloud platform, and Kubernetes leadsNetwork access for discovery, agent deployment, deployment targets10-25%, peaking in Phases 2 and 4
Application ownersService owners for in-scope applicationsClaiming certificates, approving renewals, testing after replacementA few hours each, mostly Phases 2 and 4
Security / GRCSecurity architecture, compliance, internal audit liaisonCertificate policy, cryptographic standards, audit evidence10-20%, peaking in Phase 3
ITSM ownerServiceNow or service desk platform ownerTicket routing, approval workflows, CMDB linksAbout 10% in Phases 3 and 4
Implementation partnerVendor professional services or a PKI services firmDeployment, connector configuration, knowledge transferAs defined in the statement of work

Publish this table, with real names in the second column, before Day 1. It doubles as the escalation list when something stalls.

Before Day 1: The CLM Readiness Checklist

Owner: CLM program owner, with the executive sponsor chasing approvals.

The items below take the longest to approve in most enterprises, which is why they belong before the clock starts rather than inside Phase 1:

  • Service accounts for each certificate authority. Create an account with only the permissions the CLM connector needs on each issuing CA, whether that is a Microsoft AD CS issuing CA, a private CA, or a public CA account.
  • API credentials for clouds and secret stores. Request read access to the AWS, Azure, and Google Cloud accounts, key vaults, and secret managers that hold certificates.
  • Firewall rules for discovery. Open paths from the discovery engine to the IP ranges and ports where TLS listens. Port 443 is the obvious one; 8443, 636 (LDAPS), 993, 465, and application-specific ports are where unknown certificates tend to hide.
  • Change approval for scanning. Tell the security operations center that discovery scans are coming, so they are not flagged as reconnaissance.
  • Load balancer and web server access. Arrange API or administrative accounts for appliances such as F5 BIG-IP, plus deployment rights on the web servers in scope.
  • Hosting decision. Choose SaaS, on-premises, or hybrid. For on-premises, provision the servers, database, and high availability design now.
  • Single sign-on details. Prepare the identity provider configuration for the CLM console.
  • Baseline metrics. Record how many certificates you think you have, how many certificate-related incidents you had in the last 12 months, and how long one manual renewal takes end to end. You will need these numbers for the Day 90 readout.
  • A one-page scope statement. List what is in scope for 90 days (for example, public TLS and internal web TLS) and what is explicitly deferred (for example, code signing, IoT devices, and mainframes).

Exit criteria: every credential issued and tested, firewall changes approved, scope statement signed by the sponsor.

Phase 1 (Days 1-15): Deploy and Connect

Owner: CLM program owner, working with the implementation partner. Supporting: PKI / CA administrator, Security.

Objective: The platform is running, the right people can log in, and every in-scope certificate authority is connected.

  1. Stand up the platform in the chosen deployment model. For on-premises installs, configure high availability and backups at the start; retrofitting them later means a second change window.
  2. Integrate single sign-on and define console roles. A practical starting set is platform administrator, PKI operator, application owner (self-service), and auditor (read-only). Keep the administrator group small.
  3. Connect every certificate authority in scope: internal Microsoft CAs, other private CAs, cloud CAs, and public CA accounts. Importing the issued-certificate records from each CA gives you a first inventory before a single network scan runs.
  4. Configure notification channels (email, Microsoft Teams or Slack, ITSM, SIEM) and send a test alert through each one.
  5. Document the architecture and data flows for your security review: what the platform connects to, which credentials it holds, and where logs go.

Exit criteria: all in-scope CAs connected and syncing; CA-issued inventory imported; a test alert received on every channel; security sign-off on the architecture.

Watch out for: treating the CA import as the live inventory. CA records show what was issued, not what is deployed. Many issued certificates were never installed or have since been replaced. Phase 2 tells you which ones are actually in use.

Phase 2 (Days 16-35): Discover and Baseline

Owner: CLM program owner. Supporting: Infrastructure owners run and unblock scans; application owners claim their certificates.

