Quick Check — Where we start together

• In one short paragraph, describe the specific yield gap, defect event, or KPI that prompted this conversation now.
• Which process node and product line are we focusing on for this evaluation? Options: Leading-edge node (≤7nm), Mature node (8–28nm), Legacy node (>28nm), Multiple lines, Not sure / hybrid
• How long has this gap or unexpected yield drop been visible in your metrics? Options: This week, This month, 1–3 months, 3–6 months, More than 6 months
• Roughly how many wafers per week or month would be involved in a side-by-side qualification? Options: <100 wafers/month, 100–500 wafers/month, 500–2,000 wafers/month, 2,000–10,000 wafers/month, Volume unknown
• Who on your team should we credit as the primary point-of-contact for technical and logistical details during discovery?

Why is this keeping you up at night?

• If this defect trend or ramp delay continues for two more quarters, what are the top business consequences we should know about?
• Can you estimate the revenue, margin, or customer-impact tied to the current yield shortfall? Options: > $50M/year, $10M–$50M/year, $1M–$10M/year, < $1M/year, Unknown / estimating
• Who internally is most pressured to resolve this, and what does success look like to them personally (short answer)?
• How would you describe the emotional tone: urgent firefight, measured concern, exploratory, or comfortable to wait? Options: Urgent — needs immediate action, Concerned — accelerated timeline, Exploratory — gathering options, Comfortable — longer runway
• Have you experienced a near-miss or customer escalation tied to these defects in the last 12 months? If yes, tell us what happened. Options: Yes — major escalation, Yes — manageable near-miss, No, Prefer not to say

What are you assuming that might be wrong?

• Which capabilities do you believe your current inspection and analytics stack cannot deliver? Options: High-sensitivity capture at small features, Rapid classifier training for new modes, Cross-tool correlation with metrology/e-test, Actionable root-cause hypotheses, Scalable data pipelines, Other
• What evidence supports those assumptions (logs, missed defects, past experiments)? Please point to any recent examples.
• How long do you expect it will take your team to identify a new defect mode today and push a classifier to stable accuracy? Options: <1 week, 1–2 weeks, 2–6 weeks, 6–12 weeks, Longer / unknown
• If you had to rank the weakest link in your current defect-detection workflow, which is it? Options: Capture sensitivity, Classifier speed/accuracy, Data correlation with e-test, Process/repeatability of recipes, Operational capacity
• Is there anything your team believes an external partner could never do better than the incumbent tools or in-house work? Explain briefly.

If we could deliver a measurable win, what would it feel like?

• What is the minimum yield uplift (absolute) that would make this project 'worth it' for you? Options: ≥ 2% absolute, 1–2% absolute, 0.5–1% absolute, <0.5% absolute, Not sure
• Beyond percent yield, which outcome matters most: fewer escapes to downstream test, lower scrap, faster ramp, or improved cycle time? Options: Fewer escapes, Lower scrap, Faster ramp, Improved cycle time, Other
• How important is classifier accuracy on novel defect types vs. speed-to-classify when deciding success? Options: Accuracy >> Speed, Accuracy > Speed, Balanced, Speed > Accuracy, Speed >> Accuracy
• Who on the customer side will need to say 'this is a success'—and what exact metric would make them say it?
• Tell us about a past win where an analytics or inspection change made a clear business difference. What indicators did you watch?

How will you validate — the bar for saying yes

• What are the non-negotiable acceptance criteria for a side-by-side qualification (examples: classification ≥ X, capture recall ≥ Y, false positive rate ≤ Z)?
• Statistically, what minimum sample size or wafer count will you require to be confident in the comparison? Options: <50 wafers, 50–200 wafers, 200–1,000 wafers, >1,000 wafers, Statistical plan to be defined together
• How long must a pilot run to satisfy your engineering and quality gates? Options: <2 weeks, 2–4 weeks, 1–3 months, 3+ months, Depends on results
• Which metrics will you want in the dashboard during the pilot (select all that apply)? Options: Defect count by class, Capture recall/precision, Classifier confidence score distribution, Correlation with metrology/e-test, False positive rate, Time-to-root-cause hypothesis
• What reporting cadence and format do you prefer for pilot checkpoints (daily standup, weekly scorecard, executive review)? Options: Daily standup, Twice-weekly updates, Weekly scorecard, Milestone reviews only, Ad-hoc as issues arise

Who needs to be in the room when we make the call?

