Getting the Case on the Table

• Tell me about the matter we're discussing, including the forum type and whether it is litigation, regulatory inquiry, or internal investigation Options: Federal litigation, State litigation, Regulatory inquiry, Internal investigation, Other
• How many custodians and data sources do you expect this matter will touch based on current notice or preservation steps? Options: 1-10 custodians, 11-50 custodians, 51-250 custodians, 251-1,000 custodians, More than 1,000 custodians
• When was the last time your team ran a review at this scale, and what happened that you would change if you could?
• Describe the single outcome that will make your partner or legal operations leader call this engagement a success
• Which internal stakeholders must sign off on case success signals before you can move from intake to pilot Options: Lead litigation partner, VP legal operations, Discovery counsel, Information security, Data privacy officer, Procurement, Other

Where the Deadline Really Hurts

• If the calendar shortens by half, which single timeline failure would force you to request a continuance or bring in emergency help?
• On this case, what are the court-imposed dates you cannot miss, and which deliverable maps to each date
• How quickly do you need collection to begin after custodian approval to stay on your critical path Options: Within 24 hours, 48 to 72 hours, Within 1 week, 2 weeks or longer
• Which downstream activity breaks first when collection or processing slips—review staffing, privilege review, production, or something else Options: Review staffing, Privilege review, Production schedule, Opposing counsel meet-and-confer, Trial preparation, Other
• If a vendor can commit to a guaranteed turnaround for an initial pilot, what internal approval would still be required before you can sign Options: Budget approval, Partner approval, Data access approval, Information security signoff, No additional approvals

How Confident Are You in Defensibility

• Tell me about the last defensibility challenge you faced and what evidence ultimately satisfied the court or opposing counsel
• Which defensibility artifacts do you require from a vendor to feel comfortable defending TAR and processing in motion practice or at meet-and-confer Options: Processing logs and hashes, Sampling and validation reports, TAR training and seed set documentation, Chain-of-custody records, Custodian interview notes, Other
• How important is prior reported precedent validating TAR protocols to your decision, on a scale from 1 to 5 Options: 1, 2, 3, 4, 5
• What single gap in defensibility documentation would make you stop a vendor from participating in this engagement

Where Data and Formats Tend to Betray Plans

• Describe the unusual or proprietary data sources you expect in this matter, including collaboration platforms, ephemeral chat, mobile backups, or archived systems
• Which of the following data types are present and need collection or processing Options: Email (on-prem or cloud), Enterprise chat, Cloud file shares, Mobile device images, Databases / structured exports, Third-party platform exports, Backups / archives
• How many file formats or platform connectors have historically caused processing failures on your cases Options: None, 1-3, 4-7, 8 or more, Not sure
• If we find an unprocessable format during pilot ingestion, what is your preferred escalation path and who must approve a workaround Options: Use custom parser with InfoSec approval, Extract to text-only review, Pause and negotiate scope change, Engage internal IT to export, Other
• Which cost or schedule consequence would kill the deal if addressing format issues required extra work beyond the pilot Options: Significant budget increase, Missed court deadline, Loss of confidence from partner, Inability to retain vendor, Other

Where Projects Actually Stall

• What hidden operational friction have you seen repeatedly that turns a smooth kickoff into a stalled project
• Who typically owns chain-of-custody and data access approvals on your matters, and how quickly do they respond Options: Lead partner, Discovery project manager, IT/data custodian, Information security, Outside vendor, Other
• How often do privilege or confidentiality issues require rework during review, and what currently causes most of those reworks Options: Almost always, Often, Occasionally, Rarely, Never
• If a vendor required named owners for each deployment task to proceed, who on your side would you assign in the first 72 hours
• Which of these failure scenarios would make you stop the engagement immediately Options: Data breach during transfer, Inability to reproduce processing results, Unresolved chain-of-custody, Missed critical production deadline, None of the above

