In Q1 2024, I committed $12,000 to an annual AI SDR platform contract without asking a single question about email verification. Our first campaign went to 12,000 contacts. 2,760 of them bounced—hard.
A 23% hard bounce rate. For context, healthy B2B outbound operations typically maintain hard bounce rates between 2-5% (Source: Validity, 2024). Our domain reputation took months to recover, and I had to explain to the CEO why our new "AI-powered outbound" produced worse results than the manual process it replaced.
I've been running revenue operations for four years now. In that time, I've made six significant mistakes—totaling roughly $18K in wasted budget—and this was the biggest. I'm documenting them so you don't repeat them.
Here's the thing about AI SDR platforms: the demos are impressive. The AI writes personalized emails, the workflow automation is slick, the meeting booking is seamless. And most buyers evaluate exactly what's in the demo. But the sales engagement platform features that matter most aren't the ones you play with during a demo. They're the quiet infrastructure features: email verifier functionality, bounce rate monitoring, data quality maintenance.
The question everyone asks when evaluating an AI SDR platform is: "How many contacts do you have?" The question you should ask is: "What's your verification process, and what's your median hard bounce rate?"
There's no one-size-fits-all answer—what you should evaluate depends on where your team is. Let me break this down by scenario.
Scenario A: You're Building Outbound from Scratch
If you're starting a new outbound motion—or rebuilding after a failed one—your sender reputation is a blank slate. That's both an advantage and a vulnerability. You have no domain history to protect, but you also have nothing to fall back on if things go wrong. One campaign with a 20%+ hard bounce rate and your sending domain is flagged before you've even gotten started.
In this scenario, focus on:
- Verification depth. Does the platform verify contacts at the individual email level (SMTP checks per address) or at the domain level (MX record checks)? These are very different things. Make them walk you through their exact process.
- Pre-send verification. Does the platform automatically verify emails before sending, or does it learn about bounces after the fact? Pre-send is non-negotiable. I learned this the hard way—my 23% bounce rate happened because verification was a post-hoc process, not pre-send.
- Hard bounce rate transparency. Ask the vendor for their own hard bounce rate data across their database. If they produce a number below 5%, great. If they deflect or give you a range, assume the worst.
I knew I should check these things before signing. But I thought, "what are the odds? It's an established platform." The odds caught up with me. The vendor I chose had validated that domains existed, not that individual inboxes were active. Thousands of contacts in their "verified" database were dead addresses.
Scenario B: You Have Outbound Running and Want to Add AI SDR
If you've been running outbound for a while and things are working—but you want AI SDR to help you scale—your evaluation criteria change. At this stage, you've already got a warm domain and established deliverability. The question isn't whether AI can write better emails. It's whether the platform can maintain your data quality at scale.
Here's the counter-intuitive part: at this stage, you should spend less time evaluating AI capabilities and more time evaluating data maintenance processes. Most teams do the opposite, which is why they get burned.
Focus on:
- Ongoing verification (not one-time). B2B email databases decay at an estimated 20-30% per year (Source: commonly cited industry benchmark from major email data providers, 2024). If a platform verifies contacts once on upload and never re-checks, your list quality will silently degrade month after month.
- Automatic suppression. When a contact hard bounces, does the platform automatically add them to a suppressed list across the entire workspace? Or do you need to manually update blocklists in each campaign? Automating suppression saved us from dozens of campaigns that would've blasted dead contacts.
- Deliverability analytics. The platform should show you hard bounce rate by campaign, by sending domain, by list segment—not just a single aggregate number. This insight is what lets you catch a problem before it becomes a domain reputation crisis.
Using these checks, we've caught 47 potential list issues over the past 18 months. That's 47 campaigns that didn't fire into a dying list. Roughly $8K saved compared to our worst year.
(I should add: if a platform's demo doesn't include a deliverability dashboard, that's a red flag. Every platform can demo AI personalization. Very few can effortlessly demo hard bounce rate trends. There's a reason for that.)
Scenario C: Enterprise Scale (500K+ Contacts)
For larger operations, hard bounce rate isn't an operational detail—it's a contractual concern. At this scale, a 1% difference in hard bounce rate means 5,000 wasted emails per campaign. Throw in the compounding effect of domain reputation damage across multiple sending domains and business units, and small data quality differences can cost six figures annually.
At this level, you should evaluate:
- Data quality SLAs. Can the vendor contractually commit to a hard bounce rate ceiling from their database? Top-tier B2B data providers typically guarantee 90-95% email accuracy. If they won't put it in writing, that's your answer.
- Custom verification rules. Do you have specialized requirements—like catch-all detection, spam-trap identification, or suppression rules specific to your industry? Can the platform accommodate these, or are you constrained to their standard pipeline?
- Integration infrastructure. Email verifier features only matter if they integrate cleanly with your existing stack: Salesforce, HubSpot, Outreach, Marketo. Ask about API reliability, sync frequency, and how verification results are reflected in your CRM records.
I haven't personally operated at this scale—my team sits at around 200K contacts. But I've talked to two RevOps leaders who manage databases in the millions. Their shared warning was simple: the compounding rate of bad email data at that scale destroys the ROI of even the best AI SDR tool.
How to Tell Which Scenario You're In
Here's a quick self-assessment:
- Scenario A: You've sent fewer than 50K emails in the past six months, or you're starting a new outbound motion. Focus on pre-send verification and individual-level address checking.
- Scenario B: You have active campaigns, a warm domain, and you want AI SDR to scale an already-working motion. Focus on ongoing verification and deliverability analytics.
- Scenario C: You manage hundreds of thousands of contacts, have a dedicated RevOps team, or face board-level scrutiny on data quality. Focus on SLAs and custom verification pipelines.
Most teams I talk to think they're in Scenario B when they're actually in Scenario A. There's no shame in that—it just means your evaluation checklist needs to be more rigorous in the basics.
The Checklist I Wish I'd Had
- Ask for the hard bounce rate number. If the vendor can't produce it, or says "it depends," walk away.
- Run your own test. Before signing, send a 1,000-contact test campaign into a segment from their database. Measure the hard bounce rate yourself. Above 8%? Don't sign.
- Confirm pre-send verification. This single feature would have prevented my 23% bounce disaster.
- Verify the verification. Ask them to walk you through the exact steps of how an email address gets verified. If they can't explain it in concrete terms, they don't have a systematic process.
Look, I still use AI SDR platforms. When they're set up with the right data hygiene processes, they're genuinely effective. (I've been spending time with Artisan's AI SDR lately, and their approach to email verification is noticeably more substantial than what I had before.) But no platform can save you from weak evaluation criteria.
An informed customer asks better questions. And better questions lead to better decisions. I'd rather spend 30 minutes explaining email verification to a teammate than commit another $12K to a platform chosen on demo wow-factor.
The best time to check hard bounce rate was before your last purchase. The second-best time is now.
