Score Accounts

Score Accounts

Rank a list of accounts by ICP fit + intent + trigger signals. Calls get_gtm_context(detailed: true) unconditionally, resolves mixed-identifier inputs explicitly surfacing ambiguity, scores each account on four axes, and presents both the ranking and the weight set as iteratively-refinable artifacts.

The bar

1. Resolution accuracy 100% — every input auto-resolved / verified / ambiguous / failed. Nothing silently picked.
2. Every score explainable — composite is a transparent weighted sum, never an opaque number.
3. "Why now" cites a specific signal — not the composite restated.
4. Every tier comes with a recommended action.
5. Weights and axes are exposed and overridable.

Sellers reject black-box scores. Transparency + per-account "why now" are what make this skill trusted.

Scope

Scores company-level accounts, not contacts. Persona-aware ranking is a chain target via personalize-email after tier-A is produced.

Input

• Accounts (required) — list of ZI IDs / company names / domains / mixed CSV.
• Use case (default prospecting) — prospecting, abm, territory_planning, pipeline_acceleration. Affects tier thresholds + recommended actions.
• Weight overrides (optional) — {fit, intent, trigger, engagement} summing to 100.
• Tier thresholds (optional) — {A, B}. C is the remainder.
• ICP override (optional) — natural-language refinement on top of get_gtm_context.icp.
• Intent topics (optional) — explicit list overriding GTM-derived defaults.

Four-axis framework

Each axis 0–100 independently. Composite is the weighted sum — never collapsed to an opaque number.

Default weights

fit:        45%intent:     25%trigger:    25%engagement:  5%   (redistributed if unavailable)

User overrides accepted. Weights are exposed in every output. Cache per-axis scores; recompute only the composite when weights change.

Tier thresholds

Use-case adjustments: abm → A=80/B=55 · territory_planning keeps defaults · pipeline_acceleration → A=65/B=40.

Workflow

1. Pull GTM context (always)

get_gtm_context(detailed: true). Capture ICP, personas, competitors, offerings, strategic priorities. ICP = fit-axis target; strategic priorities → intent-topic curation.

2. Honor input data first

Use user-supplied weights / thresholds / ICP refinements / intent topics. Fall back to GTM defaults only for missing fields.

3. Resolve identifiers (four-bucket routing)

• Auto-resolved — top match dwarfs alternatives. Score without confirmation.
• Verified — clear top match BUT plausible alternatives exist. Score; surface verification note.
• Ambiguous — no dominant match. Pause scoring; surface candidates.
• Failed — no match. List separately.

Routing by input type:

• Numeric ZI ID → auto-resolved.
• Domain (.com / .io / .co / .ai) → search_companies(companyWebsite). Single match → auto-resolved. Multiple → ambiguous.
• Name → search_companies(companyName).
  • Top match's size/revenue dwarfs alternatives → auto-resolved.
  • Clear top but 3+ plausible alternatives → verified with note.
  • No dominant match → ambiguous.
• No match → failed.

Surface rule. Never silently pick a winner. Present top 5 with attributes; ask user to confirm. Use GTM context as soft tiebreaker for verified (e.g., a B2B SaaS context defaults an ambiguous name to the SaaS-industry candidate over an unrelated-industry candidate, with a flag).

Domain-confirmation gate (mandatory for high-collision names). When search_companies(companyName=X) returns >100 matches AND no strong GTM tiebreaker exists, require domain confirmation. Surface top match's domain and ask. Never silently auto-pick — cost of getting it wrong is scoring the wrong company entirely.

Duplicate-record detection (mandatory). If top candidates share the same domain root (e.g., acmeco.com and acmecoinc.com) AND ≤20% revenue diff AND same metro/country → flag suspected duplicate. Surface both records and offer to union. For signal-heavy workflows, scoring both and unioning is the right default — signals may be split across records.

Resolution path must hit 100% accuracy. Score auto-resolved + verified immediately; pause ambiguous; list failed separately.

3.5. Relationship-context pre-flight (mandatory)

Tag each resolved account against GTM context. Tag visible on the row before the tier letter — sellers see relationship status BEFORE running the play.

• competitor ⚔️ — in get_gtm_context.competitors. Don't exclude from ranking (competitive intel matters) but make it impossible to miss visually.
• customer 🤝 — in get_gtm_context.customers / proof_bank. Shift recommended action to expansion / renewal.
• partner 🔗 — in get_gtm_context.partners / integration_partners. Shift to co-sell / integration angle.
• prospect — default; no tag.

In the row label: ⚔️ [Account] (B 62). Skill never silently produces "pursue this competitor" rankings.

4. Define the intent relevance set

From get_gtm_context.strategicPriorities, offerings, and competitor categories (or a user-supplied list), derive 5–10 themes you sell into. enrich_company_signals returns each company's active intent topics directly — there is no topic lookup or pre-query step — so these themes are the match set used in scoring (step 6): a returned topic counts toward intent only if it maps to one of them. Keep the themes; the matching happens per account during scoring.