What Other Options Are on the Table

• Which alternatives to an external e-discovery provider are you actively considering right now Options: Keep incumbent vendor, Build internally with existing tools, Switch to another external provider, Use platform-only self-service, Pause and negotiate with opposing counsel
• If you were to stay with your current approach, what would have to be true about its cost, timeline, or defensibility for you not to change
• Has anyone proposed solving this matter without an outside vendor, and if so who would lead that effort and what resources would they need
• Which vendor selection criteria would outweigh price when choosing between the incumbent and a new provider Options: Documented TAR precedent, Faster time to stand up, Better processing accuracy, Named references from similar cases, Stronger security controls
• If the pilot proves a targeted reduction in review hours and a defensible protocol, what internal approval would still block you from moving forward quickly Options: Budget hold, Partner objection, Information security concern, Procurement delay, Nothing would block

Operational Readiness and Constraints

• Which integration points must be available for the engagement to proceed, and who owns each connector inside your organization Options: Email provider admin, Collaboration platform admin, HR systems owner, IT operations, Security operations, No integrations required
• Do you currently have API access, export rights, or admin credentials for the primary platforms we will collect from Options: Yes for all listed sources, Partial access, some sources missing, No, we need vendor-assisted exports, Not sure
• How many dedicated internal hours or FTEs can you assign to intake and data mapping in the first two weeks Options: None available, Less than 10 hours, 10-40 hours, More than 40 hours
• Are there regulatory, privacy, or contractual approvals that could prevent data export or cross-border transfer on this case Options: Yes, approvals required, No known approvals required, Unknown, need to verify
• If any of the readiness constraints above are unmet, which single missing item would stop you from running a pilot within 14 days

Defining Acceptance Criteria and Success Signals

• Tell me what measurable signals will prove to your partner that this engagement is succeeding, for example precision, recall, time to first production, or cost per reviewed document
• Which of the following acceptance metrics do you require in a pilot to recommend enterprise adoption Options: Recall threshold, Precision threshold, Reduction in review hours, Faster time to first production, Processing accuracy by format, Security audit passed
• If a pilot meets your technical thresholds but increases short-term cost, which outcome would still justify proceeding Options: Reduced long-term cost, Clear defensibility gains, Ability to meet court deadlines, Improved scalability for future matters, Other
• If a pilot demonstrates the target savings and defensibility, what is the fastest internal route to a signed engagement and who must sign it
• Which acceptance condition would be a deal stopper even if all other metrics are met Options: Unresolved chain-of-custody, Insufficient recall, Security noncompliance, No named operational owner, Other

Pilot Scope and Early Decisions

• Which representative data set would you prefer for a pilot to prove processing and TAR workflows—small focused custodian set, cross-department sample, or a time-bounded export Options: Small focused custodian set, Cross-department sample, Time-bounded export (e.g., 6 months), Random sample across custodians
• How many reviewers and what level of partner or supervisor oversight will you commit to during a pilot Options: No reviewers from buyer, 1-3 reviewers with partner oversight, 4-10 reviewers with senior counsel oversight, Dedicated review team and daily partner check-ins
• Who will be the decision owner for pilot acceptance and how quickly can they provide a go/no-go after results are delivered
• If the pilot proves a 30 to 50 percent reduction in review volume while meeting recall targets, what internal step would accelerate signature to a week
• Which logistical constraint would prevent you from starting a pilot in the next 30 days Options: No budget allocated, No named operational owner, Data access restrictions, Vendor onboarding policies, None

Immediate Next Steps and Decision Mechanics

• Who do we need to meet next on your side to finalize pilot scope, and are they available within the next 72 hours
• Which procurement or security checks do we need to complete before a pilot contract can be issued Options: Procurement review, Security questionnaire, Insurance review, SLA negotiation, None
• How soon could you mobilize data owners and provide the first export for a pilot once access is approved Options: Within 24 hours, 48 to 72 hours, Within 1 week, Longer than 1 week
• If the pilot validates the expected savings and defensibility, what would prevent you from moving immediately to a full engagement Options: Budget constraints, Partner approval, Procurement timelines, Security concerns, Nothing would prevent
• Please list any other constraints, approvals, or expectations that we have not covered that would change how you evaluate a pilot or vendor

