Start Here: a quick network snapshot

• How many network devices are in scope for this initial evaluation, broken down by routers, switches, and firewalls? Options: <100 devices, 100–499 devices, 500–2,499 devices, 2,500–9,999 devices, 10,000+ devices
• When was your most recent severity-one outage caused by a configuration change, and what was the primary failure mode?
• Tell me about the teams who make production configuration changes, their size, and typical shift patterns
• Estimate the weekly hours your engineers spend on manual CLI changes versus time spent troubleshooting post-change incidents Options: <10 hours, 10–40 hours, 41–120 hours, 121–300 hours, 300+ hours
• Select the primary outcome your leadership will judge this project on Options: Reduce outage minutes, Achieve audit compliance baseline, Reduce manual change hours, Accelerate feature rollout, Other

When a change breaks the network: incident realities and accountability

• If a single misconfiguration caused 50,000 subscribers to be down for four hours, who in your organization would be held accountable and what immediate actions would they take?
• Give a concrete incident example from the last 18 months where an intended configuration change produced unexpected forwarding-plane impact, include the change, scope, and remediation timeline
• On average, how long from detection to full restoration after a configuration-induced outage? Options: <30 minutes, 30–120 minutes, 2–6 hours, 6–12 hours, 12+ hours
• Who first detects these incidents today, NOC, field engineers, monitoring, or customer reports, and how are they escalated? Options: NOC, Field engineers, Automated monitoring, Customer reports, Other
• Which metrics does leadership track for these events, for example customer minutes impacted, SLA penalties, or incident cost Options: Customer minutes impacted, SLA penalties, Support and overtime cost, Churn indicators, Regulatory reportables

Where configuration risk concentrates and why it matters

• Which hidden configuration gaps in your estate are most likely to cause the next large outage, by region, vendor, or device role
• How many devices are currently flagged as drifting from approved baselines, provide percentage or a device count if possible Options: <1%, 1–5%, 6–15%, 16–30%, 30%+
• List the specific artifacts you use as the approved baseline, for example canonical templates, vendor images, or golden running configs
• Which automated compliance checks do you run now, and which gaps generate the most false positives
• If a pilot found 30 percent of devices non-compliant, what would have to change internally for you to proceed with a phased automation rollout

How you validate changes today, and where testing breaks down

• Why do you still rely on manual smoke testing for broad config changes instead of automated pre-checks, staged rollouts, and rollback automation
• Do you have a lab that mirrors production for core routing and edge aggregation, including multi-vendor interoperability tests Options: Yes, full parity for core and edge, Partial parity, missing some vendor or scale elements, We have only simulation or limited testbeds, No dedicated lab
• Who owns the validation test plan and is that owner enabled to sign off on rollouts
• Which of the following validation checks do you perform after a change, BGP neighbor establishment, route convergence, ACL verification, or forwarding-plane validation Options: BGP neighbor establishment, Route convergence checks, ACL and policy verification, Forwarding-plane traffic tests, Performance and latency checks
• If pre-checks could prevent the top two failure modes you listed earlier, what internal approval or budget decision would be unlocked

Obstacles that would derail automated rollouts

• What single operational constraint today would force you to pause any automated change rollout immediately
• Which maintenance windows are acceptable for staged rollouts, nights, weekends, or limited windows during business hours Options: Night windows only, Weekend windows preferred, Limited business hours windows, Ad hoc, case-by-case
• Who must authorize emergency rollback in production, and how quickly can they be reached
• Is there any regulatory or contractual prohibition on automated changes or on granting vendor or third-party access to device configurations Options: Yes, strict prohibitions, Some restrictions requiring approvals, No known prohibitions
• If you discovered a vendor device lacked API access and required manual intervention, would that stop the pilot or require a custom integration plan Options: Stop the pilot, Proceed with custom integration plan, Workaround with limited scope

Competitive landscape, what options are you weighing

• Why would you choose to keep doing manual changes or extend a vendor-specific automation tool instead of standardizing on a vendor-neutral platform
• Which alternatives are you actively evaluating, including internal tooling, vendor-specific automation, or broader IT automation platforms Options: Internal build, Vendor-specific automation, Generic IT automation platform, Open-source tooling, Other
• For any incumbent or vendor-specific option you consider, what must be true about its multi-vendor coverage or validation to keep you from switching
• Has anyone proposed an internal build to solve automation, and if so, what is the projected time to a working pilot Options: No internal proposal, Yes, 3–6 months, Yes, 6–12 months, Yes, 12+ months
• What evaluation result or metric would make you decide to stay with your current approach and not run an external pilot