• Which stakeholders must approve the pilot and final purchase (select all that apply)? Options: VP Process Engineering, Yield Management Director, Fab Operations Manager, Quality/Metrology Lead, Finance/Capital, Procurement/Legal, Other
• What is the approval timeline you are operating under for a pilot decision and for capital procurement if the pilot succeeds? Options: Immediate (<30 days), 30–60 days, 60–120 days, 120+ days, Unknown
• Who owns the budget for pilot-related expenses (equipment time, sample wafers, engineer hours)? Options: Process Engineering, Yield, Fab Ops, Central Innovation/CapEx, Shared cost center, Not yet decided
• When technical issues arise during a pilot, who will be the primary escalation contact on your side?
• How will success be socialized internally—what audience needs to be convinced beyond the immediate project sponsors? Options: Executive leadership, Operations floor, Customer OEMs, Quality/regulatory, All of the above, Other

What could stop this from moving forward?

• Which of these potential barriers do you consider most likely to derail a pilot? Options: Recipe development takes too long, Insufficient sample wafers, Data access/security constraints, Budget/capital limits, Too many false positives, Integration complexity, Other
• Have you run a similar pilot before that failed to scale? If so, what was the primary reason?
• What mitigation would convince you we can manage the top-3 risks you selected?
• How tolerant is your team of exploratory false positives during early tuning—are we allowed to generate noisy results to accelerate learning, or must accuracy be near-production from day one? Options: High tolerance for noise (learn fast), Moderate tolerance, Low tolerance — must be accurate early, Must be production-grade from day one
• If a cost overrun or schedule slip happens, what contractual guardrails would you expect to see (caps, milestones, rollback options)?

Small first step — what would make you comfortable to start?

• Do you have representative sample wafers, labeled defect examples, or historical inspection datasets ready to share for initial testing? Options: Ready now, Can prepare within 1–2 weeks, Need 2–4 weeks, Not available / requires discussion
• Which data feeds are required and available for correlation during the pilot (select all that apply)? Options: Inspection defect CSVs, Optical metrology (CD, film), Equipment sensor logs, Wafer map history, Final electrical test results, Other
• What access constraints or data treatment (anonymization, IP controls) must we honor to proceed?
• What does an acceptable pilot kickoff timeline look like for you? Options: <2 weeks, 2–4 weeks, 1–2 months, 2+ months, TBD after scoping
• If we agreed on a minimal pilot plan now, what would be the single best next step you’d expect from us this week? Options: Draft pilot proposal and SOW, Schedule technical scoping call, Share NDA and data agreement, Bring possible pilot schedule, Other

Quick Reality Check: What Brought Us Together Today?

• In one short sentence, what's the single most urgent yield problem you're hoping to solve with external help?
• Which wafer generation or process node is this issue affecting right now? Options: Mature production node, New node ramp, Pilot line / NPI, Multiple lines/nodes, Other
• How long have you been tracking this problem? Options: Days, Weeks, Months, Quarters, Longer than a year
• Who on your team will own evaluating an inspection + analytics qualification (title or role)? Options: VP Process Engineering, Yield Management Director, Yield Engineer, Process Integration Lead, Fab Manager, Other
• What outcome would make this conversation feel like time well spent to you personally?

Are You Quietly Burning Margin?

• If your current yield trend continued unchanged for the next two quarters, how would it impact wafer cost or competitive pricing? Options: Minimal impact, Moderate margin pressure, Significant margin loss, Unsustainable—need immediate action, Unsure
• Quantitatively, what is the approximate yield gap versus target (percent or ppm defect difference)?
• How often do yield shortfalls translate into lost business or missed shipment commitments? Options: Never, Occasionally, Sometimes, Frequently, Always
• Tell me about a recent situation where yield underperformance created executive pressure—what happened and who was involved?
• Which metric matters most to your leadership right now? Options: Overall yield, Defect density per cm2, Wafer start throughput, Time-to-root-cause, Cost per good die, Other

What’s Invisible to Your Current Tools?