• Solution Experience: Collection-to-Production Walkthrough

• Confirm the current state
• You confirm the meeting's restated current state and accept the quantified cost of the problem.
• Run a pilot processing of the provided representative sample and deliver sampling metrics, TAR precision/recall estimates, and chain-of-custody artifacts before the follow-up session.
• Provide a representative dataset sample, the case success signals (key custodians, critical date ranges, and defensibility constraints), and the expected production deadlines.
• Confirm what this is costing you
• You confirm that the proposed collection, processing, and TAR workflows eliminate the primary sources of rework, cost spikes, and defensibility gaps you described.
• Walk through the collection and processing plan for your case
• You agree on the remaining evidence needed to move from evaluation to a commitment, including pilot metrics and acceptance criteria.
• Agree on measurable acceptance criteria for the pilot, including TAR thresholds, production SLAs, and evidence required for court defensibility.
• Prove the TAR-assisted review for your scenario
• Show production workflow and chain-of-custody evidence
• Validate the future state

• Solution Experience: Collection-to-Production Walkthrough
• Solution Experience Deck
• Solution Brief

• meeting
• slides
• document

• Forensic Custodian Data Collection
• Cloud and SaaS Data Extraction
• Mobile Device Forensic Imaging
• Data Processing and File Normalization
• Indexing, Deduplication, and Email Threading
• Analytics-Assisted Clustering and Concept Search
• Technology-Assisted Review (TAR) Training
• TAR Continuous Scoring and Model Tuning
• Managed Document Review and Staffing
• Privilege Identification, Redaction, and Logging
• Production Formatting, Bates Stamping, and Delivery
• Chain-of-Custody Documentation and Audit Trail
• Native File Handling for Specialized Formats
• Rapid Review Environment Standup (48-hour)
• Trial Exhibit Preparation and Export

Forensic Custodian Data Collection

• Provide the custodian roster format you can supply (for example CSV/spreadsheet with name, email, device list) or indicate if you need a template. Options: CSV/Spreadsheet provided, Need template from us, No roster yet
• How many custodians are in scope and what is the estimated aggregate mailbox or user data size in GB or TB? Options: 1-10 / <50GB, 11-50 / 50-250GB, 51-200 / 250GB-1TB, 200+ / >1TB
• Which collection methods are permissible for these custodians: remote logical collection, forensic disk image, mail export (PST/EML), or onsite collection? Options: Remote logical collection, Forensic disk image, Mail export (PST/EML), Onsite physical collection
• Indicate existing legal-preservation artifacts we should reference (for example hold notice ID, preservation custodian list, litigation hold start date).
• List any custodian constraints that may affect collection (for example BYOD policy, encrypted devices, international custodians with cross-border restrictions).

Cloud and SaaS Data Extraction

• Provide the SaaS sources to extract (for example Exchange Online mailbox, Google Drive folder IDs, Slack workspace names, Box/Dropbox workspaces) and include owner account or workspace IDs.
• Select the required export formats per source (for example PST for Exchange, MBOX for Gmail, Slack JSON export, native files for Drive/Box). Options: PST, MBOX, Slack JSON, Native files (Drive/Box)
• Indicate whether you require point-in-time exports or rolling continuous captures for custodial activity. Options: Point-in-time export, Rolling/continuous export, Unsure — advise
• Specify available integration access for each SaaS source: admin credentials, API integration endpoint, or need assistance obtaining access tokens. Options: Admin credentials available, API integration endpoint available, We need assistance obtaining access
• Identify any cross-border or data residency restrictions that apply to specific SaaS accounts or user data. Options: No restrictions, Some restrictions — will specify sources, Unknown — need assessment

Mobile Device Forensic Imaging

• State how many mobile devices require forensic imaging and list platforms and OS versions where known (for example 8 iPhones iOS 14+, 4 Android 10+).
• Choose the acquisition capability for devices: full physical/chip image, logical extraction (app data/backups), or partial logical only. Options: Physical/chip image possible, Logical extraction only, Depends on device
• Identify whether device passcodes and access credentials are documented in your custodian roster and available for collection. Options: Most passcodes available, Some passcodes available, No passcodes available
• Specify if you require carving of deleted messages, unattached media, and slack/IM app artifacts from device images. Options: Deleted content carving required, No carving required, Only specific apps — will list
• List any device management or MDM constraints (for example managed devices requiring MDM tokens or corporate wipe risk) that we must account for.

