The inquiry that has sat unread for three days: AI-powered lead qualification

The scene you know

Nils (fictitious name) runs a small IT consultancy, seven people. Business is going well, almost too well: between the contact form, LinkedIn and referral emails, fifteen to twenty inquiries now come in every week. Nils used to read them himself between calls, sort them quickly and pass them on. That worked when it was ten a month.

Now a Monday-morning inquiry is still sitting unread on Wednesday afternoon, wedged between a follow-up on an old invoice and a cold-outreach email. When Nils finally opens it, it is a company with a clear budget and a project that fits perfectly. But the prospect has already followed up twice asking if anyone will get back to them, and in the end went with someone else.

The bitter part is not that Nils is lazy. He reads every inquiry carefully, and that carefulness is exactly what costs time. Five minutes of sorting times twenty inquiries is an hour and a half a week spent sorting, not selling. And the one inquiry that really matters looks exactly the same in the inbox as the nine that do not.

Why this is a growth problem, not a diligence problem

The obvious answer is more discipline, checking the inbox more often. That misses what is actually happening here. Qualification is a repeating pattern: does the company size fit, is there a real budget signal, which offer matches the inquiry, how urgent does the text sound. This exact kind of sorting is well suited to hand to an agent that applies the criteria consistently, so they do not come out differently on a hectic Monday than on a quiet Friday.

What does not get delegated is the relationship. The agent does not decide whether you want to work with someone, and it does not run sales conversations. It reads, sorts, and tells you where to look first. The decision of who you call and what you offer stays with you and your team. What gets automated is the pre-sorting, not the sales work.

The reason this tips over exactly at the growth stage is simple: when ten inquiries came in a month, doing it by hand was not a problem, it was even an advantage, because you personally read every one. At twenty a week, that same diligence becomes a bottleneck. Not because you got worse, but because the volume outgrew the point where a person can still keep pace.

Where growth actually gets lost

The damage does not show up as an incident anyone talks about. It seeps away at three points, a little every week.

What people try that does not hold up

• ✕Everything into one shared inbox, „someone will read it“. Without clear ownership, in the end nobody feels truly responsible, and it is exactly the important inquiries that sit the longest, because they need more attention than a quick reply.
• ✕A form with ten required fields to filter right at submission. That reads like a job application and scares off exactly the good, urgent inquiries that have no patience for a form interview before they have even spoken to anyone.
• ✕A Zapier trigger that dumps every new row straight into the CRM. That tidies things up technically but sorts nothing. You end up with a neatly filled CRM full of unsorted inquiries, which needs just as much reading work as the inbox did before.
• ✕A fixed round-robin rule that hands inquiries out in turn. That is fair to the team but blind to fit: an inquiry about a highly specialized topic lands with whoever is up next, not with the person who has the matching experience.

How it is solved: the agent reads and pre-sorts, you decide

The solution is not another form and not one more tool someone has to log into separately. It is a workflow that sits between your existing lead sources and your team: a new inquiry comes in through the contact form, an email inbox or a CRM webhook, the agent reads it, sorts it against your criteria and sends a short notification to the right person, with an assessment of why this inquiry deserves priority right now, or does not.

Under the hood runs a workflow in n8n that walks through clearly defined steps: fetch the new inquiry, read the text and metadata, check it against your criteria with a language model (company size, budget signals, urgency, fit with your offer), file the result in a structured way and notify the person responsible. Each of these steps is its own traceable tool, not a black-box prompt that returns something.

You set the criteria together at the start, not the agent. What makes a good lead for you, which signals count as urgent, which offer fits which inquiry. The agent then applies those rules consistently, the same every day, no matter how full the calendar is. Edge cases where the sorting is not clear-cut get flagged instead of silently filtered out, so nothing important disappears without anyone noticing.

All results also land in Airtable or Google Sheets, which gives you an overview as a side effect: how many inquiries came in, how were they sorted, how long until the first reaction. That is not reporting for its own sake, it is the basis for sharpening the criteria after a few weeks of real use, once it becomes clear where the sorting is still off.

What matters is what the agent deliberately does not do: it does not reply to leads automatically, it does not decide who wins the deal, and it does not replace a sales conversation. It takes the reading and pre-sorting off your plate, so the person with the sales skill spends their time where it actually counts: talking to the leads worth talking to.

When you do NOT need this

• If you get fewer than five inquiries a week. Then doing it by hand takes a few minutes and the build does not pay off yet.
• If one person already reads, sorts and forwards every inquiry immediately, and it reliably works. Then you would be automating something that already runs smoothly.
• If your inquiries are so varied that no recurring criteria emerge. Then that is a criteria problem first, not an automation problem, and it needs sorting out before the workflow.
• If your team has not yet agreed on what you are actually qualifying for. Then the nicest workflow will not help, because it only consistently executes what was unclear beforehand.

How we build this concretely

This is a tailored automation workflow: a lead qualification agent that spots new inquiries, sorts them and notifies the right person. We start with your most common lead source and the standard case, so you can feel how it works in real daily life before we build out further sources and finer criteria.

The intent behind it: the agent takes the reading and pre-sorting off your team's plate, not the decision of who you want to work with. It prepares, you decide. And you set yourselves which criteria count and which cases always get flagged instead of sorted automatically.

This is a starting point, not a rigid package. Just reach out, then we look together: some things can be simplified, others expanded, all of it tailored to your sales process.

Common questions

Does the agent decide which leads we pursue?

No. The agent reads, sorts and notifies, the decision of who you call and what you offer stays with you. It takes the pre-sorting off your plate, not the sales conversation.

How does the agent know what a good lead is?

You set the criteria together beforehand, for example company size, budget signals, urgency or fit with your offer. The agent then applies those criteria consistently, the same every day, whether it is Monday morning or Friday afternoon.

Do we need to switch CRMs for this?

No. Airtable or Google Sheets is entirely enough to start, all results land there in a structured way. Connecting to an existing CRM is possible if you use one, but it is not a requirement for the pilot.

What happens to inquiries that cannot be clearly sorted?

They get flagged instead of automatically filtered out, similar to an approval step. That way nothing disappears silently in the filter just because the sorting was not clear-cut on the first pass.

How long does the build take?

The pilot for your most common lead source and the standard case stands in a manageable timeframe. Further sources, finer criteria and a CRM connection get added step by step afterward, once it is clear they are worth it.