• What defect types or process excursions do you suspect are slipping past your inspection fleet today? Options: Sub-resolution particles, Novel pattern defects, Thin-film variations, Overlay/focus drift, CMP residues, Contamination hotspots, Electrical-limited defects, Other
• Which inspection systems are in active use on the affected line(s)? Options: KLA optical inspection, Applied optical/SEM, Nikon/Hitachi optical, Dedicated SEM review stations, In-house custom tools, No inline inspection on this step, Other
• How would you rate your current fleet on sensitivity to new or rare defect modes? Options: Very sensitive, Reasonably sensitive, Limited sensitivity, Blind to novel defects, Unsure
• Give a recent concrete example of a defect mode you only discovered late—what signs were missed earlier?
• How frequently do false positives from inspection tools create wasted engineering investigations? Options: Almost never, Occasionally, Regularly, Often, Every week

When Did Your Last Surprise Happen—and Why Did It Hurt?

• Describe the most recent unexplained yield drop or excursion that cost you meaningful throughput—when was it and how big was the impact?
• How long did it take your team to form a root-cause hypothesis for that event? Options: Hours, Days, Weeks, Months, Still unresolved
• Which data streams did you consult during that investigation? Options: Inline inspection, CD-SEM, OCD or film metrology, Tool sensor logs, E-test / electrical data, SPC / run charts, Wafer history/yield DB, Other
• Was a definitive root cause established? If yes, what was it; if no, what evidence was missing?
• How did that episode change how leadership views inspection and analytics spend? Options: More willing to invest, Cautious but open, Doubled down on incumbents, No change, Other

If Fixing This Was Fast, What Would Change for Your Team?

• What's a realistic, measurable uplift you would be satisfied with from a successful qualification (e.g., percent yield, defect reduction, classification accuracy)?
• Which success criteria will you insist on during a side-by-side qualification? Options: Higher defect capture rate, 95%+ classification accuracy, Faster time-to-root-cause, Lower false positives, Seamless data integrations, Minimal recipe dev time, Other
• How soon would you need to see positive qualification signals before recommending capital purchase? Options: Within weeks, 1–2 months, During a multi-month pilot, Only after full release, Unsure
• What internal KPIs would this improvement positively affect (list concrete metrics and owners)?
• If we demonstrated the expected uplift, what downstream organizational changes would you expect (e.g., faster ramps, pricing leverage, headcount shifts)?

Who Really Decides and What Feels Good Enough?

• Who are the formal decision-makers and informal influencers for buying inspection and analytics (names, roles, and their primary success metric)?
• Which stakeholder is most risk-averse about new inspection tech, and what are they most worried about? Options: Production disruption, False positives, Recipe tuning time, Integration complexity, Capital cost, Other
• For each stakeholder, what would 'good' look like at pilot completion (be specific by role)?
• Is there an executive or board-level timeline driving this decision (e.g., pricing deadlines, customer commitments)? Options: Yes—fixed deadline, Yes—flexible, No formal deadline, Unsure
• Who needs to be in the room for a pilot sign-off conversation? Options: VP Process Engineering, Yield Director, Fab Operations Lead, IT/Data Owner, Equipment Procurement, Quality/Metrology, Other

How Ready Is Your Data to Tell the Story?

• Which data feeds are available and routinely captured for the affected wafers? Options: Inline inspection images, SEM review images, CD-SEM measurements, Film metrology (OCD/ellipsometry), Tool sensor logs / MRL, E-test / electrical, SPC / run charts, Wafer yield DB, Other
• How accessible are those feeds for a partner to ingest (APIs, SFTP, direct DB, manual export)? Options: Fully automated API access, Scheduled file drops, Manual exports only, Restricted—requires approvals, Not available
• What historical window of data can you provide for correlation work (days/weeks/lot counts)? Options: Last few lots, 4–12 weeks, 3–6 months, 1+ year, Varies by stream
• Are there known gaps or quality issues in any of these feeds we should plan around (timestamps, missing fields, image compression)?
• Who owns the data access and who will be our day-to-day contact for integrations?

What Would Make You Pull the Trigger Quickly?

