Technology
›
Semiconductor & Chip Design
›
Chip Manufacturing & Tapeout
Yield Engineering
Long-cycle design programs where IP, foundry, and ecosystem partnerships execute against tapeout and market windows.
Example organizations in this space:
PDF Solutions
KLA
Applied Materials
Onto Innovation
This interactive experience is the shipped product itself — the same application code customers run in production, mounted read-only in your browser over a real sample journey. Not a video, not a mockup: because the demo and the product are one codebase, it can never drift from the real thing.
Inside this journey
-
Pre-Sales
Qualify and diagnose before investing in a full evaluation cycle.
-
Qualification
Confirm budget range, decision authority, timeline urgency, and high-level constraints before investing in a full diagnostic cycle.
Qualification Questions
High-level evaluation fit and constraints
- To make the qualification meaningful, do you have production wafer lots available for a side-by-side run with your current fleet, or would testing be limited to non-production wafers?
- Roughly how many wafers or lots could you allocate to a multi-week qualification (this helps us estimate scope and tool time)?
- Which of these constraints are likely to affect an on-wafer qualification? Select all that apply.
- If there are any special constraints or acceptance metrics we should know now, please summarize briefly.
Budget
- Is there an allocated budget range for qualification and a possible tooling purchase?
Authority
- Who is the primary sign-off authority for an investment of this type, and which roles will influence the decision? Select all that apply.
Timeline and next step
- When do you need to reach a go or no-go decision on a qualification project?
- Would you be open to a focused 30-minute discovery to review wafer access, acceptance criteria, and a high-level project plan?
-
Process & Yield Discovery
Map current yield performance, inspection gaps, data sources (inspection, metrology, test), stakeholder roles, and target acceptance criteria.
Discovery Questions
Context and goals for this evaluation
- How often does your production line run qualification tests for new inspection systems?
- Describe your team's primary objective for this evaluation, name the target yield or defect reductions and the critical timeline you are working to meet.
- When did you first notice the current yield delta versus target on this node or line?
- Who owns the decision to change inspection tools or analytics for this line, and who signs procurement approvals?
- Estimate the annual revenue impact of a 1 percentage point yield improvement on this line, choose the closest range.
- Provide the timeline you consider acceptable for completing hands-on qualification, pick one.
Where yield problems show up, and what they cost
- If current inline inspection had flagged the root cause earlier last quarter, how much additional yield would you have recovered and what would that have been worth annualized?
- Quantify how often defects escape inspection and cause rework, scrap, or customer escapes today.
- List the top defect types on this node that are hardest to classify or correlate to process parameters.
- Identify which downstream indicator increases most reliably after an undetected defect escapes, choose one.
- Tell us which teams are pulled into investigations when an unexplained yield loss happens, pick all that apply.
- What single data gap or inspection blind spot would cause you to halt qualification immediately?
Where the current inspection chain breaks down
- Describe the most recent incident when an inspection or classifier failed to flag a defect that later showed up in electrical test, and list the immediate actions taken.
- Who was first alerted by that failure and how long before escalation reached leadership?
- Which inspection mode or tool in your current fleet tends to miss the smallest or newest feature sizes?
- How long does it typically take from suspect defect detection to a root-cause hypothesis being shared with process engineering?
- Explain the classification accuracy threshold your team requires before engineering will act on a new classifier's findings.
- If inspection exports are not available in machine-readable format within two weeks, what happens to your decision timeline?
Hidden costs and timeline risks that change the decision
- Assuming classifier accuracy could improve to 95 percent on novel defect types with fewer than 50 labeled images, what process you run today would you change first?
- Estimate the engineering hours saved per week if false positives fell by half on this line, pick a range.
- Rank the top three cost drivers when a false defect detection forces a deep investigation, select up to three.
- List typical schedule slips you see during recipe development and validation for a new inspection tool.
- Name the single highest-cost integration or validation task you expect during production rollout, pick one.
- Are there contractual or capital hurdles that would prevent you from starting a tool qualification this quarter?
The other routes you are actively weighing
- Compare the options you are actively considering, including staying with the incumbent, an internal build, or a third-party inspection plus analytics partner.
- Which evaluation criteria matter most to you right now, select up to three.
- Do you have a preferred incumbent you expect to requalify with if nothing else changes?
- Identify whether anyone on your team has proposed solving this internally instead of working with an outside vendor.
- Provide the timeline or contractual condition that would keep you committed to your current supplier for this line.
