After every lead call, Lexidesk decides whether the lead is qualified or not qualified and tags the lead record so your team can filter for the ones worth chasing.
The Lead Qualification Prompt is where you tell the AI what counts as qualified for your firm. You'll find it in Agents → Configure → Analysis Prompts → Lead Qualification Prompt.
Note: This analysis only runs for leads. Existing clients, callbacks, vendors, and wrong-number calls are skipped.
How qualification works
After every lead call, Lexidesk reads two things:
The full call transcript
Your qualification rubric (the prompt you write in this field)
It then returns one of two values: qualified or not qualified. That's it. No score, no in-between, no maybe.
A default qualification rubric is already set up when your agent is created, so this field is optional. Leave it blank and Lexidesk uses the default. Fill it in only when you want qualification tuned to your firm's specific intake criteria.
How to write a good qualification prompt
The mental model
Imagine handing the call transcript to a new paralegal and saying "qualified or not?". If they couldn't decide from the transcript alone, the AI can't either. Qualify on signals that show up in the conversation: practice area match, jurisdiction, statute of limitations, type of matter, ability to pay if relevant, whether they've already signed with another firm.
Don't qualify on things that aren't in the call (credit checks, conflict checks, internal capacity). The AI doesn't have that data here.
Keep it short
The fewer words you use, the cleaner the decision. Most firms get reliable qualification in well under half a page of text.
Pick the simplest format that fits
Most firms only need one list, not two.
Pick the pattern that matches your firm:
Pattern 1: Broad scope, few exclusions (most common) You handle most matters that come in, with a handful of clear-cut things you don't take.
List the disqualifiers, then say everything else is qualified.
## Not qualified if any of these are true:
- [matter type you don't handle]
- [jurisdiction outside your coverage]
- [conflict-type signal you can spot in the call]
Everything else is qualified.
Pattern 2: Narrow scope, specific intake You only take a tight set of cases. List what qualifies, then say everything else is out.
## Qualified if all of these are true:
- [practice area match]
- [jurisdiction match]
- [matter type or timing requirement]
Everything else is not qualified.
Pattern 3: Meaningful criteria on both sides (rare) Use two lists only when both the qualifying conditions and the disqualifiers are non-obvious and need to be spelled out. If you find yourself writing two lists, double-check whether one of them is really just the inverse of the other.
Don't use this to rank leads
Qualification is a yes/no gate, not a ranking. If your rubric is trying to capture "very qualified" vs "kind of qualified" vs "barely qualified", you want the Lead Priority Prompt instead. That one returns a 1-to-10 score and lives in the field right above this one.
The clean split:
Qualification answers: should we even call this lead back?
Priority answers: of the leads we'd call back, which ones first?
Example: a general practice firm
This firm handles a wide range of matters across North Carolina, South Carolina, and Virginia. Rather than listing everything they do, they just list what they don't.
## Not qualified if any of these are true:
- Matter is in a state we don't cover (we only handle NC, SC, VA)
- Workers' compensation matter (we refer these out)
- Class action or mass tort claim
- Caller is already represented by another firm on the same matter
- Caller is seeking only free legal advice with no intent to hire
Everything else is qualified.
That's the entire rubric. Six lines plus a closing sentence. The AI handles every other call as qualified by default, and your team filters out the disqualified ones in the dashboard.
If your firm has a tighter intake (say, plaintiff-side personal injury only), flip the format and list what qualifies instead.
Tips
Think of this as an iterative process. Many firms don’t have written qualification criteria and that’s perfectly fine. The best results come from treating it as an ongoing activity rather than simply writing it down and forgetting about it.
Avoid soft criteria. "Caller seems serious" is too vague to score reliably. "Caller has a court date in the next 60 days" is concrete and shows up in the transcript.
Test with a recent call. Pull a transcript where you already know the answer. If your rubric doesn't cleanly land on the right verdict, the wording needs work.
FAQs
Does this run for every call?
The qualification analysis only runs for calls Lexidesk classifies as a new lead. Existing client calls, callbacks, vendors, and wrong numbers are not analyzed and won't get a qualification verdict.
What happens if I leave this field blank?
If you leave the Lead Qualification Prompt blank, Lexidesk uses its default qualification logic. We do recommend writing your own criteria to maximise this feature.
What's the difference between this and the Lead Priority Prompt?
The Lead Qualification Prompt returns a yes/no answer: qualified or not qualified.
The Lead Priority Prompt returns a 1-to-10 score. Use qualification to filter out leads your team shouldn't be calling at all. Use priority to rank the ones you do call so the best ones get attention first.
Where does the result show up?
The qualified or not qualified tag appears on each lead inside your Lexidesk dashboard and optionally in your CRM integration. Your team can filter the lead list to only show qualified leads, or sort qualified leads by priority.
Can a lead be marked qualified but get a low priority score?
Yes. Qualification and priority are independent. A lead can be qualified (worth calling back) but score low on priority (less urgent or lower value than other qualified leads in the queue). That's working as intended.
What if the AI marks a call wrong?
If qualification is landing the wrong way on real calls, open the transcript and re-read your rubric against it. Usually one of two things is happening: a rule is too vague to apply from a transcript, or you've put a check in the rubric that the call wouldn't ever surface. Tighten the wording.
Can I update it later?
Yes. You can edit the Lead Qualification Prompt any time and click Save. New qualification verdicts take effect on the next lead call. Older calls keep the verdict they were given at the time.