• What are the minimum commercial or risk guarantees you need to greenlight a pilot (po clauses, capital holdbacks, trial periods)?
• How long of a pilot do you consider sufficient to validate defect capture and classification? Options: 1–2 weeks, 1 month, 2–3 months, Multi-quarter
• Which integrations are must-haves for the pilot to be meaningful (select all that apply)? Options: Recipe control, Metrology data, E-test correlation, Equipment sensor logs, MES / lot tracking, Yield DB integration, Other
• What classification accuracy and defect capture targets would be considered a clear pass?
• What maximum recipe development time or production downtime would you tolerate for pilot activation? Options: Days, 1–2 weeks, 3–4 weeks, Longer than a month, Zero tolerance

Comfort and Control During a Pilot: What Do You Need?

• What rollback controls or safety gates must be in place before any pilot work touches production wafers? Options: Immediate recipe rollback, Quarantine lots, Test-only wafers, Operator approvals, Audit logs and change control, Other
• What sample wafer types and quantities can you commit for side-by-side runs? Options: Production wafers from lots, Engineering test wafers, Known-fault seed wafers, Blend of production+seed, Unsure
• Who will dedicate time to recipe tuning, image review, and weekly qualification syncs?
• What acceptance test method do you prefer to verify classification (blind review, SEM confirmation, electrical correlate)? Options: Blind SEM confirmation, Electrical verification, Third-party review, Combined approach, Other
• How frequently do you want progress updates during the pilot and in what form (dashboards, executive briefs, daily standups)? Options: Daily standups, 2–3x weekly operational, Weekly executive summary, Bi-weekly, Ad-hoc

Next Steps — A Low-Risk Path to Insight

• Based on this conversation, what would you consider a low-effort next step that helps us prove value quickly? Options: Sample data exchange, Short side-by-side demo, Executive scoping call, Proof-of-concept plan, Other
• Who else on your side should we include in a scoping session to avoid rework later? Options: Process Engineering, Yield Engineering, Metrology, IT/Data, Fab Operations, Procurement, Other
• When would be the earliest practical date to start a light-touch proof step (data ingest or demo)? Options: Within 1 week, 1–2 weeks, This month, Next month, Unsure
• What would make you say yes to that first step—what guarantees, evidence, or reassurance do you need?
• Any final concerns or constraints we haven't touched on that would block moving forward?

Getting Oriented — What’s the single priority we should put on the table?

• In one sentence, what is the single most important yield or production outcome you need solved in the next 3–6 months?
• Which product line(s) or node(s) does this impact most? Options: Legacy mature node, Ramp node (low volume), High-volume production node, Multiple nodes, Other
• How urgent is this from a business perspective? Options: Business critical — immediate action required, High priority — within 3 months, Medium priority — 3–6 months, Low priority — >6 months
• Who will be the primary internal owner(s) for this engagement (role, not name)? Options: VP Process Eng, Yield Management Director, Fab Manager, Yield Engineer, Process Integration Lead, Data/Analytics Lead, Other
• Roughly, what’s at stake if nothing changes (choose best fit)? Options: > $50M/yr, $10–50M/yr, $1–10M/yr, < $1M/yr, Hard to quantify

Starting Small — Walk Me Through the Most Recent Incident

• Can you briefly describe the most recent yield excursion or chronic defect trend you’d like us to help with?
• When did you first notice it and how has the trajectory changed since then? Options: Just noticed (days), Weeks, 1–2 quarters, Longer than two quarters
• Which signals first alerted you (select all that applied)? Options: Inline inspection spike, Metrology deviation, E-test failures, Customer returns/complaints, Process tool alarms, Other
• How confident are you today that your incumbent inspection tools are capturing the defect modes causing this issue? Options: Very confident, Somewhat confident, Not confident, Unsure
• Share one example (anonymized) of a defect mode or wafer map that we should see during qualification.

Are You Comfortable Leaving Yield to Chance?

• If your current inspection/classification misses a new defect mode for another two quarters, what would the operational and financial consequences be?
• How often do you discover novel defect modes that require manual triage and new classifier training? Options: Weekly, Monthly, Quarterly, Rarely
• How long does it typically take your team to stabilize a classifier or rule-set for a new defect mode today? Options: Days, Weeks, Months, Not measured
• Who in your org bears the pain when classification is slow or inaccurate? Options: Yield engineers, Process integration, Fab operations, Product management, Customers, Other
• What would it mean to you emotionally and professionally if we cut that diagnosis time from weeks to hours?