- What would have to be true of your incumbent approach for you to stay with it rather than switch to a new inspection and analytics partner?
Acceptance metrics that make this a clear yes
- Pinpoint the top three metrics that would prove a pilot successful for you and rank them by importance.
- Give the numeric target for defect capture rate you require to consider switching inspection tools, pick one range.
- Choose the acceptable false-detection rate ranges for qualification, pick one.
- Specify how many production wafers or inspection hours you require in a hands-on qualification, pick the closest band.
- Project how quickly, after meeting acceptance criteria, you expect to move to procurement and install, choose one.
- Do the acceptance targets you just specified, if met, allow your team to release purchase approvals or change tool fleet within 30 days?
Operational readiness and hard constraints
- Confirm which data feeds must be available for the qualification to proceed and who owns each feed, pick all that apply.
- Name the owner or role who will serve as single-point contact for data access and API credentials, pick the role.
- Explain the current state of your defect and metrology data, is it labeled, time-stamped, and linked by wafer id?
- Are there regulatory, safety, or cleanroom access approvals that typically take longer than four weeks for new tool work?
- Quantify the headcount and skills you can dedicate to classifier training and recipe development in the next quarter, pick one.
- Would a missing named owner for cleanroom access or data exports stop the project before it starts?
Integration, data flow, and timing to production
- Outline the end to end data flow you expect between inspection, metrology, equipment logs, and test data during a pilot.
- Confirm which export or API methods are available for the systems you listed, select all that apply.
- Rank the integration tasks by difficulty and select up to three that you expect to be the hardest.
- Walk through a typical timeline from sample wafers available to completed qualification, include gating decisions and expected durations.
- Summarize the data security or anonymization constraints that must be met before raw wafer data can leave your network, select all that apply.
- When the pilot shows the required capture and classification numbers, who has authority to approve scaling to production within your organization?
Decisive next steps and timing
- Assuming the pilot hits targets, what internal approvals or purchase triggers must happen before you can sign a statement of work, select all that apply.
- Share preferred timing windows for starting classifier training and tool integration within the next quarter, pick one.
- Choose the procurement or capital route you would use for a new inspection tool, pick one.
- Pinpoint which leaders need to see pilot results and what form of evidence moves them, list title and required deliverable for each.
- Will your organization commit to a defined qualification wafer schedule within 30 days if the pilot meets the agreed acceptance metrics?
-
-
Solution Evaluation
Run hands-on qualification on production wafers to measure defect capture rates and classification accuracy against agreed acceptance criteria.
- desired_state
- current_state
- decision_readiness
- gaps
- stakeholders
- success_criteria
- current_state
- decision_readiness
- desired_state
- success_criteria
- gaps
- stakeholders
- gaps
- desired_state
- decision_readiness
- success_criteria
- stakeholders
- current_state
- decision_readiness
- decision_readiness
- decision_readiness
- decision_readiness
-
Solution Scope
Define hardware, analytics, integration modules, timelines, acceptance metrics, responsibilities, and risk mitigations for qualification and production integration.
Scope Configuration
- Install and commission wafer inspection tools
- Develop and deploy inspection recipes
- Configure defect classification stations
- Rapid few-shot defect classifier training
- Integrate inspection data with metrology logs
- Correlate inspection data with wafer electrical test
- Run production qualification and capture validation
- Deploy root-cause hypothesis engine and dashboards
- On-site operator training and workflow handover
- False-positive reduction and detection tuning
- Migrate historical defect records to platform
- Integrate feedback to process control systems
- Capital financing and incentive qualification support
Scope Questions
Install and commission wafer inspection tools
- Which inspection tool model(s) and probe station interfaces will be installed (list tool IDs or line numbers)?
- How many tools and chassis do you plan to commission on-site during this engagement?
- Where in the fab will each tool be located (fab ID, line, and cleanroom class)?
- Who will provide equipment electrical, compressed air, and exhaust hookups and confirm single-line diagram (SLD) readiness?
- When can we schedule lockout/tagout (LOTO) and tool-floor access windows for physical installation?
- How will you validate tool acceptance on delivery (e.g., vibration checks, kinematic alignment report, tool serial numbers)?
Develop and deploy inspection recipes
- Which process layers and reticle IDs must initial recipes cover (example: poly, active, metal1, via stacks)?
- How many production wafer lots per layer are available for recipe tuning during qualification?