Operational readiness and integration constraints

• If APIs, credentials, or lab parity are not available within 6 weeks, can your team still run the evaluation on schedule Options: Yes, with adjustments, No, schedule slips, Only with additional vendor support
• Which systems must we integrate with, NMS, ticketing, CMDB, device inventory, CI CD, or others Options: NMS, Ticketing system, CMDB, Device inventory, CI CD pipeline, Other
• Who owns API tokens and device credentials today, and are those stored in automated vaults or handled manually Options: Centralized vault with RBAC, Partially centralized, Manual handling by engineers
• What is the headcount you can commit in full-time equivalent engineers for the pilot Options: No dedicated FTEs, 0.5–1 FTE, 2–3 FTEs, 4+ FTEs
• Are there compliance approvals, legal reviews, or change board gates that must be satisfied before device access or production testing Options: Yes, strict gating, Yes, but administrative, No
• If a required integration points to a third-party system that has no API, would you be willing to grant temporary SSH access for the pilot or is that a deal blocker Options: Willing to grant temporary access, Requires a custom integration plan, Deal blocker

Success metrics, acceptance, and next steps

• If the pilot proves a measurable reduction in configuration-caused outages, what would realistically stop you from signing a production deployment agreement within 30 days
• Select the primary acceptance metrics we should measure in the lab pilot, mean time to detect, mean time to restore, percent of successful rollbacks, compliance delta Options: Mean time to detect, Mean time to restore, Percent successful rollbacks, Compliance delta, Number of manual changes removed
• Who will be the decision maker with authority to sign off on pilot success and commit budget for rollout Options: VP Network Operations, Director Network Engineering, CIO/CTO, Procurement with exec sponsor
• What budget or procurement constraints determine the latest date you can commit to a purchase decision
• If the lab pilot achieves the agreed acceptance criteria, when could you begin a staged production rollout, immediately, within 1 to 3 months, or later Options: Immediately, 1–3 months, 3–6 months, Later or unsure
• Which single missing approval or condition would prevent you from moving to production even if the pilot met all technical criteria

• Solution Walkthrough: Intent-Based Automation

• Confirm the current state and its cost to your team
• You confirm the demonstrated workflow prevents the specific manual-change outage scenario from Discovery.
• Seller prepares a mapped lab runbook that ties each buyer scenario to the exact pre-checks, staged rollout steps, and validation checks to be executed in the demo.
• You agree that the pre-checks and validation checks in the demo would catch the edge cases you worry about in production.
• Map your scenarios to the intent model
• Buyer provide the lab test plan and a list of representative devices with OS versions and access method for the demo.
• Execute the lab pre-check and staged rollout for the scenario
• Buyer define the acceptance criteria metrics for the lab runs, including rollback timeout, acceptable convergence time, and pass/fail conditions.
• You accept a concrete set of lab acceptance criteria and the next evaluation runs required to prove multi-vendor coverage.
• Demonstrate automatic rollback and remediation
• You identify any vendor feature gaps that must be resolved before production rollout.
• Seller produce a short gap log after the demo listing any vendor integrations or feature gaps requiring custom work and estimated effort.
• Confirm multi-vendor coverage and integration gaps
• Validate that this maps to your needs
• Agree remaining evidence and next evaluation steps

• Solution Walkthrough: Intent-Based Automation
• Solution Experience Deck
• Solution Brief - Intent-Based Automation Walkthrough

• meeting
• slides
• document

• Vendor-Neutral Intent Model Templates
• Convert CLI Runbooks to Intent Templates
• Automated Provisioning Workflow Deployment
• Pre-Change Validation and Pre-Check Rulesets
• Staged Rollout and Canary Deployment Orchestration
• Automatic Rollback and Snapshot Configuration
• Closed-Loop Change Validation Tests (Lab)
• Multi-Vendor Device Adapter Development
• Compliance Baseline Enforcement and Remediation
• Continuous Compliance Scanning and Reporting
• Forwarding-Plane Convergence Verification Scripts
• Auto-Remediation Policy Tuning and Thresholds
• Engineer Hands-On Automation Training