Objective: One inventory, every certificate classified, and a named owner on every production certificate.

Use More Than One Discovery Source

Network scanning only finds certificates that are being served on a listening port. Certificates sitting in Windows certificate stores, Java keystores, PEM files, secret managers, and container images never answer a scan. A complete baseline combines several sources:

  • CA imports from Phase 1, which show everything issued by connected CAs.
  • Network scans across in-scope IP ranges and ports, which show what is actually served, including certificates from CAs you have not connected.
  • Cloud and secret store connectors for certificates in cloud load balancers, key vaults, and secret managers.
  • Agent or file-based discovery on servers, which finds client authentication certificates, keystores, and files that are not exposed on a port.
  • Load balancer and web server integrations, which tie each certificate to the exact virtual server or site that uses it.
  • Container and Kubernetes discovery for certificates held in cluster secrets and ingress configurations.

Triage Into Buckets, Not One Long List

A raw inventory of thousands of rows is not actionable. Sort it into buckets with a clear next action for each:

BucketWhat it meansNext action
Expiring within 30 daysLive certificates with no renewal planRenew now through the current process; add to pilot candidates
Unknown ownerNo application or team claims the certificateRoute to the owner of the IP range or host; escalate to the sponsor after one week
Weak cryptographyRSA keys under 2048 bits, SHA-1 signatures, deprecated TLS versionsFlag to Security; replace at next renewal or sooner if exposed
Self-signed or untrusted issuerNot chained to an approved CAReplace with a certificate from an approved CA, or record a dated exception
Wildcards and shared keysOne certificate or key installed on many hostsMap every host using it; plan to split before automating
Issued but not found deployedIn the CA database, not seen anywhere in the estateConfirm with the owner; revoke if unused

Make Ownership Stick

Tag each certificate with an owner, the application it serves, its environment (production, staging, development), and its business criticality. If you have a CMDB, link certificates to configuration items so ownership survives team changes. Ask application owners to claim their certificates themselves; a PKI team that becomes the default owner of everything becomes the bottleneck for every renewal.

PQC tip: capture the key algorithm, key size, and signature algorithm for every certificate now. NIST IR 8547, finalized in October 2026, deprecates quantum-vulnerable algorithms such as RSA and ECDSA after 2030 and disallows them after 2035. The certificate inventory you build in Phase 2 is the first input to a full cryptographic inventory, which tools such as CBOM Secure extend beyond certificates to keys, libraries, and protocols.

Exit criteria: discovery complete across every in-scope source; an owner on a target share of production certificates (90% is a reasonable bar); every bucket has a remediation date; baseline report shared with the sponsor.

Phase 3 (Days 36-55): Policy, Alerting, and Workflows

Owner: Security / GRC for the policy. Supporting: PKI / CA administrator enforces it in the platform; ITSM owner builds the workflows.

Objective: New certificate requests follow written policy, expiry alerts reach the right person at the right time, and requests flow through your service management process.

Write a Certificate Policy the Platform Can Enforce

A certificate policy only works if the platform can check it on every request. Cover these decisions at minimum:

  • Approved CAs per use case: public TLS, internal TLS, client authentication, S/MIME, and code signing.
  • Key algorithms and sizes: for example, RSA 2048 or 3072 and ECDSA P-256 or P-384, according to your internal standard, plus approved signature hashes.
  • Maximum validity per certificate type: public TLS is set by the CA/B Forum schedule; internal lifetimes are your choice, and shorter is safer once renewal is automated.
  • Naming and SAN rules: which domains each team may request, and when wildcards are allowed.
  • Approval rules: which requests need one or more approvals, such as wildcards, external-facing names, and code signing.
  • Exceptions: who can grant one, how it is recorded, and the date it expires.

Write the policy you can enforce today and tighten it each quarter. A strict policy that nobody follows turns exceptions into the norm.