Data Processing and File Normalization

• Provide estimated ingest volume and an inventory of container types (for example PST count, ZIP archives, OST, Slack JSON) to be processed.
• Select the processing outputs you require: extracted text, OCR for images/PDF, full metadata fields, and preservation of native files. Options: Extracted text, OCR for images/PDF, Full metadata set, Preserve native files
• Indicate language and OCR needs including specific non-English languages (for example Spanish, Chinese Simplified, German) that require tuned OCR. Options: English only, Spanish, Chinese, Other — specify
• What processing accuracy acceptance criteria will validate completion (for example >99% OCR accuracy on searchable PDFs, <1% failed conversions)?
• List any specialized file types requiring custom normalization (for example CAD .dwg, GIS shapefiles, source code repositories) and provide examples if available.

Indexing, Deduplication, and Email Threading

• Choose your preferred deduplication strategy: exact hash deduplication, near-duplicate similarity, or family-level deduplication tied to email families. Options: Exact hash deduplication, Near-duplicate detection, Family-level (email families)
• State retention rules for deduplication and families (for example retain one parent per family, keep most recent within date range, preserve all in privileged sets).
• Specify email threading requirements including whether attachments should be threaded as separate review items or collapsed with parent messages. Options: Include attachments as separate items, Collapse attachments with parent message, Other — specify
• Indicate which indexing metadata fields are mandatory for search and review (for example From, To, CC, Subject, Message-ID, Sent Date, SHA-1).
• Estimate the expected reduction in reviewer volume from deduplication and threading and indicate if you want these reductions reported in processing metrics. Options: <10%, 10-30%, 30-60%, >60%

Analytics-Assisted Clustering and Concept Search

• Select analytics features to enable: unsupervised clustering, topic modeling, near-duplicate grouping, concept-based search or all. Options: Unsupervised clustering, Topic modeling, Near-duplicate grouping, Concept-based search
• State whether you can provide an initial concept taxonomy, keyword seed list, or privilege markers to seed models. Options: Taxonomy/seed list provided, No taxonomy — need assistance
• Choose desired cluster granularity for reviewers (for example coarse 10-50 clusters, medium 50-200, fine 200+). Options: Coarse (10-50), Medium (50-200), Fine (200+)
• Specify whether analytics should surface results as review tags, custom metadata fields, or reviewer suggestions for TAR training. Options: Review tags, Custom metadata fields, Reviewer suggestions, All of the above
• Provide any known sensitive patterns or legal categories (for example contract numbers, HIPAA PHI patterns) to prioritize in concept modeling.

Technology-Assisted Review (TAR) Training

• Select the TAR protocol you prefer: continuous active learning, seeded sample-and-review, or rank-and-review. Options: Continuous active learning, Seeded sample-and-review, Rank-and-review
• Specify the desired initial training set size for seeding the model (for example 500-1,000 labeled documents, 1,000-5,000). Options: 500-1,000, 1,001-5,000, 5,001+
• Indicate availability of subject matter experts to label seed sets and ongoing training batches (for example dedicated SME(s), part-time SMEs, none). Options: Dedicated SME(s) available, Part-time SMEs, No SMEs — need assistance
• What measurable acceptance criteria will you use to confirm TAR training success (for example recall ≥75% at a specified precision, or precision at top decile)?
• Identify whether you require a documented sampling audit trail that includes seed set labels, reviewer IDs, and timestamped training snapshots for defensibility. Options: Yes — require audit trail, No — not required

TAR Continuous Scoring and Model Tuning

• Specify frequency for scoring and model tuning: continuous, per-review batch, daily, or weekly. Options: Continuous, Per review batch, Daily, Weekly
• List the performance metrics you want reported each tuning cycle (for example precision@k, recall, AUC, training-set stability).
• Choose whether automatic model promotions to reviewers are allowed or require human approval before promotion. Options: Automatic promotions allowed, Human approval required
• Specify a score threshold approach that should trigger sampling or production consideration (for example score ≥0.85 or top X percentile).
• Identify which tuning artifacts you want captured in the audit log for challenge response (for example training iterations, labeled examples, scorer versions).