Vendor-Neutral Intent Model Templates

• Which device families in your control plane and core edge inventory should be modeled as intent templates (for example: core routers, PE routers, aggregation switches)?
• How many distinct service intents (for example: BGP PE-to-PE connectivity, segment-routing L3VPN, access VLAN provisioning) do you need templated for the evaluation? Options: 1-3, 4-10, More than 10
• Who is the owner of the canonical single-line diagram (SLD) or network topology document we should use to derive the intent templates?
• When an intent template changes, what configuration elements must be treated as immutable versus parameterized (examples: ASN, loopback addresses, VRF names)?
• Provide the expected template input parameters you require for each service intent (for example: peer ASN, neighbor IP, route-targets, interface list).

Convert CLI Runbooks to Intent Templates

• List the existing CLI runbook procedures (by name and last-updated date) you want converted into intent-driven workflows.
• Specify typical runbook preconditions we must encode as pre-checks (examples: no active BGP flaps, interface admin-up, CPU < 70%).
• Identify any runbook steps that require manual confirmation from your network operations center (NOC) before automated execution. Options: No manual steps, One manual confirmation required, Multiple manual confirmations required, Custom - describe
• Describe the canonical CLI snippets or configuration fragments that must be parameterized versus those that should remain verbatim in the template.
• Confirm whether archived CLI runbooks are stored in a versioned repository (for example a git repo, change ticket system) and provide access method. Options: Yes - git, Yes - ticket archive, No - not versioned, Other

Automated Provisioning Workflow Deployment

• Which provisioning scenarios do you prioritize for the initial deployment (examples: new VRF creation, subscriber access VLAN assignment, MPLS label distribution)?
• How many devices will participate in the first staged provisioning run (provide a count split by device role: core, aggregation, access)?
• Specify the integration endpoints for orchestration actions (examples: network management system API, ticketing system webhook, CMDB read endpoint).
• Identify any credentials or vault method you require for provisioning (examples: vault path, role-based service account, SSH key rotation cadence). Options: Central secrets vault, Local device credentials, Hybrid - describe, Not determined
• Provide your desired rollout schedule characteristics for provisioning workflows (for example: batch size, quiet hours, inter-batch delay).

Pre-Change Validation and Pre-Check Rulesets

• Specify the exact pre-change checks you require for production-affecting changes (examples: BGP neighbor state is Established, route count within expected range, interface counters below threshold).
• Identify the authoritative sources for pre-change state (examples: live routing table, telemetry feed, syslog events, netflow samples).
• List the acceptable device health thresholds that must be met before a change proceeds (examples: CPU < 70%, memory < 80%, control-plane process up for 24h).
• Confirm whether you require a dry-run validation that renders the post-change desired config and a diff against the running-config before commit. Options: Yes - dry-run with diff, No - no dry-run, Optional - depends on change type
• What failure thresholds or evidence will define a pre-check rejection that stops execution (for example: BGP not Established, more than 5% route delta)?

Staged Rollout and Canary Deployment Orchestration

• Provide the staging strategy you prefer for canaries (examples: single device, single site, percentage of population, per-population by role). Options: Single device, Single site, Percentage-based, Role-based batches, Custom - describe
• How many incremental stages do you want for a typical rollout (examples: canary -> small batch -> regional -> full)? Options: 1 (canary only), 2-3 stages, 4+ stages
• Identify metrics to watch during a canary (examples: BGP session count, forwarding-plane drop rate, traffic loss percentage) and the telemetry source for each.
• Specify the automatic pause or manual approval points you require between stages (examples: automatic pause for 30 minutes, manual sign-off after canary). Options: Automatic pause only, Manual approval required, Hybrid - describe
• List the outage impact tolerance for canary failures (examples: rollback on any BGP flap, allow transient flap < 30s, allow 0.1% traffic loss).