Where Hidden Defects Live — Tell Me About Your Current Signals

• Which inspection tools and vendors are in your current fleet (list top 3 by usage)?
• What metrology and test data streams can we correlate during qualification? Options: CD metrology, Film thickness, Ellipsometry, E-test wafer maps, Tool logs/recipes, Environmental sensors, Other
• How accessible are those data streams for a short-term pilot (days to weeks)? Options: Full access via API/DB, Manual export only, Requires IT project >4 weeks, Not available
• Describe the wafer types and process layers you’ll want us to inspect during qualification (e.g., BEOL metal, via, CMP, STI).
• On average, what defect density and distribution are we likely to see on the sample wafers you’ll provide? Options: Very low (sparse), Low, Moderate, High

What’s Getting in the Way of Faster Root Cause?

• When you try to accelerate root-cause, what single operational constraint blocks you most often? Options: Insufficient labeled images, Tool recipe variability, Lack of correlated data, Organizational approvals, Budget/capital, Other
• How many labeled defect images (approx) do you have today for the defect types you care about? Options: <50, 50–200, 200–1000, >1000, Unknown
• Have past vendor pilots failed because of recipe development time, or some other integration issue? Tell us a specific story.
• If recipe tuning is a bottleneck, which resources are available internally to accelerate it? Options: In-house recipe engineers, Vendor service team, Shared script libraries, None currently
• How comfortable would you be letting us run an initial tuning session on non-production wafers to speed pipeline readiness? Options: Very comfortable, Somewhat comfortable, Uncomfortable, Depends on controls

What Would Production-Grade Confidence Actually Look Like?

• Beyond headline accuracy, what specific quantitative acceptance criteria matter to you (e.g., capture rate, false positive rate, classification recall)?
• Which of these metrics do you require as a pass/fail for qualification? Options: Defect capture relative to incumbent, Classification accuracy (per-class), False positive rate, Time to root-cause, Correlation with e-test
• Our platform aims for 95%+ classification on novel defects — how would you validate that claim internally? Options: Blind side-by-side run, Third-party audit, Statistical sampling plan, Correlated e-test verification, Other
• What sample size (wafers or defects) would you need to feel statistically comfortable with results? Options: Small pilot (10–20 wafers), Moderate (20–100 wafers), Large (>100 wafers), Prefer vendor recommendation
• Who signs off on qualification results and what governance or committee is involved? Options: Yield Director, Process Integration Lead, Fab Manager, Procurement, Cross-functional review board, Other

If We Could Snap Our Fingers — What’s the Minimal Viable Win?

• If this pilot delivered a single measurable improvement, what outcome would first convince leadership we’re onto something? Options: % yield improvement, Reduction in escapes/customer defects, Faster root-cause time, Lower investigation hours, Better correlation to e-test/metrology
• What minimum % yield improvement would justify a capital expenditure on a new inspection system for this line? Options: ≥1%, ≥0.5%, Depends on revenue per wafer, Need ROI model first
• What’s an acceptable timeline from pilot start to production acceptance in your org? Options: <1 month, 1–3 months, 3–6 months, >6 months
• How important is preserving existing process recipes and tool behavior during qualification (i.e., low vs high risk to production)? Options: Critical — no changes allowed, Prefer minimal changes, Moderate changes acceptable, Open to full tuning
• If we hit the minimal win, what would you expect us to deliver next (scale, integration, handover)? Options: Scale to more lines, Full production install, Integrate with MES/ELN, Knowledge transfer/training, Other

What Would Make Procurement and Capital Say Yes?

