| name | qualify-inbound |
| description | Qualify an inbound lead by running signals against their company domain and reading native fit + urgency scores. Falls back to manual scoring when the Saber CLI or a scoring profile isn't available.
|
Qualify Inbound
Use this skill to qualify an inbound lead fast — run the signal questions against their company, then read calibrated fit and urgency scores.
For scoring concepts, see _shared/scoring.md.
Saber CLI check
Run saber --help. With the CLI you get native scoring — without it, the skill still works using a manual qualification fallback (Path B).
Step 1 — Get the lead
Ask for the inbound company's domain (e.g. acme.com). If only a name is provided, ask for the domain or attempt to infer it.
Step 2 — Choose the path
saber scoring profile list
saber scoring assignment list --type company --object <domain>
- At least one company scoring profile assigned to this domain → Path A (read native scores).
- A profile exists but isn't assigned to this domain → Path A, with one extra step: assign first.
- No company scoring profile at all → Path A is unavailable. Either route to
configure-scoring (best one-time investment) or take Path B.
- No CLI → Path B.
Path A — Native scoring
A1. Check for fresh signal data
Before running new signals (which costs credits), check for recent results on this domain:
saber subscription list
If subscriptions covering the relevant signals have already run for the domain's list, scores already exist and you can skip straight to A3.
A2. Run signals if needed
Show the credit balance:
saber credits
Tell the user: "Running [N] signals against [domain] will use [N] credits." Confirm, then run:
saber signal --domain <domain> --question "<signal question>" --answer-type boolean --no-wait
saber signal get <signalId>
Auto-trigger will update assigned scores as each signal completes.
A3. Assign profile if needed
If the lead's domain isn't already assigned to the profile:
saber scoring assignment create --profile <profileId> --type company --object <domain>
Assignment triggers immediate compute.
A4. Read scores
saber scoring scores --type company --object <domain> --detailed
You get one row per dimension (fit, urgency) with the contributing rules.
A5. Map to inbound tiering
| Fit | Urgency | Tier | Recommended action |
|---|
| ≥ 70 | ≥ 50 | Hot | Fast-track to AE; reach out within hours |
| ≥ 70 | < 50 | Right fit, slow timing | Keep warm — light-touch nurture, re-check on next signal cycle |
| 40–69 | ≥ 70 | Reactive — wrong fit, hot moment | Discovery call to verify; could be a one-off rather than a long-term fit |
| 40–69 | < 70 | Possible fit | Use research-account before reaching out |
| < 40 | any | Not a fit | Polite nurture or disqualify |
If a disqualifier rule contributed strongly negative points, surface it explicitly — that's usually a hard "no fit" regardless of the rest.
A6. Present qualification summary
## Qualification: [Company] ([domain])
**Tier:** Hot / Right fit, slow timing / Reactive / Possible fit / Not a fit
**Fit:** 84 (Δ +6)
- New VP Sales hired (+20)
- 51–200 employees (+15)
- HubSpot user (+10)
**Urgency:** 72 (Δ +12)
- Series B in last 6 months (+25)
- Hiring SDRs (+15)
**Recommendation:** [1–2 sentences explaining the tier and suggested next action]
Path B — Manual qualification (fallback)
Used when there's no CLI or no scoring profile.
B1. Load signal definitions
Check conversation context for approved signals (from signal-discovery or informal). If none, ask:
"Do you have defined signal questions for your ICP? If not, run signal-discovery first to set them up."
B2. Run the signals manually
Without the CLI, research each signal question against the domain — web search, LinkedIn, news, the company's own careers page.
B3. Score
Count positive signals; weight by the user's stated priorities. Apply disqualifiers as in generate-signals. Map to:
- Strong fit — majority positive, no disqualifiers
- Possible fit — mixed
- Weak fit — few positives or a disqualifier fired
- Not a fit — disqualifier or zero positive
B4. Present
## Qualification: [Company] ([domain])
**Score:** Strong fit / Possible fit / Weak fit / Not a fit
| Signal | Result | Notes |
|---|---|---|
| [question] | Yes / No | [context] |
**Recommendation:** [1–2 sentences]
Step 3 — Suggest next steps
- Hot / Strong fit: use
write-outreach to draft a personalised first touch
- Right fit, slow timing: add to nurture; re-check when the next signal cycle runs
- Reactive / Possible fit: use
research-account to validate before investing AE time
- Not a fit / Weak fit: route to a generic nurture sequence or disqualify
Related
_shared/scoring.md — scoring concepts and CLI referenceconfigure-scoring — set up scoring once, qualify manyresearch-account — deeper context for borderline leadswrite-outreach — turn a hot qualification into a message
