Practice area as a qualification axis: routing inbound at multi-practice law firms

In short

1. Multi-practice law firms route inbound enquiries badly. Not for lack of effort. The routing variable is which practice, and most websites and intake teams cannot identify the practice before a human partner reads the email.
2. Practice can be inferred with high reliability from website behavior alone. A visitor who reads a post-merger integration piece, then your industrials case, then your M&A team page, is an M&A inquiry before they fill out anything.
3. Conflict checks have to happen at the front door, not at the partner-handoff stage. A lead that turns out to be a conflict after three days of internal triage costs the firm credibility and a future referral.
4. Generic intake forms qualify in the worst direction. They surface every enquiry equally, route nothing, and push the qualification cost to the partner who is the most expensive person in the chain.
5. The fix is a 60-day setup: identification, conflict integration, practice inference, partner routing. In that order, with partner routing last because it depends on the others working.

A managing partner at a London full-service firm tells the story like this. A FTSE-250 manufacturer's general counsel called the firm's switchboard at 14:20 on a Thursday with a cross-border antitrust query. By 18:00 the call had been forwarded twice, summarised once, and added to the inbox of a corporate partner who handles M&A. The corporate partner read it Monday morning, recognised it was not a corporate matter, forwarded it to a competition partner. By Tuesday afternoon the GC had given the work to another firm.

This is the routing tax. Multi-practice law firms pay it on a measurable percentage of inbound, and they pay it especially hard on the highest-value enquiries because those enquiries are the ones that get the most-considered referrals to other firms. The lost engagements are usually six figures and up.

Practice area is a qualification axis. Most inbound systems treat firm size, region, or "matter type" as the primary qualifier. For multi-practice firms, the practice is the variable that determines whether the lead gets to the right partner inside the same firm or never gets to a partner at all. (For the wider framing on identifying the visit before the partner sees it, see the senior partner is researching your competitor at midnight.)

The mis-routing tax

How often does inbound get routed wrong at a multi-practice law firm? More often than the partnership thinks.

The numbers depend on the firm, but the pattern is consistent across the firms we have spoken to. A meaningful percentage of inbound enquiries lands at the wrong practice on first pass. Of those, roughly half eventually reach the right partner. The other half either bounce or die in an inbox. Annualised, the lost-engagement value scales with firm size and is usually in the high six figures or low seven figures for a 100-lawyer full-service firm.

Two patterns drive the misses. First, the enquiry comes in with a corporate firm name on the caller ID and the receptionist defaults to corporate, even when the matter is regulatory or competition or tax. Second, the enquiry comes in via a partner who happens to know the visitor and forwards it to their own practice even when the matter is elsewhere. Both patterns are human and reasonable. Both are correctable.

The deeper problem is that nobody is measuring it. Mis-routing is invisible at most firms because the lost lead never enters the CRM, and the partner who finally reads the email does not log "this should not have come to me." Until you can see the loss, you cannot fix it.

Why multi-practice firms have it harder

A boutique tax firm has a routing problem with one variable: tax. Inbound enquiries either are tax matters or are forwarded to friendly outside firms. The qualification is binary.

A multi-practice firm has the same problem multiplied by the number of practices. Eight practice groups means eight gateways, each with its own partners, conflict register, pricing model, and intake conventions. The probability that an inbound lands at the right gateway on the first try is, mathematically, lower. The probability that the wrong-gateway partner has time to forward correctly is also lower because they are busy with their own pipeline.

Multi-practice firms also have a positioning problem that single-practice firms don't. The firm's website covers a broad surface. A visitor reading the home page sees "M&A, litigation, tax, employment, regulatory, IP" and clicks the one closest to their question, but the click is approximate. Half the time the clicker is in the right practice and the matter is in another. The website confirmed an interest band. The practice itself is undetermined.

The third structural problem is internal politics. Practices compete for credit on cross-practice matters. A lead that genuinely spans M&A and tax should not be claimed by the M&A partner just because the website's M&A page got the click. Routing logic that ignores this creates real friction inside the partnership, and the partnership notices.