• Which procurement model is easiest for your organization for pilot-to-production transitions? Options: Standard CapEx purchase, Leasing/finance, Pilot-as-a-service, Subscription/consumption, Other
• What internal approval gates typically slow inspection tool purchases (select all that apply)? Options: Capital committee, Technical review, Vendor security/IT, Facility constraints, Budget cycle timing
• When does your next capital planning window open? Options: Immediately, This quarter, Next quarter, Later this year, Not on schedule
• What commercial risks do you need to mitigate in a pilot to say yes (e.g., ROI guarantees, limited capital exposure, rollback plan)?
• Would a short-term capital protection (e.g., refundable deposit, success-based payment) make you more comfortable? If so, how much? Options: Yes — small refundable deposit, Yes — success-based payment, No — not necessary, Unsure

How Will This Live Inside Your Fab Day-to-Day?

• Who will own day-to-day operations of the new inspection/analytics stack if we move beyond pilot? Options: Fab operations, Yield engineering, Process integration, Central analytics team, Third-party services, Other
• Which team will be responsible for providing sample wafers, recipes, and tooling access during the pilot? Options: Process integration, Yield engineering, Fab operations, Tool vendor, Other
• What are your expectations for training and knowledge transfer during the pilot? Options: Hands-on onsite training, Remote workshops, Documentation and playbooks, Train-the-trainer, Other
• How tolerant is your fab of false positives during the pilot—how much extra engineering triage can you absorb? Options: Very tolerant, Somewhat tolerant, Low tolerance, Zero tolerance
• What rollback controls must be in place before any inspection changes touch production? Options: Immediate recipe revert, Quarantine wafers, No production access until signoff, Audit trail and approvals, Other

Designing a Side‑by‑Side That Actually Proves It

• How do you typically run side‑by‑side qualifications today and where do they fall short?
• Which acceptance gates would be non-negotiable for you (choose top 3)? Options: Capture parity or improvement vs incumbent, Per-class ≥95% classification, False positive below X%, Demonstrable root-cause trace to process, Integration with metrology/e-test
• Do you prefer blind comparison runs, or paired runs with operators aware of which tool is which? Options: Blind comparison, Paired runs (not blind), Mixed approach, Undecided
• What statistical confidence level would you require to accept pilot results (e.g., 95% confidence)? Options: 90%, 95%, 99%, Prefer vendor recommendation
• Who needs to be present for final acceptance testing and signoff? Options: Yield lead, Process integration, Fab manager, Procurement, Quality assurance, Other

What Would Smooth the Transition to Production?

• Which integrations are must-haves before a production handover (MES, SECS/GEM, ELN, e-test DB, other)? Options: MES, SECS/GEM, ELN, E-test DB, Equipment logs, Cloud analytics, Other
• How long do you expect recipe tuning and validation to take before steady-state production behavior is achieved? Options: <1 week, 1–4 weeks, 1–3 months, >3 months
• What operational KPIs should we monitor during the first 90 days after deployment? Options: Yield trend, Defect density, Classification accuracy, False positives per wafer, Time to root-cause, Other
• What success communication cadence do you want during the pilot and handover (e.g., daily standups, weekly reviews)? Options: Daily standups, Twice-weekly sync, Weekly review, Bi-weekly, Ad hoc
• What ongoing governance will be in place to triage edge cases after handover? Options: Escalation to vendor support, Cross-functional review board, Dedicated on-site team, Remote support SLA, Other

What Would Make Saying Yes Easy — Low-Risk First Steps

• Would you be open to a limited-scope pilot focused on one critical layer or tool to de-risk the program? Options: Yes — single layer/tool, Yes — small wafer set across layers, No — need broader scope
• What minimum commitments (time, wafers, staff) can you realistically provide for a high-quality pilot?
• Which week(s) in the next quarter are best for an initial kick-off and sample collection? Options: Next 2 weeks, This month, Next month, Next quarter, Need to confirm internally
• Who should be our primary day-to-day contact and who is the executive sponsor for decisioning?
• What one concern would prevent you from moving forward right now, and what would need to change to remove it?

Quick Snapshot: The Line We're Talking About

• Which single production line or process node should we focus on for this qualification conversation? Options: N7 / N5 node, N16 / N12 node, Mature node (≥65nm), Legacy high-volume line, Multiple lines / cross-node, Other (please specify)
• Who will be our primary contact for day-to-day qualification coordination, and what role do they play?
• Briefly describe the trigger that brought this to the table right now (pick the closest and add details): Options: New node ramp falling behind target, Sudden unexplained yield drop, Recurring defect mode emerging, Price competitiveness / margin pressure, Routine technology refresh / exploratory, Other (please describe)
• Approximately how many wafers per week flow through this line today? Options: <100, 100–500, 500–2,000, 2,000–10,000, >10,000
• How urgent is resolving this problem from a business perspective? Options: Critical — revenue at immediate risk, High — schedule or contracts impacted, Medium — margins and forecasts affected, Low — proactive/longer-term

If We Do Nothing, What Breaks First?

