We Analyzed 68,520 Scored Immigration Inquiries. Here's the Intake Funnel Nobody Publishes.

Most law firms can tell you two numbers: how many leads came in, and how many consultations happened. Everything between those two numbers, the place where revenue is actually decided, is a black box.

The black box is an industry-wide condition, not a firm-level failing. Clio's secret-shopper study of 500 US law firms found that only 33% respond to email inquiries at all, and 48% are effectively unreachable by phone¹, meaning most firms lose prospects before any funnel stage could even be measured. Harvard Business Review's audit of 2,241 companies found an average first-response time of 42 hours, with 23% never responding². Whatever happens after "the lead came in" is, at most firms, invisible.

So we opened our black box. Over 21 months, our AI intake system logged and scored 68,520 conversations with prospective clients of a high-volume US immigration firm, every inquiry graded 1 to 5 on engagement, case fit, contact quality, and booking intent. This article publishes what that dataset shows: how many internet inquiries are real, which channels send cases versus noise, and what happened in one October when the firm tracked the funnel all the way to paid consultation fees. The firm is anonymized; the numbers are as they ran.

What 68,520 Scored Conversations Look Like

Every conversation was graded on how far it actually went, and the grades translate into three tiers any managing partner can read at a glance. Engaged: the prospect held a real conversation about a genuine legal problem. Qualified case: they went further, a real case plus working contact details, actively exploring their options. Consult-ready: the top tier, they asked to book a consultation and accepted that it comes with a fee. Here's how 68,520 conversations distributed:

Sit with that middle number. The folk wisdom says internet leads are junk. Across 68,520 scored conversations, more than four in ten inquiries were qualified cases with contact details, and nearly one in four was consult-ready. Leads look like junk when qualification happens days later, handled by a busy human. When it happens in the first conversation, at any hour, the quality was there all along.

The Channel Table: Where the Real Cases Come From

Because every conversation carries a grade and a source, we can answer the question most firms guess at: which channels send cases, and which send noise. (Qualified case here means the same as above: a real case with working contact details.) The gap is enormous.

Facebook sent the most volume by far, and the lowest quality by far: barely half the qualified rate of Instagram. For a firm buying leads at legal-industry prices (the highest cost-per-lead of any advertising category at $131.63 on average, with immigration typically $50-150⁵), this table is a budget-reallocation memo. It also explains why firms that only count "leads" keep funding their worst channel: on raw volume, Facebook looks like the winner.

One more cut worth knowing: prospects with US phone numbers turned out to be qualified cases 43.4% of the time, versus 38.9% for international numbers. Smaller gap than most immigration firms assume, and an argument for answering the 2am international WhatsApp message instead of writing it off.

One Month, Followed to the Money

Scores tell you about quality. To follow the funnel all the way to revenue, here is the firm's own operating report for October 2024, when 1,910 total inquiries were tracked from first touch to consultation-fee agreement:

8% of all raw inquiries, unfiltered, everything that came in, agreed to pay for a consultation within the month. At the stated fee range, that's $31,000-$42,600 in agreed consultation fees in a single month, before a single case retainer is counted.

Two design choices did the heavy lifting at the money stage. First, the fee was framed after value, never in the opening message; notably, Clio's shopper study found only 12% of responding firms provide cost estimates at all¹. Second, a human took the baton at the money moment: the system's job ended at agreement, a client services rep made the confirming call, and every conversation summary and score was pushed into the firm's CRM before the rep dialed. That mirrors the settled industry pattern; Gartner finds 95% of service leaders are retaining human agents alongside AI rather than replacing them⁶.

Why Any of This Worked: The Speed and Persistence Layer

None of the numbers above are achievable at industry-standard response behavior. The MIT/InsideSales study of 15,000+ web leads found the odds of contacting a lead are 100x higher when the first attempt happens within 5 minutes versus 30 minutes, and the odds of qualifying it are 21x higher³. HBR's research found firms responding within an hour were ~7x more likely to qualify the lead than those an hour slower².

Persistence compounds the same way: 80% of conversions require five or more follow-ups, yet 44% of salespeople stop after one⁴. This dataset added a counterweight the research rarely mentions: early on, follow-ups sent too close together read as pushy and suppressed replies. Spacing them out lifted engagement. Persistent and patient beat either one alone.

Benchmark Yourself: The Numbers to Pull This Week

Pull these from your own CRM this week. If you can't, that's the finding. A funnel you can't measure is a funnel you can't manage, and at thousands of inquiries a month, a five-point improvement at any single stage is worth six figures a year at immigration fee levels.

