Every enquiry in the CRM with a name against it.

The problem

An enquiry lands in a shared inbox. Two people read it. Neither actions it. On Thursday somebody answers it, not knowing it was already answered on Tuesday. This is not carelessness. A shared inbox has no concept of ownership, so every message in it is simultaneously everybody's job and nobody's, and the default outcome is silence.

Underneath that sits a sorting problem nobody has time for. Out of forty enquiries a month, maybe six are the work you actually want. Twelve are plausible. Eight are the wrong size or the wrong country. The rest are suppliers, students and recruiters. Telling them apart takes a few minutes each, and those minutes get spent by whoever happens to open the inbox first, which is frequently the most expensive person in the building.

So the good ones queue behind the bad ones. One audit of 2,241 companies clocked an average first response of forty-two hours, with nearly a quarter never replying at all. Those companies were not lazy. They had an inbox, the inbox had a queue, and the queue was sorted by arrival time rather than by anything whatsoever to do with value.

The thing everybody buys is a form-to-Slack notification. It is free and it genuinely helps. It also produces a channel full of identical-looking messages, none of them assigned to anyone, which moves your triage problem from the inbox to Slack and moves your ownership problem precisely nowhere. What is missing is not a notification. It is a decision, made the moment something arrives, by something that never gets busy.

How it works

Six things happen between the form being submitted and somebody picking it up, and the whole sequence runs in under a minute.

1. 01Every digital door lands in one pipelineThe website forms, the shared inbox, the chat widget, the ads lead forms and the LinkedIn enquiries all terminate in the same place rather than in five. The inboxes stay exactly where they are and the team keeps working in them. What changes is that arriving somewhere now also means being recorded somewhere, which is what makes a queue countable.
2. 02It is read for intent before it is scoredWhat does this person actually want, how urgent is it, and is it the work you want. That is a reading job rather than a keyword job: a message saying we are opening a second site in March is high intent and contains no buying words at all, and a message full of buying words from a marketing agency selling you SEO is not a lead. Scoring a message you have not understood is how good enquiries end up filed as spam.
3. 03It is enriched where there is something to enrichA company name or a work email address is matched against public company data for size, sector and location, so the score is working from what the business is rather than only from what the message says. This matters most at the two ends: the enquiry that sounds small and is not, and the one that sounds enormous and is a one-person reseller.
4. 04It is checked against everybody already in the systemMatched on email domain, phone and company so a returning customer is recognised as one, an open deal is not duplicated into a second record, and the third enquiry from the same firm this month lands against the first two. Without this step the CRM fills with near-duplicates and within a year nobody trusts the numbers coming out of it.
5. 05It is routed to a person, not to a teamRouted by whatever actually decides it in your business: territory, service line, deal size, language, or who is on duty this week. The record is created with an owner, a next step and a due time against it. A team is not an owner. An enquiry that can be routed to a group can be ignored by a group.
6. 06Somebody is told, where they already areThe owner gets it in Slack with the score, the two-line summary, the enrichment and a link straight to the record, so the first thing they do is reply rather than read. Anything scored as urgent pages whoever is on duty. Anything unclaimed past its window escalates, because the point is not only to assign an enquiry but to notice when the assignment did not take.

What it's built from

Four pieces, which is why this is one of the shorter builds. Any of them swaps for the equivalent you already run.

• n8n, the part that decidesSelf-hosted on European infrastructure, and the actual product. Every routing rule, score threshold, deduplication window and escalation timer lives here as something you can open and read. When the rule changes because somebody has left or a new service line opens, one person changes it in one place rather than in four tools that each hold a quarter of the logic.
• Claude, for the reading and the summaryReads each enquiry for intent, urgency and fit against the work you want, scores it against criteria you set in plain language rather than in a points table, and writes the two-line summary that sits on the record. The scoring criteria being readable English is the part that matters: it means you can argue with a score and change the rule the same afternoon.
• HubSpot, or whatever CRM you haveWhere the record lands, with an owner, a next step, a source and a timer. Pipedrive, Zoho, Salesforce or the one you already pay for and do not use all work. If there genuinely is no CRM, putting the simplest thing that holds an owner and a date in place is part of the build rather than a prerequisite for it.
• Slack, for the handoff to a personThe enquiry arrives in the channel with its context already attached and a link to the record, and genuinely urgent ones page whoever is on duty rather than waiting to be noticed. Teams or email work the same way. A handoff that requires somebody to go and look somewhere else is not a handoff, it is a notification.

