Most failed software-delivery engagements were predictable on day one. Not because the vendor lied, but because the buyer didn't ask the questions that would have surfaced the mismatch. The contract looked clean, the rate card was competitive, the references checked out — and six months later velocity was flat, scope was contested, and the relationship was on its third escalation call.
A Virtual Delivery Center engagement isn't different — except that VDC contracts are smaller, faster, and more frequent than traditional outsourcing, which means a single bad question costs you less but a systematic lack of due diligence costs you the same. This is the 12-question checklist we use to help enterprises pressure-test a VDC vendor — including AiDOOS — before signing. It's also the checklist we use internally when evaluating other vendors for our own infrastructure.
For each question: why it matters, what a good answer looks like, what a bad answer looks like, and a one-line example. At the end of the post, a sample response document showing what answering all 12 well actually reads like.
1. How fast can a pod be operational, and what's the SLA on that?
Why it matters: "Time to first commit" is the single best leading indicator of how the rest of the engagement runs. A vendor that takes 90 days to stand up a 5-person pod is signaling friction at every layer — security review, contracting, talent matching, tooling provisioning — that won't disappear once the pod is live.
Good answer: A specific number with an SLA attached. "5–10 business days from scope alignment, with a contractual 14-day backstop." Bonus points if they share the breakdown: NDA + security review (day 1–3), pod composition (day 2–5), tooling and integration setup (day 5–8), first commit (day 8–10).
Bad answer: "It depends on the engagement." Translation: they don't have a repeatable process; each pod is bespoke, which means slow.
One-line example: "AiDOOS typical: 7 business days from signed scope to first PR merged. Contractual SLA: 14 days."
2. What's your replacement clause when a pod member underperforms?
Why it matters: Underperformance is the most predictable risk in any engagement. The replacement clause determines whether handling it is a contract amendment or a platform default.
Good answer: "Pod-level SLAs cover quality. Underperformers are rotated automatically without client-vendor escalation. Replacement happens within the current sprint or the next one. We absorb the rotation cost; you don't pay for the swap."
Bad answer: "We work with you to find a replacement." Translation: the burden's on you to detect the problem, raise it, prove it, and wait through the replacement cycle while billing continues.
One-line example: "Same-sprint rotation, automatic. Last 12 months: 3.2% of pod members rotated, average rotation time 4.7 days."
3. Who owns the IP, and at what point does it transfer?
Why it matters: Most enterprise legal teams will catch this before signing, but the answer reveals how the vendor thinks about your work. A vendor that retains IP "until full payment" is treating the engagement as transactional. A vendor that transfers IP at commit is treating it as embedded.
Good answer: "All IP transfers to you at commit. The pod is a work-for-hire arrangement; nothing the pod produces is retained by the platform or the individual contributors. The contract carries explicit work-product assignment and waiver of moral rights."
Bad answer: "IP transfers on full payment." That's leverage masquerading as a clause. Don't sign it.
One-line example: "Commit-time transfer. No platform retention. Standard work-for-hire with explicit waiver."
4. What does "outcome-based" mean in your contract — milestones, hours, hybrid?
Why it matters: "Outcome-based" is a spectrum. At one end, it means full milestone billing — the vendor only gets paid for shipped, accepted work. At the other, it means time-and-materials with a thin "outcome" wrapper. The middle is hybrid: T&M with milestone bonuses or penalties.
Good answer: "Milestone-based with platform-bench economics. The pod runs at a milestone rate; we absorb bench time. Scope changes recompose the milestone schedule, not the rate. T&M is available as a fallback only when scope can't be defined upfront — which is rare."
Bad answer: "Outcome-aligned T&M with quarterly true-up." Translation: it's T&M and the "alignment" is marketing copy. Compare to true VDC pricing mechanics.
One-line example: "Milestone billing for 90% of engagements. Platform absorbs bench. Rate cards visible at pricing."
5. What governance does the platform impose vs. what we own?
Why it matters: Governance is the single largest hidden cost of any delivery engagement. If you own standups, code-review SLAs, milestone reporting, and quality gates, you're paying for the engagement and the management overhead. If the platform owns it, your engineering managers are freed up.