The four data points for clean routing

You can route correctly with four pieces of information. Three of them you can collect without asking the visitor anything. The fourth needs one or two intake questions.

The firm. Identified at the firm level from the IP and enrichment. Acme Industries plc, manufacturing, headquartered in the West Midlands, 2,400 employees. This tells you almost nothing about practice but everything about counterparty, sector, and conflict potential.

Web behavior. Which pages the visitor read, in which order, for how long. A visitor who reads two M&A articles and an industrials case is almost certainly an M&A enquiry. A visitor who reads three competition articles is almost certainly competition. The path is usually clear within four pages.

Existing relationship. Is this firm a current client of any practice? Is it on the conflict register? Has any partner worked on a matter for or against this counterparty in the last three years? The answer determines whether the routing should go to a relationship partner, a different partner, or nowhere.

The matter, briefly stated. The one piece you usually need to ask. Not a thousand-word brief. One sentence. "We are evaluating an acquisition of a UK competitor and want a competition law sense-check before approach." That sentence, plus the first three data points, gets the lead to the right partner with very high reliability.

Three of the four come from the visit itself. One comes from a thirty-second conversation. That conversation can be with a human at the front desk or with an AI agent that asks the question in plain language and writes the answer to the partner's inbox.

Inferring practice from web behavior

Practice inference from a reading path is more reliable than most law firm partnerships expect. The reason is that legal content is highly clustered. An article about post-merger integration sits inside a corporate / M&A vocabulary. An article about cartel enforcement sits inside a competition vocabulary. The two vocabularies barely overlap.

When a visitor reads three articles in twelve minutes, the practice is usually identifiable from the article tags, the keywords, and the case studies linked from each piece. The exceptions are the genuinely cross-practice matters, and those are exactly the cases the routing logic must handle differently rather than guess.

The cross-practice cases are where naive routing causes the most damage. A visitor who reads a tax article and then a litigation article is probably looking at a tax dispute, not "either tax or litigation." A visitor who reads an employment article and then a regulatory article is probably looking at a sector-regulator employment matter. The system should allow a multi-practice flag instead of forcing a primary label.

What does not work is treating reading depth as a "lead score" the way SaaS marketing teams do. The depth signal is qualitative for legal services. Eight minutes on a single technical article is a strong signal regardless of bounce rate or scroll percentage. A visitor who returns three days later for a related article is a strong signal. A visitor who clicked five articles in two minutes is probably a competitor doing market research and the alert should be muted entirely. (For more on why light pipeline beats lead scoring, the longer argument lives in that piece.)

The intake conversation

The conversation that fills in the matter has to be short, contextual, and partner-grade. A SaaS chat widget that says "How can I help?" is the wrong shape because the answer is rarely a single sentence the visitor wants to type into a popup.

A better conversation looks like this. The agent recognises the firm, recognises the practice path, and opens with one specific question that confirms the practice. "We see you've been reading our recent pieces on competition compliance. Are you exploring a notification matter, an enforcement response, or something else?" Three options as a framing makes the question answerable in one click. The visitor's answer (or refusal to answer) is itself diagnostic.

Two follow-up questions, maximum. One on jurisdiction (UK only, EU, multi-jurisdictional). One on urgency (week, month, quarter). Anything beyond that belongs in the partner-to-partner call, not in the intake conversation.

What the agent must not do: simulate legal advice, ask for confidential matter details, or push for a meeting before the routing has happened. The job of the agent is to fill in the practice variable and the timing variable. The partner conversation is what closes, and the agent's job ends when the brief is on the partner's desk.

Conflict at the front door

Multi-practice firms run conflict checks. Most run the check at the wrong stage of the funnel.

The standard model is: lead arrives, partner reads, partner decides if a conflict check is needed, conflict team runs the check, lead either proceeds or doesn't. This works for outbound business development where the partner is initiating contact. It fails for inbound because the lead has already invested time in your firm by the time the conflict check runs.

