Most incident response contracts read the same way. Twenty-four-seven coverage. Rapid response. Experienced team. Every vendor pitch deck uses the same four adjectives, and none of them tell you what actually happens the night your EDR lights up at 2 a.m.
The gap between what an IR contract promises and what it delivers usually shows up only at the worst possible moment. By then you’re not evaluating a vendor. You’re stuck with one.
The good news: you don’t need to wait for a breach to find the gap. You just need to ask better questions before you sign. Below are ten. Ask them of any firm you’re evaluating, including us. The answers will tell you more about the engagement than any slide deck will.
1. What actually starts your SLA clock, and what happens if you miss it?
“20-minute response” and “1-hour response” sound like commitments, but often they’re marketing jargon. The question that matters is what triggers the clock and what it measures.
Does the clock start when you page the vendor, or when a human acknowledges the page? Does “response” mean a qualified practitioner begins triaging your environment, or a ticket gets opened and someone calls you back to schedule a kickoff? CrowdStrike’s well-known 1-10-60 framework (detect in one minute, investigate in ten, contain in sixty) is widely cited as an industry benchmark, but it describes a technical detection target, not a contractual guarantee most firms will put in writing. Ask your vendor to define the clock in the contract, not the pitch, and ask what remedy exists if they miss it.
2. Is the SLA about response, or about resolution?
A fast phone call means nothing if containment still takes days. IBM’s 2025 Cost of a Data Breach Report puts the global average breach lifecycle at 241 days: roughly 181 days to identify a breach and 60 more to contain it once found. That average includes many organizations with a retainer in place. A retainer that guarantees someone will pick up the phone quickly, but says nothing about time to containment, is solving the easy 10% of the problem.
Ask what your prospective firm measures and reports on: response time, containment time, or both. If they only talk about the first call, that’s the answer.
3. What do you already know about our environment before day one?
This is the single biggest driver of how a breach plays out. A firm that has never seen your cloud architecture, identity provider, or logging setup has to map your environment while actively trying to stop an attacker. That’s hours, sometimes days, spent on context instead of containment.
Ask directly: do you maintain any ongoing visibility into our environment, or does the engagement start from a blank page the moment we call? Some firms offer a genuinely pre-emptive model, where they build readiness assessments and environment context continuously so the responder isn’t starting cold. Others are pure break-glass: dormant until the breach, then learning your systems live. Both models exist in the market, and both can be legitimate, but you should know which one you’re buying, because the difference shows up in hours, not features.
4. What happens to unused retainer hours?
Most IR retainers are sold as pre-purchased hours. The uncomfortable question is what happens to the hours you don’t use. Some vendors run “use it or lose it,” which quietly incentivizes you to hope for a bad year to get your money’s worth. Others let you roll hours into tabletop exercises, readiness assessments, or threat hunting, so you get value from the retainer even in a quiet year.
Ask specifically how unused hours are treated, and get it in writing. It’s a small contract clause with an outsized effect on whether the retainer earns its cost every year, or only in the year you’re breached.
5. Are you pre-approved on our cyber insurance panel, or will we find out mid-incident?
If your organization carries cyber insurance, this question can be the difference between a covered claim and a denied one. Insurance carriers often require policyholders to use panel-approved vendors for forensic investigation and incident response, and using an unapproved vendor without prior authorization can exclude those costs from the claim. Approval as a non-panel provider is a formal process, and carriers generally will not fast-track it once an incident is already underway.
Ask your IR firm whether they’re listed on your specific carrier’s panel today, not whether they’ve worked with insurers “in general.” If they’re not on the panel, ask what it takes to get them added, and start that process before you need it, not during the incident.
6. Who actually shows up: the person on the call, or a subcontractor?
Large consultancies and MSSPs often staff incidents from a bench or rotation. The person who sold you the retainer and the practitioner who shows up at 2 a.m. can be two different people who have never spoken. Managed detection and response providers, in particular, tend to detect and escalate, then hand off to a separate forensics team, which means you’re re-explaining your environment to a second group mid-crisis.
Ask who staffs your specific engagement, whether that team is dedicated or pooled across clients, and whether there’s a handoff between detection and deep-dive forensics. A “no handoff” answer, where the same team carries the incident from alert to remediation, is rarer than the marketing suggests.
7. What does your evidence handling and chain-of-custody process look like?
If this incident ends up before regulators, a court, or your cyber insurer’s claims team, how you collect and preserve evidence matters as much as how fast you collect it. Sloppy chain-of-custody documentation can undermine an otherwise solid investigation.
Ask how the firm handles forensic artifact collection, whether documentation is court-ready by default or only produced on request, and how they transfer sensitive evidence, such as malware samples or memory dumps, without exposing it to standard file-sharing risk.
8. Do you run tabletop exercises, and how often should we?
A retainer is not a plan, and a plan that’s never been tested is a hypothesis. NIST SP 800-84 is the standard federal methodology for designing these exercises, and most compliance frameworks, including PCI DSS 4.0’s Requirement 12.10.2, call for testing the incident response plan at least annually. Organizations in regulated or high-risk industries often run tabletop exercises quarterly rather than once a year.
Ask whether tabletop exercises are included in the retainer or billed separately, who facilitates them, and whether the exercise produces a documented list of gaps with an actual remediation owner, not just a slide deck that gets filed away.
9. What happens to our data, and does any of it train a model we don’t control?
AI is now part of nearly every IR platform’s pitch, and that raises a question worth asking plainly: where does our environment data go, and does any of it train a model outside our control? If a vendor’s AI is a licensed third-party model, your telemetry may be flowing somewhere you can’t audit. If it’s proprietary and trained only on the vendor’s own casework, that’s a materially different data governance story.
Ask for a straight answer on model provenance, not a reassurance. Also ask what certifications back up their data handling claims. SOC 2 Type II and ISO 27001 are the two most commonly requested by enterprise buyers, auditors, and insurers, and a vendor should be able to point you to a live trust portal, not a PDF from two years ago.
10. What does your team actually do between incidents?
This question separates a retainer that works every day from one that sits dormant until disaster. Ask what the firm does with your account when nothing is on fire. Do they track your environment’s readiness on any kind of ongoing basis? Do they proactively flag new exposure? Or does the relationship go quiet the day the contract is signed, only to resurface when you page them?
The honest answer to this question, more than any SLA number, tells you whether you’re buying insurance you hope to never use, or a partner who’s actively reducing the odds you’ll need them.
The pattern behind all ten questions
None of these questions are trick questions. They’re the ones a board, an auditor, or an insurer will eventually ask you, after the fact, if your IR provider comes up short. Asking them now, while you still have leverage and time, costs you an afternoon. Finding the answers out mid-breach costs considerably more: IBM’s research shows organizations with a tested incident response plan save an average of $2.66 million per breach compared to those without one.
If you want a structured way to compare providers side by side, our breakdown of pre-emptive IR versus the break-glass retainer model walks through the operational differences in more detail, and our IRT page lays out exactly how our own team answers each of these questions, including staffing, evidence handling, and what happens to your data.
We also explain why real-time alerting and predefined response processes are the real gap in most organizations’ security stack, and what our IR project manager sees from inside a 24/7 global response operation. For the platform-side story of why we rebuilt our own retainer model instead of buying into the break-glass status quo, see Why We Reforged Rapid-IR. Our own certifications are documented at trust.profero.io and in our SOC 2 and ISO 27001 announcement.
If you’re evaluating a retainer, or reevaluating the one you already have, talk to our IR team. Bring the ten questions. We’ll answer all of them.