Managed Document Review and Staffing

• Select your preferred review staffing model: vendor-managed review teams, your internal reviewers, or a blended team approach. Options: Vendor-managed review teams, Internal reviewers, Blended team
• Estimate peak reviewer headcount and concurrency needs during the busiest review period. Options: 1-10 reviewers, 11-50, 51-200, 200+
• Specify required review workflows such as single-pass relevance, multi-pass relevance plus privilege, or issue-coding taxonomy. Options: Single-pass relevance, Multi-pass relevance + privilege, Issue-coding taxonomy
• Indicate whether you want reviewer quality control sampling and reporting and, if yes, provide desired sampling frequency (for example 5% weekly). Options: Yes — will specify frequency, No — not required
• List any restrictions on reviewer access for privileged custodians or matter-specific confidentiality queues. Options: Privileged queues required, No special restrictions

Privilege Identification, Redaction, and Logging

• Specify whether automated privilege-identification rules are required (for example attorney domain rules, privileged header patterns) in addition to manual review. Options: Automated rules + manual review, Manual review only, Automated only
• Select the redaction approach needed: burn-in blackline redaction for images/PDF, native redaction, or metadata redaction. Options: Burn-in blackline (PDF/TIFF), Native redaction, Metadata redaction, Combination
• Choose the privilege-log output format you require for productions (for example CSV load file, detailed narrative entries, or a hybrid loadable format). Options: CSV load file, Detailed narrative, Hybrid
• Indicate whether post-redaction OCR validation and redaction logs per document are required. Options: Yes — require OCR validation and logs, No
• List any special legal privilege categories or confidentiality designations that must be captured as structured metadata in the privilege log.

Production Formatting, Bates Stamping, and Delivery

• Select required production formats: native file production, TIFF with load file, or PDF OCR with load file. Options: Native, TIFF + load file, PDF + OCR + load file
• Provide Bates stamp format and prefix requirements and indicate whether Bates numbering should be per-document or per-page.
• Specify required load file standard(s) for ingestion by opposing counsel (for example Concordance DAT, Relativity load file, CSV) and include version if known. Options: Concordance DAT, Relativity load file, CSV, Other
• List acceptable delivery methods for productions and provide endpoint details if pre-configured (for example secure FTP address, SFTP credentials, or physical drive requirements).
• What production acceptance criteria will validate delivery (for example file count reconciliation, checksum verification, load-file validation)?

Chain-of-Custody Documentation and Audit Trail

• Confirm whether you will provide an existing chain-of-custody template or require a standard form to be used. Options: Existing template provided, Need standard template from us
• Specify which artifacts must be recorded in the audit trail (for example MD5/SHA checksums, collector identity, collection timestamps, device serial numbers). Options: All listed artifacts, Subset — will specify
• Choose the preferred delivery format for chain-of-custody records: PDF evidence logs, CSV export, or continuous audit dashboard access. Options: PDF evidence logs, CSV export, Audit dashboard access
• Indicate whether court-prescribed chain-of-custody formats or third-party compliance requirements (for example regulator-specific logs) apply to this matter. Options: Yes — they apply, No — none apply
• State whether notarized or witnessed collection statements are needed for any custodians or specific devices. Options: Yes — required for some, No

• Master Services Agreement (MSA)
• Statement of Work (SOW)
• Service Level Agreement (SLA)
• Order Form & Fee Schedule
• Data Processing Agreement (DPA)
• Chain-of-Custody & Evidence Handling Addendum
• Data Access Authorization
• Change Order Agreement
• Security & Regulatory Compliance Addendum (conditional)