What it's worth

Do this with your own numbers rather than ours. It takes a minute, and it is the only version of it worth believing.

• You get 60 enquiries a month and somebody spends four minutes triaging each one.Four hours a month of reading, which sounds small until you notice it is four hours of your most senior person, taken in two-minute pieces that cost more than they look.
• A quarter of enquiries are not real work at all, and they currently sit in the same queue as the ones that are.15 a month ahead of or mixed in with the 45 that matter. Sorting is not the win here. Ordering is.
• Your average first response is a working day and the published research holds that replying inside the hour makes qualifying nearly seven times more likely.Most of the effect is available for free. The enquiries are already arriving, they are already yours, and the only thing changing is the gap before somebody answers.
• Two enquiries a month currently fall through the shared inbox entirely.24 a year, at whatever your average job is worth. These are the ones that never appear in any report, because nothing ever recorded that they existed.

These are illustrations using published averages and round numbers, not results we are reporting. Your volume, mix and average value will move every line, which is why the first week of the build counts them rather than assuming them.

The research behind those assumptions

Published figures rather than ours. The first three come from one audit of 2,241 companies, which is the most useful study on this that exists.

• 42hwas the average time those companies took to respond to a web enquiry. The Short Life of Online Sales Leads, HBR
• 23%of them never responded at all. Not slowly. At all. The Short Life of Online Sales Leads, HBR
• ~7xmore likely to qualify a lead when the first contact happens inside an hour rather than later. The Short Life of Online Sales Leads, HBR
• 70%of a sales person's time goes on things that are not selling, mostly admin and preparation. Salesforce, State of Sales

What it costs

This is sold as one system at the automation price on the services deck, and it sits at the lower end of that range because it is four tools and two to three weeks rather than a stack. What moves it is how many doors there are and how complicated the routing rule genuinely is. One form, one inbox and routing by service line is the bottom. Six sources, three languages, two trading names and routing that depends on who is on holiday is the top.

The running cost is a small hosting line plus the per-enquiry cost of the reading step, which is fractions of a cent per message. There is no seat licence, which matters here more than usual: the people who need the handoff are often not the people anybody would buy CRM seats for.

Outside the build price is somebody watching it, which is the care retainer. Here that means watching the scoring rather than the plumbing. Scoring rules go out of date as the business changes its mind about what it wants, and a score that is quietly wrong is worse than no score at all. People trust scores. That is the whole problem with them.

What goes wrong, and what happens then

Every one of these is in the build because it happens, not because somebody raised it once in a workshop.

• The score is wrong about a good enquiryIt will be, occasionally. So nothing is ever discarded on a score: a low score changes where it goes and how loudly, never whether it exists. Every scored enquiry is still a record somebody can find, and misses are reviewed weekly against what actually closed.
• The same company enquires three times in a monthDeduplicated on email domain, phone and company name within a rolling window, so that is one account with three touches rather than three leads and three people ringing them. The second and third touches raise the urgency rather than creating competition.
• The owner is on holidayRouting reads the rota rather than a static list. An unclaimed record escalates on a timer to a named fallback, not to a channel. Nothing is assigned to somebody who is not there, which is the most common way a well-built routing system quietly stops working in July.
• A supplier pitch is scored as a leadClassified before it is scored, and routed out rather than down. Recruiters, suppliers and students are a category with their own destination, because a sales pipeline with recruiters in it is a pipeline people stop opening.
• The CRM rejects the writeRetried with backoff, and a persistent failure raises an alert and holds the enquiry in a queue rather than dropping it. The enquiry is never only in flight: it is recorded before any downstream system is asked to accept it.
• Enrichment finds nothingCommon with personal email addresses and sole traders, and it is not an error. The record is created with what is known and flagged as unenriched, because an enquiry nobody can look up is still an enquiry. Scoring falls back to the message alone.
• Everything gets scored highThe usual failure after three months, when the criteria have been loosened once too often. The weekly number includes the score distribution for exactly this reason, so drift is visible as a shape rather than discovered when somebody stops believing the labels.