Automatic Rollback and Snapshot Configuration

• Which snapshot cadence do you require for config backups prior to changes (options: immediate pre-change, hourly, daily, both running and startup configs)? Options: Immediate pre-change, Hourly, Daily, Both running and startup, Custom - describe
• Describe your preferred rollback triggers and their thresholds (examples: persistent BGP down > 30s, forwarding-plane loss > 0.5%, control-plane PID crash).
• Identify whether rollback should be automatic, automatic with notification, or manual with operator confirmation for production changes. Options: Automatic, Automatic with notification, Manual confirmation required
• Specify how configuration snapshots must be stored and retained (examples: encrypted object store with 90-day retention, versioned git repo with signed commits).
• Confirm the maximum acceptable time-to-rollback from trigger detection to completed rollback in production. Options: < 30 seconds, 30-120 seconds, 2-10 minutes, Custom - describe

Closed-Loop Change Validation Tests (Lab)

• Provide a lab topology reference or SLD for the validation tests including device roles, link capacities, and test traffic generators.
• Which test cases must be executed in the lab to mirror your production scenarios (examples: BGP session bring-up, LDP/segment-routing label distribution, multicast join and leave)?
• Who on your engineering team will observe and sign off on lab test runs and where will you capture test evidence (examples: pcap, route table dumps, telemetry graphs)?
• List the test traffic profiles and volumes the lab should generate to validate forwarding-plane behavior (examples: 1 Gbps IMIX, ramp to 10 Gbps, BFD session churn).
• What measurable acceptance criteria will confirm the closed-loop validation passed in the lab (examples: BGP convergence < 60s, 0 packet loss for control-plane management traffic, forwarding-table delta < 0.1%)?

Multi-Vendor Device Adapter Development

• Identify the device OS families and firmware ranges that currently lack adapter support and require development.
• Specify access methods for each device family (examples: NETCONF, gNMI, SSH CLI, RESTCONF) and whether out-of-band management exists.
• Describe any custom configuration constructs or local templates on devices that an adapter must translate (examples: proprietary ACL macros, custom TCL scripts).
• Confirm whether you can provide lab devices or device images for adapter development and functional testing. Options: Yes - physical devices, Yes - virtual images, No - we need vendor-supplied test access, Other - describe
• List the telemetry and state artifacts an adapter must surface to the platform (examples: RIB/adj-RIB-ins/out, interface counters, BGP neighbors).

Compliance Baseline Enforcement and Remediation

• Which compliance frameworks or standards govern your configs (examples: internal approved baseline, NERC CIP control list, ISO network configuration guidelines)?
• How many baseline templates do you maintain and where are they stored (examples: central S3 bucket, git repo, ticketing attachments)?
• Specify the remediation actions you permit for noncompliant devices (examples: auto-apply baseline, create a remediation ticket, quarantine device). Options: Auto-apply baseline, Open remediation ticket, Quarantine device, Manual remediation
• Identify audit frequency and reporting cadence you require for baseline compliance (examples: real-time alerting, daily report, weekly executive summary). Options: Real-time alerts, Daily report, Weekly summary, Monthly audit
• Describe exceptions process for devices that intentionally diverge from baseline (examples: documented exception in CMDB, temporary change ticket ID).

Continuous Compliance Scanning and Reporting

• Which telemetry streams will be available for continuous scanning (examples: syslog, streaming telemetry, periodic config snapshots)?
• Provide the set of reports you require out of the box (examples: noncompliant devices list, drift trend by device group, remediation time-to-fix).
• Specify role-based access to compliance reports and whether exports to your ticketing system are required. Options: RBAC required, Exports to ticketing required, Both, None
• Identify acceptable false-positive tolerance for automated compliance alerts (for example: allow up to 5% false positives before requiring manual review).
• Describe your retention and archival requirements for compliance audit artifacts (examples: keep snapshots for 1 year, signed change logs for 7 years).

Forwarding-Plane Convergence Verification Scripts

• Provide the convergence metrics you need verified after a change (examples: route propagation time to all peers, next-hop reachability, traffic microburst recovery time).
• Which test endpoints or traffic flows should be used to validate forwarding-plane convergence (examples: probe IPs, service prefixes, telemetry collectors)?
• List the scriptable artifacts we can use in the lab and production for verification (examples: traceroute scripts, BFD session monitors, flow collector queries).
• What evidence will validate forwarding-plane convergence to your standards (examples: traceroute showing expected path within X ms, no packet loss across N probes)?
• Confirm acceptable measurement windows after a change for declaring convergence (examples: 30s, 60s, 5 minutes). Options: 30 seconds, 60 seconds, 5 minutes, Custom - describe

Auto-Remediation Policy Tuning and Thresholds

• Specify the categories of incidents you allow auto-remediation to act on (examples: config drift, interface admin-down, flapping BGP neighbor).
• Which threshold values should trigger auto-remediation for transient faults versus persistent faults (examples: flap count, error-rate percentage, packet-loss duration)?
• Identify notification routes and escalation windows when auto-remediation runs (examples: Slack channel, paging list, 5-minute escalation).
• Describe the rollback-to snapshot policy when an auto-remediation action fails to restore desired state.
• List any change categories that must be excluded from auto-remediation for compliance or safety reasons (examples: firmware upgrade, license changes).

