I recently sat in a boardroom with a company most in the industry consider "well-managed." Their CRM was spotless, the pipeline neatly segmented by stages. Marketing reported MQLs, sales reported SQLs. The Head of Sales proudly showed me a list of "golden leads" - 27 inbound inquiries, prioritized A, B, C. In his mind, it was a victory. In mine, it was a list of missed opportunities, only no one had measured it yet.
When I asked when these people submitted their inquiries, he didn't hesitate: "Yesterday afternoon. This morning, we started reviewing them so we're prepared for contact." You could feel the room expecting a nod of approval because he did the "right thing": research first, then personalization, then the call. The problem is, the market doesn't operate on your schedule, and interest isn't something that patiently waits in your CRM.
Meanwhile, the lead is solving their problem. If someone else offers the next step sooner, you're behind before you even start. The most expensive part of inbound isn't the ad, SEO, or landing page. It's the silence between "Submit" and the first meaningfully relevant response.
Your marketing isn't broken. Your response is.
Most companies today aren't losing leads in marketing. They're losing them in the gap between a submitted form and the first coherent contact. That's where we comfort ourselves with process, statuses, and "tomorrow morning" because it sounds professional. In reality, money isn't lost in an explosion; it leaks - minute by minute.
InsideSales/MIT analysis is old, but it's still brutally relevant in one aspect: the key signal isn't the day of the week or the hour, but the speed of response. In the first few minutes, you have a window when the buyer is still "in context" - they've just thought about their problem, just clicked, just expressed intent. When you close that window and respond an hour later or the next day, you're not continuing a conversation; you're trying to reopen it.
If you feel "the next day" still counts as fast, you're living in 2012. The market has moved on, and so has your customer.
Five minutes isn't a recommendation. It's a survival threshold.
The same study contains a number teams often overlook because it sounds too extreme: responding in 5 minutes versus 30 minutes isn't "a little better." At 30 minutes, the likelihood of contact is drastically worse, and the probability of qualification is many times lower than if you respond within five minutes. This isn't marginal optimization; it's the difference between even getting a chance to talk.
Even more uncomfortably, after about 20 hours, additional persistence can start working against you. Not because follow-up itself is bad, but because the customer has already made a decision or shifted their attention. You look like you're chasing someone who's already left the store. Then, in the internal report, we write: "The lead was unqualified." Often, they were perfectly viable - you were just slow.
The key takeaway isn't "call everyone who fills out a form." The key takeaway is: build a response system that can choose the right next step in minutes. Speed without relevance is spam. Relevance without speed is literature.
Plot twist: your salespeople aren't lazy. They're too diligent.
When you start measuring where minutes are lost, you usually don't find ignorance or laziness. You find an obsession with perfect preparation. A salesperson wants to make the right impression, so they open LinkedIn, check the company, read the news, look at the tech stack, find common ground, and only then compose the "first email." In practice, researching one lead often takes around 40 minutes - and that's with good people who care.
This virtue - thoroughness - becomes the main reason sales loses. While the salesperson manually crafts the perfect introduction, the customer has already received a response from a competitor who wasn't necessarily smarter, but was timely. And then we make a typical corporate move: add more rules, more fields in the CRM, more approvals, and "quality control" to protect quality. The wheel spins in place, but the sweating increases.
Orchestration by intent: the same lead doesn't mean the same response.
Inbound today isn't a game of capturing leads; it's a game of speed and relevance in the first response. That's why at FrodX, we don't view "SLA between marketing and sales" as a Confluence table. We see it as orchestration by intent: what did the person want the moment they clicked "Submit"? If you don't differentiate this, you'll either be too slow with hot signals or too aggressive with cold ones.
The model that proves useful in practice is simple but requires discipline. For content downloads (e-book, guide), it's often more sensible to quickly send a personalized follow-up and guide them to the next content step, rather than a three-minute call that misses the signal. For a demo or pricing request, there's no debate: an immediate, personalized response, or you're billing yourself. For a "call me now" or callback request, the time value of the signal is highest, and a meaningful voice response should be immediate - human or Voice AI. For outbound, research-heavy follow-up, there's room for automated preparation so the salesperson doesn't burn 40 minutes on a lead that just needs quick, correct redirection.
The question for most companies isn't whether they understand this "in theory," but whether the model is documented, implemented, and measured. If the answer is "we usually call as soon as possible," then you don't really have a system; you have hope.
AI isn't magic. It's a discipline accelerator.
When people hear "AI in prospecting," they imagine generic spam - and they're right if it's set up incorrectly. An agent isn't useful because it's "AI," but because it's correctly configured: knowledge sources, guardrails, localization, approval logic, and clear outputs. Without that, you get beautifully packaged chaos that you'll just spread faster.
At FrodX, we use the HubSpot Prospecting Agent for four tasks that directly solve the time problem: preparing a research brief, choosing the right angle, drafting a personalized outreach message, and integrating into a sequence/task flow. The result isn't a "robot that sells," but a salesperson who gets most of the preparation done in five minutes instead of forty. The 35-minute difference per lead sounds small until you multiply it by 30 leads a week: that's 1,050 minutes, or 17.5 hours - almost half a person's work week spent on tasks you can standardize.
For cases where an immediate voice response makes sense, there's another category: Voice AI (e.g., for missed call recovery, callback requests, simple qualification). If a large percentage of your inbound calls come outside business hours, the question is very concrete: will you let these people wait until tomorrow, or will you build a system that catches them the minute their intent is still alive?
One thing must be clear: AI doesn't fix a bad process. It only accelerates it. If you have chaos today, you'll have faster chaos tomorrow.
Self-check: where is your money leaking?
If you want to solve this without motivational speeches, first answer some brutally concrete questions. How many minutes, on average, pass from a "demo request" to the first personalized response - 5, 30, or 18 hours? Which signals trigger an immediate response, and which go into nurture: is this documented or "up to the salesperson"? How much time does a salesperson spend manually researching one lead - 10 minutes or 40 minutes? Do you have approval logic to prevent generic AI spam, or do you send everything automatically? And what percentage of leads do you get when the team isn't working (evenings, weekends), and what actually happens to them?
If you don't know the number for three out of five questions, you have a measurement problem. If you know the number and it's bad, you have a response problem - and that's fixable with a system, not with additional pressure on people.
Igor Pauletič