Good answer: "Platform-managed delivery. Each pod has an embedded delivery manager who runs sprint cadence, code-review SLAs, and milestone reporting. Client owns: scope definition, outcome acceptance, integration approvals, escalation final-call. Everything in between is the platform."
Bad answer: "Flexible — we adapt to your processes." Translation: you'll run it. Adaptive sounds nice; it's actually expensive.
One-line example: "Embedded delivery manager per pod. Standups, code reviews, milestone reports — platform-default. You review outcomes, not tickets."
6. How are technical disagreements escalated and resolved?
Why it matters: Disagreements happen — over architecture, over acceptance criteria, over what "done" means. The escalation pattern reveals whether the vendor has thought about adversarial cases or only happy paths.
Good answer: "Three levels: pod-level resolution (delivery manager), platform-level resolution (engagement architect), and contract-level resolution (legal/exec). Each level has a documented turnaround SLA. Most disputes resolve at level 1 within 48 hours."
Bad answer: "We work it out collaboratively." Translation: there's no defined path, so disputes drift.
One-line example: "L1: 48-hour SLA. L2: 5-day SLA. L3: contract escalation. Last 12 months: 87% resolved at L1."
7. What's the data-handling addendum — and is it co-authored or take-it-or-leave-it?
Why it matters: If you're in a regulated industry — healthcare, financial services, defense, insurance — the data-handling addendum is the single most important clause in the contract. A vendor that hands you a boilerplate "we comply with GDPR" addendum is signaling that compliance is a checkbox, not an operational discipline.
Good answer: "Co-authored during contracting. We start with a baseline NDA and platform-default data-handling rules; we extend to your sector-specific requirements (HIPAA, SOX, FINRA, FedRAMP, regional data residency). The addendum is reviewed by both legal teams before pod activation."
Bad answer: "Standard MSA. We sign your NDA." That's not a data-handling addendum. That's an NDA.
One-line example: "Co-authored addendum, sector-tailored. See industry solutions for the operating-model details."
8. What integrations are native vs. need engineering work to enable?
Why it matters: If the platform claims "GitHub integration" but you discover at day 8 that it's a manual export-import flow, your time-to-first-commit slips. Integration depth is a leading indicator of operational maturity.
Good answer: "Native: GitHub, Jira, Monday, Linear, Slack, Microsoft Teams. Each surfaces pod activity in your existing tooling without extra integration work. Custom integrations (your internal review system, your ticketing) take 2–5 days to wire up at platform expense."
Bad answer: "We support all major tools." Vague claim, no specifics, no timeline. Press for a list.
One-line example: "5 native integrations live. Custom integrations: 2–5 days, no client engineering required."
9. What's the audit trail — who can request what, when?
Why it matters: If you're regulated, your auditors will ask. If you're not regulated, you'll still want it the first time something goes wrong and you need to reconstruct who made what decision when. Audit-trail availability is binary — either it's automatic or it's manual export.
Good answer: "Built-in. Every pod activity (commits, milestone sign-offs, escalations, replacements, scope changes) is logged with timestamps and actor identity. Audit reports are generated on demand by client admins. Retention: minimum 7 years."
Bad answer: "We can run reports for you on request." Translation: it's manual, it's slow, and it's the platform's discretion what to share.
One-line example: "Self-service audit report from client admin console. 7-year retention. Included in platform fee."
10. What's the termination clause — and what happens to in-flight work?
Why it matters: The termination clause determines whether the engagement is reversible. A multi-quarter wind-down means you're locked in for the duration even if it's not working. A milestone-bounded exit means you can pivot at the next milestone.
Good answer: "Milestone-bounded termination. Either side can exit at the next completed milestone with 30-day notice. In-flight work is wrapped to the milestone boundary; nothing mid-feature gets abandoned. No decommissioning fees, no buyout clauses."
Bad answer: "Quarterly termination with 90-day notice." That's a captive-style clause adapted to look modern.