If the conflict check happens at the front door, the firm's response to a conflicted lead is fast and graceful. The agent says "We need to do a quick conflicts check before we discuss the matter further. Can we get back to you tomorrow?" The conflict team runs the check that night. By morning the lead either gets a routing to the right partner or a polite "we have to recuse on this one, here are two firms we'd recommend." The polite recusal is a relationship-builder. The slow recusal three weeks later is not.

This requires the conflict register to be queryable in real time, or near it. Most firms still run the register as a batch process. Updating the workflow to query at the front door is a one-time IT project that pays for itself within the first six conflicted enquiries. (The privacy posture for any vendor doing this query is non-trivial, so worth pre-reading the partner-level read of Article 28 before the procurement gate opens.)

A 60-day routing setup

If your firm has a website, an inbound queue, and zero practice routing intelligence, here is the order of operations.

Days 1 to 14: identification. Firm-level identification on every visitor. No practice inference yet, just the firmographic signal. Most firms can be live inside two weeks.

Days 15 to 28: conflict integration. Pipe the identified-firm signal into your conflict register query. The output is a status flag (clear / current client / conflict / unknown) on every visitor session. This is the highest-return fortnight of the whole project for most firms, because the gain is direct: no more slow recusals.

Days 29 to 45: practice inference. Layer the reading-path model on top of the firm signal. Tag each session with the inferred practice plus a confidence band. Cross-practice signatures get flagged as such, not collapsed.

Days 46 to 60: partner routing. Each practice group nominates a triage partner who receives the daily summary of identified enquiries for that practice. The triage partner forwards to the matter-running partner the same day. The handoff is internal and invisible to the visitor.

By day 61 the firm has visibility on its inbound, conflict-clearance at the front door, and practice-routed alerts that arrive while the visitor is still warm. The mis-routing tax is meaningfully smaller. Most of the gain comes from the conflict integration plus the daily-summary discipline. The practice inference adds a long tail of correctly-routed cross-practice matters that the human eye misses, which is exactly the tail that contains the biggest engagements.

Frequently asked questions

Our reception team already triages enquiries. Why automate the routing?

Reception triages on the visible information, the caller's company name, what they say, the partner they ask for. The web behavior signal is invisible at the desk. Automation is not replacing reception, it is enriching the file before reception sees it. The receptionist still answers the phone. The note attached to the call now reads 'visited four competition law articles last week, viewed your cartel-defence team page on Tuesday' rather than 'caller wants to discuss a matter'.

What if the enquiry doesn't fit a single practice cleanly?

Many don't, and that is the most important case to handle correctly. Litigation matters often have a tax angle. Regulatory enquiries often overlap with employment. The system should flag a cross-practice signature rather than force a single label. The triage partner reviews the flag and decides whether to involve a second practice from the start, which is exactly the judgement call partners should be making and which mis-routed alerts deny them.

Aren't conflict checks the firm's compliance job, not something the website should touch?

The conflict register is the firm's responsibility, run by compliance, owned at partnership level. The website is not running the check. The website is querying the register at the front door so a conflicted enquiry can be routed gracefully (recusal with a referral) rather than disgracefully (silence for three weeks). The querying logic is read-only against the existing register and audited the same way any other compliance read is.

Will this work for a 50-lawyer firm or only the magic-circle ones?

Both. The volume is lower at smaller firms, the practice list is shorter, the routing logic is simpler. The cost of a mis-routed enquiry is also smaller per incident, but more visible because there are fewer enquiries to lose. Most 50-lawyer firms find the conflict-integration step delivers more value than the practice-inference step, because at smaller firms practice ambiguity is rarer and conflicts are denser.

How does practice inference handle ambiguous reading paths?

Often the path is mixed because the buyer hasn't named their problem yet. The system falls back to a short clarifying question rather than guessing. Two questions in the intake conversation usually settle it. If the visitor refuses to answer the questions, the alert routes to the practice the reading path most strongly suggests, with the cross-practice flag set so the receiving partner knows to read the file before responding.