• What’s the single most consequential thing that will fail if current defect trends continue for another quarter?
• Can you quantify the recent yield impact in absolute terms (wafer starts lost, yield percentage, or estimated revenue)?
• How long has the team been tolerating this condition before escalating it to leadership? Options: Days, Weeks, Months, Over a quarter
• When defects escalate, what downstream activities get delayed or reworked most often (e.g., tape-outs, product ramps, customer shipments)? Options: Tape-out delays, Customer shipments delayed, Rework cycles increase, Metrology backlogs, Other
• If you had to assign a financial urgency level to fixing this problem right now, how would you categorize it? Options: Immediate multi-million impact, Material but manageable, Small impact, Primarily reputational

What’s Invisible Today That Should Make Us Nervous?

• Which defect modes or excursions do you suspect are slipping past your current inspection fleet entirely? Options: Sub-resolution particle defects, Novel morphological defects, Intermittent process excursions, Layer-to-layer correlated defects, We don't know—that's the problem, Other (please specify)
• Give a concrete example of a time you found a root cause late—what was missed initially and how long until it was identified?
• Where do you see the biggest gaps today: defect capture (sensitivity), classification (accuracy), or root-cause correlation with metrology/e-test? Options: Capture sensitivity, Classification accuracy, Correlation with metrology/e-test, Data integration and lineage, All of the above
• How often do false positives from the current inspection tools lead to wasted engineering investigations? Options: Daily, Weekly, Monthly, Rarely
• Who on your team currently owns triage of new defect modes and how do they escalate something unusual?

Who Holds the Keys — and Who Actually Pulls Them?

• If you mapped every decision required to approve a new inspection tool/qualification, where do real bottlenecks or veto points exist?
• Which stakeholders must sign off on a side-by-side qualification and production acceptance? Options: VP Process Engineering / Yield, Fab Manager / Plant Leadership, Equipment Purchasing / Capital, Metrology & Electrical Test, Quality / Reliability, IT / Data Security, Operations / Line Engineers, Other
• For each stakeholder group you've selected, what is their single most important success metric for this project (e.g., dollars saved, ramp schedule, classification accuracy)?
• Who controls the capital and what is the typical approval horizon for equipment purchases at the required scale? Options: Local fab director (weeks), Corporate procurement (months), Capital planning cycle (quarters), Already budgeted, Other
• Who would be the internal champion driving day-to-day work during the pilot, and how much of their time can be allocated? Options: Dedicated full-time, Part-time (~50%), Ad-hoc as issues arise, Not yet identified

If We Could Wave a Wand — What Would Change?

• Imagine yield improved by 1% in six months: what would that unlock for product, pricing, or customer commitments?
• What specific measurable targets would make you call the qualification a success (e.g., defect capture rate lift, classification accuracy, time-to-root-cause)? Options: 95%+ classification accuracy, X% increase in capture rate (please specify), Reduction in time-to-root-cause to hours, Fewer false positives by Y%, Seamless data correlation with metrology/e-test
• What trade-offs are acceptable during qualification—longer pilot time vs higher confidence, or faster pilot with tighter acceptance thresholds? Options: Longer pilot for higher confidence, Faster pilot with stricter pass/fail, Hybrid approach, Undecided
• Beyond metrics, how would you describe the ideal day-to-day experience after deployment (e.g., fewer emergency meetings, automated alerts you trust)?
• Which of these would make your leadership most comfortable signing off on production acceptance? Options: Repeatable capture & classification results, Clear integration with metrology/e-test, Limited capital exposure, Short rollback plan, Demonstrated ROI

What’s Honestly Been Holding Change Back?