Tier Your Alerts

Sending every expiry notice to a shared mailbox guarantees that nobody reads them. Route alerts by urgency. For 100-day certificates, a sensible starting point is:

  • 30 days before expiry: notify the certificate owner and log to the SIEM.
  • 14 days: notify the owning team’s channel and open an ITSM ticket.
  • 7 days: escalate the ticket priority and notify the team lead.
  • 3 days: page the on-call engineer.

When certificates reach 47-day lifetimes in 2029, compress the tiers. Our post on certificate management best practices covers renewal windows in more depth.

Connect Workflows and Reporting

Route certificate requests through your ITSM platform so approvals, change records, and audit evidence live in one place. Schedule a weekly expiring-certificates report for owners, a high-risk certificate report for Security, and confirm the audit trail retention matches your compliance requirements.

Exit criteria: policy approved and enforced on new requests; alert tiers live and tested; ITSM integration tested end to end; first audit report generated.

Phase 4 (Days 56-75): Automation Pilot

Owner: CLM program owner. Supporting: Infrastructure owners and application owners for each pilot system.

Objective: Prove zero-touch renewal in production on a small number of systems that matter, including what happens when automation fails.

Pick Pilots That Prove Something

Choose two to four pilot use cases. Score candidates on business impact if the certificate expires, how often it renews, whether the platform supports the target natively, how willing the owner is, and how easy it is to roll back. Systems that usually make good first pilots:

  • A customer-facing web tier behind a load balancer such as F5 BIG-IP.
  • Linux web servers running Apache or Nginx.
  • Windows servers running IIS.
  • A Kubernetes ingress using ACME.

Leave code signing, IoT fleets, mainframes, and anything with certificate pinning for later waves. Those need more design work than a 20-day pilot allows.

Match Each Target to the Right Protocol

  • ACME (RFC 8555) for web servers, reverse proxies, and Kubernetes.
  • EST (RFC 7030) and SCEP (RFC 8894) for network devices and mobile device management.
  • REST APIs and configuration management such as Ansible for custom pipelines and applications.
  • Renewal agents for servers and appliances that have no native enrollment client.

Verify Deployment, Not Only Issuance

A renewal is finished when the new certificate is live on the endpoint, not when the CA signs it. After each pilot renewal, confirm that the endpoint serves the new certificate with a complete chain and that the service picked it up without a restart someone had to remember. Then break something on purpose: revoke a credential or block a port, and confirm the alert reaches the right person within the tier you designed in Phase 3.

Exit criteria: every pilot system has completed at least one zero-touch renewal in production; one failure scenario tested end to end; rollback steps documented; application owners sign off.

Phase 5 (Days 76-90): Scale, Operationalize, and Hand Over

Owner: CLM program owner. Supporting: executive sponsor for the Day 90 readout; all owners for the wave plan.

Objective: The platform runs as business as usual, and the rest of the estate has a dated plan.

  • Build the wave plan. Group the remaining certificates by technology and owner, order the waves by business impact, and date them. Put every public TLS certificate on automated renewal before March 15, 2027.
  • Publish runbooks for a failed renewal, onboarding a new application, and emergency mass replacement. Mass replacement is not hypothetical: Google stopped trusting new Entrust-issued TLS certificates in Chrome 131 for certificates with an earliest SCT after November 11, 2024, and affected organizations had to move to a new CA in months.
  • Make the CLM platform the front door. New applications request certificates through the platform by default, so the inventory stays complete without another discovery project.
  • Stream events to your SIEM and put the KPIs below on a dashboard the sponsor can see.
  • Run knowledge transfer so the operations team, not the implementation partner, owns daily work.
  • Hold the Day 90 readout. Compare baseline against today, show the pilot results, and get the wave plan approved.

Exit criteria: wave plan approved with dates; runbooks published; KPI dashboard live; the operations team handling day-to-day CLM work.