• Non-Disclosure Agreement (NDA)
• Subscription & Order Form
• Master Services Agreement (MSA)
• Statement of Work (SOW)
• Service Level Agreement (SLA)
• Data Processing Agreement (DPA)
• Access & Data-Sharing Agreement
• Acceptance Criteria & Sign-Off Certificate
• Change Order Agreement
• Regulatory Compliance Addendum
• Commercial Summary & Payment Terms

• Go-live Health Check
• First Measurement Review
• Acceptance Gate and Incumbent Wind-down
• Operational Retrospective and Runbook Handoff
• Quarterly Operational Review

• Confirm training plan and a shared channel plus escalation path for ongoing issues.
• Restate acceptance criteria and numeric targets
• Produce a documented pass/fail outcome for each acceptance criterion recorded in Solution Scope.
• Capture a formal acceptance decision with the buyer's named signatory and date, or record the conditional acceptance plan.
• Agree the incumbent wind-down plan or retention terms and schedule any required data migration or archiving tasks.
• Publish the formal acceptance record with per-criterion pass/fail results and the buyer signatory information.
• Create a remediation tracker for any failed criteria with tasks and hard completion dates.
• Document the incumbent wind-down steps and schedule the final decommission or retention review.
• Document lessons learned from deployment and early operations
• Capture a prioritized list of operational lessons and concrete runbook updates.
• Finalize and publish operator runbooks and validation checklists for steady-state operations.
• Re-confirm acceptance criteria and owners
• Publish the operational retrospective document including prioritized lessons and runbook edits.
• Deliver updated runbooks and validation checklists for day-to-day operations and rollback procedures.
• Distribute the training plan and schedule hands-on sessions to validate operator proficiency.
• Present the quarterly metric dashboard
• Confirm whether quarterly metrics meet the Solution Scope targets or require additional remediation.
• Ensure persistent incidents are tracked with remediation owners and deadlines until closure.
• Maintain a prioritized backlog of enhancements and vendor gaps with agreed delivery timelines.
• Publish the quarterly performance report with trend charts and the remediation tracker status.
• Update the enhancement backlog with committed timelines for the next quarter.
• Prepare compliance artifacts required for the next audit window and distribute to auditors or internal stakeholders.
• Confirm all acceptance criteria and owners recorded in Solution Scope are accurate and acknowledged.
• Validate that the initial rollout shows no critical forwarding-plane failures and identify any urgent technical fixes.
• Create a short remediation list with target dates for all high-priority blockers.
• Publish the go-live health summary including open issues and timelines for remediation.
• Collect and share logs and validation artifacts for any failed staged rollout attempts.
• Update the escalation path document and circulate it to on-call teams.
• Present first measurement data
• Establish whether each named metric is trending toward the Solution Scope targets or requires corrective action.
• Document a concrete remediation plan with resolution dates that will bring off-target metrics back in line.
• Confirm whether the onboarding timeline to the acceptance gate remains realistic.
• Publish the measurement report including raw metric calculations and supporting validation artifacts.
• List specific pre-check or test additions required to reduce change failure rate and assign resolution dates.
• Schedule any necessary vendor integration work to close device support gaps identified during diagnosis.
• Deployment and rollout validation
• Runbook and playbook review and sign-off
• Present outcome data against each criterion
• Root-cause diagnosis for gaps
• Persistent incidents and root-cause updates
• Training and competency gaps
• Agree corrective actions and timelines
• Pass/fail determination and signatory capture
• Enhancement and bug backlog review
• Early adoption signals and usage patterns
• Confirm timeline to acceptance gate
• Remediation plan for any failed criteria
• Compliance and audit readiness check
• Monitoring, alerts, and validation checks tuning
• Blockers and open technical issues
• Agree immediate remediation actions
• Agree ongoing shared channel and escalation path
• Short meeting option when stable
• Incumbent system and process wind-down check
• Update risk and issue register
• Close and next steps