One-line example: "Milestone exit, 30-day notice. No buyout. Last completed milestone is the exit boundary."
11. What's the price escalation rule across multi-quarter engagements?
Why it matters: Most contracts have annual escalation built in. Some are CPI-linked, some are vendor-discretion, some are flat. The structure matters for multi-year procurement planning.
Good answer: "CPI-linked annual escalation, capped at 5%. Rate card visible publicly. Custom rates negotiated only for engagements over 100 pod-months annually, and even then within a documented framework."
Bad answer: "Rates renegotiated annually based on market conditions." Translation: opaque, vendor-favorable.
One-line example: "CPI-linked, 5% cap. Public rate card. See pricing."
12. Who's accountable when the pod misses a milestone — the platform or the talent?
Why it matters: This is the question that separates a true VDC from a staff-aug rebrand. In staff aug, individual contractors are accountable to you. In a VDC, the platform is accountable to you, and the platform manages the pod. If the answer dodges, you're in staff-aug territory dressed up.
Good answer: "The platform is accountable. Missed milestones trigger pod-level review, root-cause analysis, and remediation — at platform expense. The platform either delivers or compensates per the milestone clause. Individual talent isn't your problem; pod outcomes are the platform's problem."
Bad answer: "The pod members work hard but ultimately delivery depends on requirements clarity and dependencies." Translation: missed milestones become your fault.
One-line example: "Platform-accountable. Missed milestone = platform-funded remediation OR proportional credit per contract."
What a "good" answer to all 12 looks like — sample response document
If you're sitting on a vendor RFP and want to see what a clean answer document reads like, here's a compressed version of what AiDOOS would write back. Use it as a benchmark — your final vendor's answers should match this density and specificity, even if the details differ.
| # | Question | One-line answer |
|---|---|---|
| 1 | Time-to-pod-operational | 1-2 business days. 14-day SLA backstop. |
| 2 | Replacement clause | Same-sprint rotation, automatic. Platform-funded. |
| 3 | IP ownership | Commit-time transfer. No retention. |
| 4 | Outcome-based meaning | Milestone billing, platform absorbs bench. Public rate card. |
| 5 | Governance ownership | Platform-managed. Embedded DM. Client reviews outcomes only. |
| 6 | Escalation path | 3 levels with SLAs. 87% resolved at L1. |
| 7 | Data-handling addendum | Co-authored, sector-tailored, both legal teams review. |
| 8 | Integrations | 5 native, custom in 2–5 days. |
| 9 | Audit trail | Built-in, self-service reports, 7-year retention. |
| 10 | Termination | Milestone exit, 30-day notice. No buyout. |
| 11 | Price escalation | CPI-linked, 5% cap. |
| 12 | Milestone accountability | Platform-accountable. Funded remediation. |
If a vendor's answers are this dense and this specific, they've operationalized their delivery model. If they're vague or "we customize per client," they're closer to staff augmentation than a true VDC. The difference matters.
How to use this checklist
Three practical patterns for putting the checklist to work:
- Pre-RFP filtering. Send the 12 questions to candidate vendors before issuing a full RFP. Vendors who answer well are worth the RFP cycle; vendors who can't articulate their answers cleanly aren't.
- Procurement scorecard. Convert the 12 questions into a scoring rubric (good answer = 2, partial = 1, bad = 0). A 24-point total. Anything below 18 fails the procurement gate. Forces objective comparison across vendors.
- Renewal review. If you have an existing VDC engagement up for renewal, run the 12 questions against your current vendor's actual behavior over the last 12 months. The gap between what they promised and what they delivered is the negotiating position for the renewal.
Where to start
If you're inside a VDC procurement decision and want our written answers to all 12 questions for your specific scope, schedule a 30-minute call. We'll walk through the answers, share the actual contract clauses where they exist, and identify where AiDOOS doesn't fit your particular constraints — that honesty up-front beats discovering it three months in.
For broader procurement context, see the VDC vs Outsourcing framework and VDC vs ODC decision matrix. To see what an active engagement looks like operationally, the build-your-VDC page covers the engagement lifecycle.