The 90-Day CLM Plan at a Glance

PhaseDaysAccountable ownerKey deliverablesExit gate
ReadinessBefore Day 1CLM program ownerCredentials, firewall rules, scope, baseline metricsAll access tested; scope signed
1. Deploy and connect1-15CLM program ownerPlatform live, SSO, CAs connected, alert channelsCA inventory imported; security sign-off
2. Discover and baseline16-35CLM program ownerMulti-source discovery, triage buckets, owners assignedTarget ownership reached; remediation dates set
3. Policy and workflows36-55Security / GRCEnforced policy, alert tiers, ITSM integration, reportsPolicy enforced on new requests
4. Automation pilot56-75CLM program ownerZero-touch renewals on 2-4 pilots, failure testPilot renewals verified in production
5. Scale and hand over76-90CLM program ownerWave plan, runbooks, dashboard, Day 90 readoutOperations team running BAU

Certificate Management

Prevent certificate outages, streamline IT operations, and achieve agility with our certificate management solution.

KPIs to Track From Day 0 to Day 90

Measure the same handful of numbers at Day 0, at each phase gate, and at Day 90. The targets below are examples to adapt, not industry benchmarks.

KPIWhy it mattersExample Day 90 target
Certificates discovered vs. known at Day 0Shows how much was invisible beforeFully reconciled, with every new find triaged
Share of production certificates with a named ownerOwnership is the precondition for automation90% or higher
Certificates expiring within 30 days with no renewal planThe direct predictor of the next outageZero
Share of in-scope renewals completed zero-touchProves automation works in production100% of pilot scope
Time to replace one certificate end to endDetermines how fast you can respond to a CA distrust or key compromiseHours, not days
Certificates using weak cryptographyTracks risk and PQC readinessAll have a remediation date
Certificate-related outages in scopeThe number the business cares aboutZero

Common Pitfalls in the First 90 Days

  • Trying to automate everything at once. A 90-day plan that includes code signing, IoT, and every legacy appliance usually finishes none of them. Prove the model on a few systems, then scale in waves.
  • Treating discovery as a one-time event. New certificates appear every week. Schedule recurring scans and route new finds to owners automatically.
  • Automating before ownership. Replacing a certificate on a system nobody owns means nobody tests it afterward.
  • Ignoring private PKI. Internal certificates often outnumber public ones and cause just as many outages. Keep internal CAs in scope from Day 1.
  • Stopping at issuance. A certificate that was issued but never deployed still expires on the server.
  • Alert fatigue. One shared mailbox for every notice means none get read. Tier and route alerts by owner and urgency.
  • No active sponsor. Firewall requests and change approvals stall without someone senior asking why.

Our Take: What We Would Actually Recommend

We have delivered PKI and certificate management programs for enterprises across many industries. This is the advice we give most often in the first week of a rollout.

Treat March 15, 2027 as your real go-live date. The 90-day plan is the vehicle; the 100-day TLS cap is the deadline. Every public TLS certificate that is still renewed by hand on that date will need replacing five or six times a year, and the 47-day schedule in 2029 doubles that again.

Make application owners do the claiming. The PKI team should run the platform, not own every certificate in it. Self-service claiming in Phase 2 is the single biggest predictor we see of whether automation scales in Phase 5.

Report hours and outages, not certificate counts. “We manage 12,000 certificates” does not move a budget conversation. “Renewals that took two hours of engineer time now take zero, and we had no certificate outages this quarter” does.

Start collecting post-quantum data now. The algorithm fields you capture in Phase 2 are free to collect today and expensive to reconstruct later, when the NIST IR 8547 deadlines arrive.

Do not wait for a perfect inventory to begin automating. Phases 2 and 4 can overlap for well-understood systems. A clean pilot on your most visible web tier builds more support for the program than another week of scanning.