• When similar initiatives have stalled in the past, what was the real reason—not the official reason—that they stopped?
• Which of the following constraints are you most worried will derail a qualification? Options: Recipe development time, Capital budget limits, Integration with data pipelines, Disruption to production schedule, Internal politics / competing priorities, Data access/security approval
• How many engineering hours are you realistically able to commit from your team for recipe tuning and investigation during a pilot? Options: >2 FTEs, 1–2 FTEs, Less than 1 FTE, Unclear yet
• What would be a deal-breaker for you in commercial terms or deployment risk? Options: High capital without trial, No rollback plan, Proprietary data lock-in, Unclear acceptance criteria, Long integration timelines
• If you could remove one internal obstacle right now, what would it be?

Data, Recipes, and Experiment Readiness — Can We Run the Test?

• If we tried to start a side-by-side qualification next week, what single logistical or technical roadblock would stop us immediately?
• Do you have representative sample wafers and known-defect wafers available for a side-by-side within the next 30 days? Options: Yes, both representative and known-defect wafers ready, Representative wafers only, Known-defect wafers only, No, need to collect samples
• What data feeds do we need to integrate for meaningful correlation (select all that apply)? Options: Inspection defect lists, Overlay/metrology, Process tool sensor logs, EDA/test chip electrical results, Tool recipes / lot history, Other
• Who will own data access and security approvals on your side, and how long does that typically take?
• How much recipe tuning lead time do you expect will be required before we can run production-like wafers on our system? Options: <1 week, 1–2 weeks, 2–4 weeks, >4 weeks

Decision Gate: What Would Make This a Clear Yes?

• What are the non-negotiable acceptance criteria you would require to approve production acceptance after qualification? Options: Minimum capture rate (please specify), ≥95% classification accuracy, Demonstrated root-cause correlation with metrology/e-test, No increase in false investigations, Integration with existing MES/data systems, Defined rollback controls
• What pilot duration would you consider sufficient to validate those criteria? Options: 2–4 weeks, 1–3 months, 3–6 months, Depends on result stability
• What level of capital exposure or financial commitment is acceptable to start a pilot (e.g., loaner tool, capital purchase, subscription)? Options: Loaner/tool-on-loan, Small capital deposit, Subscription/consumption model, Full capital purchase required, Need to discuss
• Which go/no-go gates should be in place during the pilot (technical, schedule, financial)? Please list the minimal gates you require.
• Who on your side will make the final go/no-go sign-off and what evidence do they insist on seeing?

Emotional Check: How This Is Landing on Your Team

• On a scale from calm to crisis, where does this issue sit for your leadership team right now? Options: Crisis — immediate action required, High concern — elevated attention, Manageable — watching closely, Low priority
• How confident are you that your current vendors or internal tools will detect and classify a new, novel defect mode quickly? Options: Very confident, Somewhat confident, Not confident, Unsure
• What would keep you up at night if we switched inspection platforms and something unexpected appeared?
• What emotions—relief, skepticism, excitement—do you expect from the extended team if the pilot shows early success? Options: Relief, Excitement, Skepticism, Indifference, Other
• Who do you trust inside or outside the company to validate the technical claims (benchmarks, classification performance)? Options: Internal metrology team, Independent test lab, Customer-facing engineering, Third-party consultant, Other

Small Steps That Create Momentum — Practical Next Moves

• What is one low-risk experiment you would be willing to commit to in the next 30 days to build momentum? Options: Run 1–2 known-defect wafers side-by-side, Share historical defect data for a dry-run analysis, Authorize a week of recipe tuning, Schedule stakeholder alignment meeting, Other (please specify)
• Who needs to be in the room for that first experiment to be meaningful (names, roles)?
• What is the earliest practical date you could make the wafers and access available for that experiment? Options: Within 1 week, Within 2–4 weeks, In 1–2 months, Later / need discussion
• What would you need from us to make that first experiment feel low-risk (e.g., loaner hardware, data-only pilot, strict rollback plan)? Options: Loaner hardware, Data-only analysis, Short-duration on-site pilot, Written rollback & risk plan, Other
• After this conversation, what is the single most important follow-up we should do to keep momentum? Options: Send proposed pilot plan & timeline, Schedule technical kickoff, Provide a data ingestion checklist, Prepare commercial pilot terms, Other