- Who will own recipe sign-off for each layer (name and role on your team)?
- What time budget should we assume for recipe development per layer (days)?
- Which process control limits or acceptance windows from your process control plan should recipes respect (e.g., CD tolerance, overlay thresholds)?
- Indicate any locked recipes or tool recipes that cannot be changed during the engagement and must be integrated instead of modified.
Configure defect classification stations
- Which classifier deployment mode do you prefer for station configuration: local workstation, on-prem server, or cloud-edge hybrid?
- How many parallel classification stations are required to match your wafer throughput targets (units per shift)?
- What file formats and metadata schema must the station accept from the inspection tool (example: proprietary wafer map + JSON with die coordinates)?
- Who is the named owner for classifier configuration and maintenance on your team (name, role, contact)?
- Which defect label taxonomy should the station use (provide the layer-specific defect classes or reference document)?
- Specify the network and security requirements for classifier station integration (e.g., VLAN, firewall ports, certificate provision).
Rapid few-shot defect classifier training
- What minimum classifier accuracy on novel defect classes measured on held-out production dies will you accept as a go/no-go threshold?
- How many labeled examples per novel defect class can you provide during training runs (actual wafer captures or SEM images)?
- Which imaging modalities will you supply for training (optical brightfield, darkfield, SEM images, EDX overlays)?
- Who will perform the initial label validation on training images and provide disputed-label resolution?
- What turnaround time do you require from initial image submission to an updated classifier deployed on a station?
- Specify any regulatory or internal data handling rules for training images (e.g., IP redaction, export controls).
Integrate inspection data with metrology logs
- Which metrology systems and data types must be correlated (CD-SEM, ellipsometer, AFM, film-thickness logs)?
- How frequently must inspection results be joined with metrology records (real-time, hourly batch, daily batch)?
- What identifier will be used to join records (wafer ID, lot ID, die coordinates, mask set ID)?
- Who manages metrology data exports and can provide sample CSV/CSV schema and a recent data extract for mapping?
- Specify data retention and time-range for correlation (e.g., last 6 months of metrology runs, lifetime of a lot).
- Are there any calibration or timestamp synchronization constraints between inspection logs and metrology equipment we should plan for?
Correlate inspection data with wafer electrical test
- Which electrical test outputs must be linked (parametric test logs, yield maps, failing vector IDs)?
- How will electrical test results be exposed for ingestion (file export, API, direct DB access)?
- What wafer traceability fields exist on your test floor to match inspected dies to electrical test results (probe card ID, handler serial, wafer slot)?
- Who from your test engineering team will validate correlation hypotheses between defect density and failure modes?
- Specify the acceptable latency between inspection capture and availability of electrical test data for correlation workflows.
- Indicate any restrictions on sharing failing vector contents or test program source code for root-cause analysis.
Run production qualification and capture validation
- What defined acceptance metrics will confirm successful qualification (specific defect capture rate delta, classifier precision/recall thresholds, throughput targets)?
- How many qualification lots and wafer samples must be processed to satisfy your production acceptance protocols?
- Who is authorized to sign the final qualification acceptance and provide the official production handover document?
- When do you require qualified tools to be available for production (target calendar date)?
- Which KPI reports and evidence artifacts do you require for validation (wafer-level maps, ROC curves, SEM validation images)?
- Are there specific false-positive or false-negative thresholds tied to production release for particular layers or test programs?
Deploy root-cause hypothesis engine and dashboards
- Which dashboards and KPIs do you need on day one (defect density by die, layer heatmaps, trend of defect classes by lot)?
- How frequently should the hypothesis engine rerun correlations between inspection, metrology, and test logs (on-demand, nightly, hourly)?
- Who are the named dashboard consumers (yield engineer, process owner, fab manager) and what access levels do they require?
- What alerting thresholds and notification channels should the platform use when a new root-cause hypothesis crosses confidence thresholds?
- Specify any existing visualization templates or corporate report formats the dashboards must conform to.
- Indicate whether the hypothesis engine needs to preserve audit trails for each automated hypothesis run for regulatory or quality review.
On-site operator training and workflow handover
- How many operators and engineers require hands-on training sessions for inspection and classification workflows?
- Which shift patterns must training cover (day shift only, all three shifts, weekends included)?
- Who will be the named workflow owner responsible for day-to-day operations after handover?
- What operator certification artifacts do you require at handover (checklists, signed runbooks, competency sign-offs)?