The Trend Line: Where Intake Measurement Is Heading

Four shifts visible in the broader data suggest funnels like this one are about to become the norm rather than the exception:

1. AI-run intake is crossing from edge to default. Legal AI adoption jumped from 19% of professionals in 2023 to 79% a year later and has held there⁷, and firms with wide adoption are nearly 3x more likely to report revenue growth than non-adopters⁸. The differentiator is no longer using AI; it's deploying it at the revenue point.
2. Consumers are meeting the machines halfway. Gartner projects that by 2028, at least 70% of customers will begin their journey through a conversational AI interface⁹, and Clio finds 70% of legal clients already prefer or are neutral toward firms using AI⁷. What clients punish is silence, not automation.
3. Commitment is moving earlier in the funnel. In the October cohort, fee agreement happened before any human conversation, a pattern healthcare's revenue-activation platforms have already normalized, where booking and payment intent are captured in the first interaction rather than after days of phone tag.
4. Instrumentation is becoming a leadership expectation. The same shift healthcare call centers went through, from measuring handle time to measuring revenue per stage, is arriving in legal. Firms that invest consistently in workflow technology sustain growth over five years, while flat-utilization firms see declining revenue per lawyer⁸. The managing partner's Monday dashboard is moving from "how many leads?" to "where did they stop?"

Methodology & Honest Caveats

Two data sources, both from a single high-volume US immigration firm, published in aggregate and anonymized. First: a conversation-level log of 74,114 inquiries from July 2024 through July 2026, of which 68,520 carried a full 1-5 engagement score after excluding 4,869 backfilled records with no transcript and a small number of unscored rows; all distribution, channel, and origin figures are computed directly from those rows. The reader-facing tiers map to that rubric as follows: engaged = scores 3 and above, qualified case = scores 4-5, consult-ready = score 5. Second: the firm's October 2024 operating report, which tracked 1,910 total inquiries through to consultation-fee agreement; stage counts and the 70%/95% close estimates in that section are the firm's own pipeline figures, and the month's row-level log was still ramping up at the time, so treat that funnel as the firm's reported numbers rather than independently recomputed ones. Channel labels were normalized (for example, several Instagram referrer variants combined). One firm's numbers are a benchmark, not a guarantee; practice area, fee structure, and lead sources all move these rates. What doesn't change: these stages exist in every consumer practice, and almost nobody measures them.

Where Ignition fits

The system behind these numbers is what Ignition runs as a managed service: instant engagement on every channel, in-conversation qualification and scoring, human handoff at the money moment, and stage-by-stage reporting so you see this funnel for your own firm, live. It starts with a free 14-day pilot, and you only pay when we book a meeting.

References

1. Clio. (2024). Legal Trends Report, secret shopper study of 500 US law firms. 33% of firms responded to email inquiries (down from 40% in 2019); 40% answered calls; 48% unreachable by phone in total; of firms that responded, only 12% provided cost estimates.
2. Oldroyd, J., McElheran, K., Elkington, D. (March 2011). "The Short Life of Online Sales Leads." Harvard Business Review 89, no. 3. Audit of 2,241 US companies: 42-hour average response among responders; 23% never responded; firms contacting within an hour were ~7x more likely to qualify the lead than an hour later and 60x more likely than those waiting 24+ hours.
3. Oldroyd, J. / MIT & InsideSales.com. (2007). Lead Response Management Study. Six companies, 15,000+ leads, 100,000+ call attempts: odds of contacting a lead called within 5 minutes vs 30 minutes are 100x; odds of qualifying are 21x.
4. The Brevet Group; Spotio; Peak Sales Recruiting (compiled sales-persistence research). 80% of sales require 5+ follow-ups; 44% of salespeople give up after one attempt; 92% stop by the fourth.
5. WordStream (2025 benchmarks); Taqtics analysis of $21.4M in personal-injury ad spend (2025); My Legal Academy (2026). Legal services average cost-per-lead $131.63, the highest of any industry; immigration CPL typically $50-$150; personal injury $312-$512 by case type.
6. Gartner. (2025). 95% of customer-service leaders plan to retain human agents alongside AI; by 2027, 50% of organizations planning significant service-workforce reductions are expected to abandon those plans.
7. Clio. (2024-2025). Legal Trends Report. AI usage among legal professionals rose from 19% (2023) to 79% (2024), holding at 79% in 2025; 70% of clients prefer or are neutral toward firms using AI.
8. Clio. (2025). Legal Trends Report, tenth edition. Firms with wide AI adoption are nearly 3x more likely to report revenue growth; firms investing consistently in technology sustain growth over five years while flat-utilization firms see declining revenue per lawyer.
9. Gartner (via IBM, January 2026). Conversational AI adoption forecast. By 2028, at least 70% of customers will use a conversational AI interface to begin their customer journey.
10. Internal dataset: 74,114 logged inquiries (68,520 scored) for a high-volume US immigration firm, July 2024 to July 2026, plus the firm's October 2024 operating report (1,910 inquiries tracked to consultation-fee agreement). Published in aggregate, anonymized.