How CertSecure Manager Supports Each Phase

CertSecure Manager, Encryption Consulting’s certificate lifecycle management platform, maps directly onto this playbook:

PhaseCertSecure Manager capability
Deploy and connectSaaS, on-premises, or hybrid deployment, plus a PKI-as-a-Service tier with managed PKI operations; connections to Microsoft, public, and private CAs, including AWS Private CA; SSO integration
Discover and baselineContinuous discovery that flags unknown and rogue certificates, including containers, cloud environments, HashiCorp Vault, Azure Key Vault, and Java KeyStore
Policy and workflowsOrganization-wide enrollment policies, restricted encryption algorithms, FIPS compliance enforcement, multi-level approval workflows, M of N approval for CA access, and ServiceNow ticket-based issuance
Automation pilotACME, SCEP, EST, REST APIs, and Ansible; renewal agents for servers, load balancers, and internal applications; integrations with F5 BIG-IP, Apache, Nginx, IIS, and Tomcat
Scale and operateExpiry alerts, tickets, and incidents; events streamed to Splunk, Datadog, and OpenTelemetry; audit trails and on-demand high-risk certificate reports

If your team needs hands-on help with readiness, CA design, or the pilot, Encryption Consulting’s PKI Services and PKI Assessment teams can run the phases alongside you.

Conclusion

The first 90 days with a CLM platform are an organizational project as much as a technical one. Assign one owner per role, finish readiness before Day 1, build an inventory from more than one source, write a policy the platform can enforce, and prove zero-touch renewal on systems that matter before scaling. Teams that follow that sequence reach Day 90 with fewer surprises, a dashboard the sponsor trusts, and a dated plan to have every public TLS certificate on automated renewal before the 100-day cap arrives on March 15, 2027.

The most useful thing you can do this week is fill in the owner table with real names and send it to your sponsor. Everything else in this playbook moves faster once those names exist.

Want to see how the phases look in practice? Start a 15-day free trial or book a CertSecure Manager demo, or estimate your own savings first with the CertSecure Manager ROI calculator. For background on the fundamentals, see our Education Center guide to certificate lifecycle management and our post on moving from manual certificate tracking to automation.

Frequently Asked Questions

How long does a CLM implementation take?

Most enterprises can deploy a CLM platform, build a complete certificate inventory, enforce policy, and prove automated renewal on a few production systems in about 90 days. Full automation across the estate usually continues in waves for several months after that, depending on how many certificate types, CAs, and legacy systems are in scope. Readiness work such as service accounts and firewall approvals is the most common cause of delay.

Who should own a CLM implementation?

A single CLM program owner, usually the PKI lead or a security engineering manager, should be accountable for the plan, schedule, and phase gates. An executive sponsor such as the CISO unblocks approvals. PKI administrators, infrastructure owners, application owners, Security or GRC, and the ITSM owner each own specific deliverables. Naming one person per role, rather than a team, prevents tasks from stalling between groups.

What should we automate first after deploying a CLM platform?

Automate first where a certificate outage would cost the most, such as customer-facing web tiers and load balancers, and where the platform supports the target natively through ACME, agents, or a direct integration. Choose two to four pilots with engaged owners and easy rollback. Leave code signing, IoT devices, mainframes, and systems with certificate pinning for later waves, because they need more design work.

Do we need a complete certificate inventory before automating renewals?

No. You need complete discovery and confirmed ownership for the systems you automate, not for the entire estate. Well-understood systems can enter the automation pilot while discovery continues elsewhere. You should still finish organization-wide discovery within the first 90 days, because unknown certificates are the ones most likely to expire without warning and cause an outage.

How does the CA/B Forum 47-day schedule affect a CLM rollout?

CA/B Forum Ballot SC-081v3 cuts the maximum validity of publicly trusted TLS certificates to 200 days from March 15, 2026, 100 days from March 15, 2027, and 47 days from March 15, 2029. Each step multiplies renewal volume, so a CLM rollout should prioritize automated renewal for public TLS certificates and set alert tiers that leave enough time to fix a failed renewal before expiry.