- When should shadowing and paired-run sessions be scheduled relative to tool commissioning?
- Specify any language, safety, or cleanroom training prerequisites operators must complete before we deliver on-site sessions.
False-positive reduction and detection tuning
- Which defect classes currently generate the highest false-positive investigation load (list by layer and defect label)?
- How many engineering hours per week can you allocate to iterative tuning and manual review during the tuning window?
- What target reduction in false-positive rate is required to consider tuning successful (percent or absolute volume reduction)?
- Who will adjudicate borderline calls during tuning and provide final labels for retraining?
- Which automated pre-filters or recipe-level thresholds are allowed to be altered to reduce false positives (give layer-specific allowances)?
- Are there any defect classes where false-positive suppression is not permitted due to safety or yield risk?
-
Mutual Commit
Finalize commercial terms, data-access authorizations, warranty and service commitments, and the acceptance criteria that trigger payment or handover.
Agreement Modules
- Purchase Agreement
- Order Form & Payment Schedule
- Software License & Hosting Agreement
- Master Services Agreement (MSA)
- Statement of Work (SOW)
- Service Level Agreement (SLA) & Warranty Addendum
- Acceptance Test Plan and Payment Trigger
- Data Processing and Access Agreement (DPA)
- Change Order Agreement
-
Deployment
Lock readiness facts and configuration values before execution begins.
-
Pre-Deployment Readiness
Capture concrete readiness facts — wafer schedules, cleanroom access, safety approvals, data feeds, and named owners required before execution.
Pre-Deployment Questions
Environment and site access
- Primary deployment site and cleanroom room(s) where equipment will be installed (name the fab site and room). This tells us where to schedule onsite teams.
- Cleanroom access status for the seller's field engineers (badges, escorted access, required training). Choose the current state so we can plan lead time.
- Site facilities readiness for install (power circuit availability, cleanroom load/rack space, and network access for telemetry). State exactly: 'Ready', 'Partial — list constraint', or 'Not ready'.
Data and integration
- Which data feeds will be accessible at deployment start (select all that apply). These feeds are required for qualification and analytics.
- Named owner(s) for data access and approvals (inspection, metrology, test). Provide name, role, and primary contact so we can request access.
- Integration readiness for third-party systems the platform will ingest (is an integration engineer assigned and are scoped API/transfer methods approved?).
People and ownership
- Named onsite deployment owner who will coordinate wafers, tool access, and daily logistics (name, role, phone/email). This person will be our single point of contact onsite.
- Safety, change-control, or EHS approver required before install (select current state). We need the approver to be identified to schedule pre-install checks.
Timing and constraints
- In-scope wafer lots or production line names for initial qualification (list per site). This defines which wafers we will schedule for qualification.
- Blackout windows or restricted production periods in the next 90 days that would prevent install or qualification (select one). If 'Yes', we will request exact dates after submission.
- Earliest available start date for installation and qualification (enter a date or 'TBD'). This sets the deployment timeline baseline.
-
Configuration Details
Lock exact configuration values the deployment team will use — tool recipes, classifier thresholds, integration endpoints, credentials, and validation test plans.
Configuration Details
Environments & Endpoints
- Production inspection instance name (enter the exact instance name the deployment will configure; format example: "prod-us-west-1")
- Inspection API endpoint URL (enter full URL the platform will call; format: https://<host>[:port]/<path> — Default: https://inspection.platform.local/api/v1)
- Integration endpoint type for the buyer's defect-tracking or MES system (select one) — this selects the connector type the integration endpoint will use
Features & Modules (enable/disable)
- Which modules should be enabled for this deployment? (select all that apply)
Mappings & Ownership
- Defect class field name in the buyer's defect database or MES (enter the exact field/key the integration will map to)
- Role that will own providing credential secrets via their secrets manager (enter role name; DO NOT paste secrets — e.g., "Yield Engineer" or "IT Integrations")
Limits & Policies (thresholds)
- Classifier acceptance threshold (probability between 0.0 and 1.0) used to mark a defect class as "confirmed" — Default: 0.95
- Maximum false-detection rate allowed during qualification (percentage). Default: 2.0
Validation & Acceptance
- Minimum defect capture rate required for go/no-go (percentage). Default: 95
-
Deployment Execution
Install equipment, develop recipes, train classifiers, run qualification lots, and execute the rollout plan with clear owners and escalation paths.
-
Go-Live Acceptance
Formal acceptance gate: verify KPI thresholds, false-detection rates, integration stability, and obtain named sign-offs before production handover.
