The AI read: a plain-English second opinion on every verdict
MailCull's AI read turns a deterministic email-verification result into plain English: a summary, a recommended next step, and a confidence read. The engine still decides; the AI just explains it.
On this page · 4 sections
Most email validators hand you a one-word verdict and a number. "Risky, 62." Then it's on you to figure out what that actually means for your next send.
MailCull already shows its work on every check: the SMTP reply, the MX record, the confidence score, the full signal chain. Now there's one more option when you want it: the AI read.
What the AI read is
Open any check, click Analyse with AI, and about three seconds later you get a plain-English briefing alongside the evidence:
- A summary: what the result means, in one or two sentences.
- Signal: the single strongest reason behind the verdict.
- Next: a concrete recommended next step for your campaign.
- Why: a read on the confidence number, so you know whether it's a true confirmation or a hedge.
- A recommendation: Send, Send with caution, Do not send, or Manual review, with the reasoning.
It reads the exact same signals you can see on screen. Nothing more, nothing hidden.
What the AI read is not
This is the important part. The AI does not decide the verdict. The verdict is deterministic: the same address gets the same answer every time, produced by the validation engine, not a language model. The AI sits on top of that result and translates it. If the engine says undeliverable, the AI explains why it's undeliverable; it can't talk it into being deliverable.
That separation is deliberate. You should never have to wonder whether your "valid" result came from a coin flip. The engine is the source of truth. The AI is the colleague who reads the receipt and tells you what they'd do.
When to use it
Use it on the calls you're unsure about: the catch-all domains, the role inboxes, the results sitting in the middle of the confidence range. Skip it on the obvious ones. It's optional, one click, and you're never charged for a read that fails to generate.
What it costs
The AI read is included on every plan: 25 reads a month on Free, 500 on Pro. No separate add-on, no per-call surprise. When you're out, the verdict and the full evidence chain are still right there. The AI read is the bonus, not the product.
Paste an address, get a verdict, and when you want the human take, ask the AI.
Start with 500 free validation credits. No credit card.
Both Free and Pro run the same scan engine — full SMTP probe, MX lookup, typo, disposable, domain checks, and the evidence chain on every verdict. The difference is the monthly credit pool (Free=500, Pro=100,000) plus Pro's API and MCP access.