• Go-live Health Check (Weeks 1-4)
• First Measurement Review (Weeks 4-10)
• Acceptance Gate, Outcome Validation (Around Day 90)
• Operational Monthly Review (First 3-6 Months)
• Quarterly Realization Review (Ongoing Quarterly)

• Publish a resolved-issue report that shows closure evidence and any residual actions for partially mitigated items.
• Produce the formal acceptance record showing pass/fail per criterion, the evidence used, and the named signatory or buyer owner for the decision.
• Publish a remediation schedule for conditional or failed criteria with specific deliverables and completion dates.
• Provide documentation that incumbent data migration or archive is complete and that fallback access has been closed or approved read-only.
• Operational KPI review
• Verify production SLA compliance and mean time to resolve critical issues are within acceptable variance and document corrective steps for any outliers.
• Reduce the critical issue backlog with clear resolution dates and actions, or escalate items that cannot be resolved in the next period.
• Ensure defensibility artifacts for the period are complete and accessible via the shared evidence channel.
• Reconfirm acceptance criteria and ownership
• Update the enhancement backlog with priority scores and estimated delivery windows.
• Verify and archive the period's defensibility artifact package into the agreed shared channel.
• Quarter-to-date realization vs targets
• Confirm whether cumulative SLA compliance and average cost per reviewed document meet the targets in Engagement Scope and identify gap drivers if not.
• Agree a prioritized, time-bound plan for enhancements or process changes to address trend-level issues.
• Ensure support and escalation paths are current and that the buyer has a maintained channel for issues and enhancement requests.
• Publish the quarterly realization report showing SLA compliance, cost per reviewed document, and trend commentary.
• Create a prioritized quarter plan that lists enhancements, expected delivery windows, and acceptance criteria for each item.
• Confirm and document the support channel details and escalation matrix in the shared workspace.
• Confirm the platform environment and data ingestion are complete and chain-of-custody artifacts exist for the initial collections.
• Produce a prioritized list of blockers with owners and target resolution dates for the next 14 days.
• Verify reviewer access and basic workflow execution for initial reviewer cohort.
• Deliver an ingestion validation report that includes evidence of chain-of-custody artifacts and processing logs.
• Publish the prioritized blocker list with owners and target dates for remediation.
• Enable reviewer access for the remaining named reviewers and confirm login success within 48 hours.
• Present first-period performance data
• Determine whether mean reviewer throughput and TAR precision are on track relative to Engagement Scope targets and document variance magnitude.
• Identify root causes for any metric gaps and agree a prioritized remediation plan with target dates.
• Confirm the evidence package and timeline required for the Acceptance Gate meeting.
• Deliver a remediation plan that lists each corrective action, the acceptance evidence to be produced, and resolution dates.
• Provide a sample of TAR-reviewed documents and associated scoring reports for independent spot-checking.
• Publish an updated measurement dashboard that includes reviewer throughput and TAR precision, refreshed weekly.
• Restate acceptance criteria and numeric targets
• Produce a documented acceptance decision that records pass or fail for each acceptance criterion from Engagement Scope and includes a named signatory where required.
• Agree remediation tasks and deadlines for any failed or conditional acceptance items so the project can move to steady-state support.
• Confirm the incumbent system wind-down status and actions required to stop parallel usage.
• Open issues and remediation burn-down
• Trend analysis and long-tail issues
• Present outcome data and evidence package
• Diagnose root causes for gaps
• Deployment and ingestion validation
• Enhancement requests and prioritization
• Document pass/fail per criterion and acceptance decision
• Prioritized enhancements and resolution plan
• Agree corrective actions and timelines
• Early adoption and usage signals
• Support channel and escalation confirmation
• Open issues and blockers
• Confirm timeline to Acceptance Gate
• Defensibility artifact audit
• Remediation plan for any failed or conditional items
• Immediate remediation actions
• Incumbent system wind-down confirmation