Checklist items
- Receive signed Go‑Live acceptance document from buyer's designated approver
- Deliver final acceptance test (FAT/SAT) report and obtain dual sign‑off
- Validate false‑detection rate test report meets agreed acceptance criteria
- Confirm integration stability test report is complete and approved
- Verify end‑to‑end data traceability and correlation test
- Document and verify rollback plan and restore point
- Obtain site safety and cleanroom approvals required for production handover
- Handover operations runbook and escalation contact sheet with buyer acknowledgement
- Schedule and confirm first production acceptance run with named owners
- Resolve or formally accept mitigation for all critical open issues prior to handover
-
-
Success
Run recurring outcome reviews, track issues and enhancement requests, and maintain a shared ticketing and improvement cadence to sustain yield gains.
Success Reviews
- Go-live Health Check (weeks 1-4)
- First Measurement Review (weeks 4-10)
- 90-day Outcome Review and Incumbent Wind-down
- Quarterly Operational Review
- Annual Outcomes and Improvement Cadence Review
Issues & Enhancements
- Update the ticket backlog with prioritized fixes and target completion dates.
- Establish the remediation timeline and single owner-less task list for any outstanding high-severity items.
- Archive or migrate incumbent system data and record the archive location in the shared workspace.
- Disable write access to the incumbent workflow or formally set it to read-only per the wind-down decision.
- Publish the 90-day outcome packet with metric dashboards, unresolved ticket list, and remediation timeline.
- Operational metrics and trends
- Confirm ongoing operational health by reviewing classification accuracy and mean time-to-root-cause trends.
- Agree the prioritized list of operational fixes and enhancements for the next quarter.
- Ensure ticket burn-down is progressing and set corrective steps if it is not.
- Re-confirm success criteria and owners
- Schedule the next classifier retraining or recipe freeze window based on prioritized fixes.
- Publish a one-page operational health snapshot to the shared workspace after the meeting.
- 12-month outcomes summary
- Validate whether the solution sustained the agreed yield improvements and false-detection targets over 12 months.
- Document persistent systemic issues and agree a mitigation or investment plan for each.
- Set the year-ahead maintenance schedule, classifier audit cadence, and ticketing SLAs.
- Publish the annual outcomes report with metric dashboards and the agreed maintenance calendar.
- Schedule the annual classifier audit and any required recipe validation lots.
- Formalize ticket triage rules and target SLAs for the upcoming year in the shared operations playbook.
- Confirm deployment components (tools, data feeds, integrations) are live and functioning to the baseline checklist.
- Identify and log the top 3 operational blockers with resolution dates in the shared ticketing system.
- Agree immediate remediation actions and the timing for the first measurement meeting.
- Publish and circulate the go-live health summary to the shared workspace within 24 hours.
- Log all identified blockers in the shared ticketing system and assign target resolution dates.
- Schedule targeted troubleshooting sessions for any critical integration or recipe failures discovered.
- Present measured data vs targets
- Determine whether defect capture rate and classifier accuracy are trending toward the Solution Evaluation stage targets.
- Produce a short remediation plan with discrete tasks and resolution dates for any metric gaps.
- Confirm the timeline and prerequisites for the acceptance gate in the onboarding window.
- Publish the first-measurement report, including per-lot metrics and exemplar images, to the shared workspace.
- Create classifier retraining and recipe-adjustment tasks in the ticketing system with target dates.
- Schedule the follow-up diagnostic session to validate corrective actions before the acceptance window.
- Present 90-day outcome data
- Confirm which outcome metrics meet the Solution Evaluation stage targets and which require remediation.
- Complete the incumbent wind-down checklist, ensuring data archive or migration and closure of dual-work workflows.
- Deployment and integration validation
- Root-cause diagnosis for metric gaps
- Persistent root causes and systemic risks
- Document per-metric status and remediation items
- Ticket and enhancement backlog review
- Incumbent system wind-down checklist
- Ticketing and improvement cadence assessment
- Agree corrective actions and timelines
- Prioritize next-quarter fixes
- Early adoption and usage signals
- Open issues, risk register, and remediation timeline
- Open issues and blockers
- Agree year-ahead maintenance and improvement plan
- Confirm timeline to acceptance gate
- Integration stability and data feed health
- Immediate remediation actions
- Shorten meeting if no changes
- Agree next cadence and owners for ongoing tracking