Speed.ai has now analyzed more than 5 million intake calls for personal injury law firms — and we’re currently analyzing about half a million more every month.
That number matters for one reason: pattern density. At 5 million calls, you stop guessing what goes wrong at intake and start knowing. You know which qualifying questions get skipped when the phones are busy. You know what a caller with a serious injury actually sounds like in the first ninety seconds. You know the difference between an agent who’s underperforming and an agent who’s been handed an impossible script.
This release is what we built with that knowledge. It’s the largest set of upgrades we’ve ever shipped at once. Here’s everything, in detail.
A Rebuilt Scoring Engine — Customizable Down to the Case-Type Qualifier
The core of Speed.ai — the analysis that reads every call and grades what happened — has been overhauled from the ground up.
Two things changed.
First, the analysis is now evidence-first. Every conclusion Speed.ai draws is anchored to what was actually said on the call. Findings have to matter to be reported: the engine is built to surface what genuinely affects case quality and intake performance, not to nitpick. When Speed.ai says a qualifier was missed or a case was under-evaluated, it can point to the moment in the call that proves it. The result is scoring your intake team can’t argue with — because the receipts are attached.
Second, the criteria are now fully yours. Every firm defines a signable case differently. Your MVA qualifiers aren’t the same as your premises qualifiers, and neither matches the firm across town. Speed.ai now scores against your criteria, customizable down to the individual qualifiers for each case type you take. Your minimum treatment thresholds, your coverage requirements, your red flags, your referral rules — codified into the engine, so a “good call” in Speed.ai means a good call by your firm’s definition, not a generic one.
Every analyzed call now also carries coaching notes: what the agent did well, what to do differently next time, tied to specific moments in the conversation.
Detailed Qualifier & Script Adherence
Knowing a call scored a 72 is useful. Knowing why is what changes behavior.
Every call now includes a detailed qualifier and script adherence breakdown: which qualifying questions were asked, which were answered, which were skipped entirely, and how closely the agent followed your intake script. No more re-listening to a twenty-minute call to figure out whether anyone asked about insurance coverage. It’s right there — asked, answered, or missed.
For intake managers, this turns coaching from an opinion into a checklist. For firm owners, it answers the question that’s been unanswerable at scale: are the questions we pay people to ask actually being asked?
Agent Score Breakdowns & the Agent Scorecard
Agent performance reporting got the same treatment, at two levels.
On every single call, you can now see exactly how an agent’s score was earned — the specific behaviors that added to it and the specific misses that cost points. It’s the difference between telling an agent “your scores are low” and showing them the three habits that are lowering them.
Across calls, the new Agent Scorecard view rolls performance up per agent: score trends, qualifier coverage, adherence, and the patterns that only appear over dozens of calls. Line your team up side by side, see who’s improving, and know precisely what to work on with each person — before your monthly numbers tell you something went wrong four weeks ago.
TBI Spotting
The most valuable cases are the easiest ones to lose at intake.
Callers with a potential traumatic brain injury almost never say “TBI.” They say they hit their head. That they blacked out for a second. That they’ve had headaches since the wreck, or they feel foggy, or their memory’s been off. A busy intake agent hears a routine MVA — and a potentially seven-figure case gets a routine follow-up, or none at all.
Speed.ai now flags language consistent with head trauma automatically — powered by Braincussion. When a call contains indicators like loss of consciousness, memory gaps, persistent headaches, or confusion, the lead is flagged so your team can escalate it for attorney review — before it gets treated like a fender bender.
No workflow change, no new screens to check. Same calls, same dashboard. Speed.ai just catches what a rushed first pass can’t.
Speed.ai MCP: Ask Your Intake Data Anything
Your firm’s intake data now works wherever you already work with AI.
Speed.ai MCP is a direct, secure connection between your Speed.ai account and AI assistants like Claude and ChatGPT, built on MCP — the open standard for connecting AI tools to live data. Connect once, and your intake data becomes something you can question in plain English:
“Which leads from last week scored 80+ and still haven’t been contacted?”
“How did after-hours conversion compare to business hours in August?”
“Draft a follow-up plan for my five highest-value open leads.”
The answers come from your actual call data — scores, summaries, outcomes — not generic AI guesses. No exports, no spreadsheets, no dashboard-hopping. A dashboard you check becomes a colleague you ask.
And it’s read-only by design: connected assistants can query your data, never change it. Access is tied to your account and revocable at any time.
How to Get All of This
The rebuilt scoring engine, qualifier & script adherence detail, and agent score breakdowns with the Scorecard view are live now.
TBI Spotting and Speed.ai MCP are enabled on request — message us through the in-app messenger and we’ll turn them on for your firm, usually the same day. That’s also where to go to get your custom criteria and case-type qualifiers dialed in; our team will work through your intake playbook with you.
Not a Speed.ai customer yet? Five million calls went into what you just read about. Book a demo and see what your intake calls have been trying to tell you.