5. Fetch data per account (parallel, batched ≤10; chunked for large lists)

Both calls below batch multiple accounts per request, so fetch a chunk of accounts together rather than one-by-one:

• enrich_companies(zoominfoCompanyIds: [chunk], fields: industries, employeeCount, revenue, country, metroArea, businessModel, employeeCountByDepartment, foundedYear) — up to 25 per call.
• enrich_company_signals(zoominfoCompanyIds: [chunk], signalTypes: ["INTENT", "NEWS", "SCOOP"]) — up to 10 per call. Returns each account's recent intent topics (each with signalScore and audienceStrength), news (with category), and scoops (with scoopType), plus a date on every signal. Do not pre-filter on score, topic, category, or date — that is applied during scoring (step 6).

Hard batch limit: ≤10 accounts per enrich_company_signals call (≤25 per enrich_companies call).

Batch + context-window discipline. For lists >25 accounts, process in chunks of ~25 accounts end-to-end (resolve → fetch → score → compose row → write chunk → discard raw payloads) before moving to the next chunk. Don't accumulate full raw enrichment payloads for hundreds of accounts in working context — once per-axis scores + the winning trigger event + the winning intent topic are captured per account, drop the rest. For >100-account lists, summarize completed chunks into running totals (tier distribution, top-A list, multi-product anomalies, duplicate-suspected flags, missing-axes counts) and discard the per-account breakdowns from context. Output is built incrementally chunk-by-chunk so a long list doesn't blow context.

Skip account_research here; fire selectively in §7.5 for tier-A.

6. Score each axis

Fit (0–100) — compare enrich_companies to get_gtm_context.icp:

Cache per account; reuse across weight changes.

Intent (0–100) — from the enrich_company_signals intent topics, keep those that map to the relevance set (step 4) with signalScore ≥ 60 in roughly the last 30 days (use each signal's date). Score max(signalScore × audienceStrengthFactor) over the survivors. A=1.0 · B=0.85 · C=0.7 · D=0.55 · E=0.4 (from audienceStrength). Cap 100. Record the winning topic for "why now." If no relevant intent survives → 0 with "no relevant intent activity" flag.

Trigger (0–100) — from the enrich_company_signals news and scoop signals, kept to the last 90 days by each signal's date (drop older). Map each signal's news category or scoop scoopType to the weight below:

event_score = signal_type_weight × recency_factor

Recency: 0-14d=1.0 · 14-30d=0.7 · 30-60d=0.4 · 60-90d=0.2 · >90d=0.

Account trigger = max(event_score) capped at 100. Record winning event for "why now."

Engagement (0–100) — if account_research returns rich CRM context: active deal/renewal/champion = 80–100 · past meeting/known stakeholder = 40–70 · no history = null. If null, redistribute weight and surface gap.

7. Compute composite + assign tier

composite = round((fit × w_fit + intent × w_intent + trigger × w_trigger + engagement × w_engagement) / 100)

Assign per thresholds. Default A≥75 / B 50–74 / C<50 (use-case overrides apply).

7.5. Auto-pull account_research on tier-A rows (mandatory)

Tier A = "route to AE in 24h." Engagement-axis gap on tier-A is the highest-cost gap to close.

For each tier-A account (and ONLY tier-A — cost control): account_research(zoominfoCompanyId, query="Open opportunities, active deal stages, named champion or blocker, last activity date, renewal timing"). Parse for:

• Open deal status — stage, value, next step.
• Renewal date — surface prominently if within 90 days.
• Named champion / blocker — source-tag [from account_research].
• Last activity — flag if >60 days old.

Append inline beneath the why-now:

| 1 | [Account] | 🤝 A | 84 | ... | [Trigger event] X days ago — [pain-bridge]                                     ↳ Engagement: open deal $XXXk, champion [Name], last activity Xd ago [from account_research]

If no CRM history → annotate "no engagement signal — cold open."

For tier-B/C: skip — cost-to-value doesn't justify.

8. Compose "why now" per account

One sentence anchored on the strongest signal:

• Trigger + in-tier fit → cite event + date. "Closed [counterparty] acquisition 20 days ago."
• High intent → cite topic + score + recency. "Spiked on '[topic]' (score 92, audience A) over 14 days."
• Strong fit, no fresh signal → "Perfect-fit ICP — no fresh trigger; pursue on fit alone."
• Engagement-driven → "Active deal in flight; renewal due in 47 days."
• Strong trigger BUT C-tier (fit mismatch) → be explicit about routing: "Do not pursue — strong trigger (new CEO 10 days ago) but ICP mismatch ([reason]) keeps this low priority." Don't bury the trigger; surface BOTH signal and recommendation.
• Low signal across all axes → "Low signal — monitor only."

Never restate the composite as the why-now. Always cite the underlying axis driver.

9. Self-check before output

• ☑ Composite shown with component breakdown (fit / intent / trigger / engagement).
• ☑ "Why now" cites a specific signal, not the composite.
• ☑ Tier has a recommended next action.
• ☑ Weights + axes used exposed.
• ☑ Every input bucketed (resolved / ambiguous / failed) — none silently dropped.
• ☑ Ambiguous surfaced, not silently picked.
• ☑ Stale signals (>90d) contribute 0; not padded.
• ☑ Missing axes flagged + weights redistributed transparently.
• ☑ Iteration options offered.

10. Present + offer iteration

1. Accept ranking; save filter+weight set.
2. Adjust weights — re-rank without recomputing axes.
3. Tighten / loosen tier thresholds.
4. Refilter — remove tier C / specific industries.
5. Swap ICP — different ICP definition.
6. Drill into one account — chain to personalize-email.
7. Add accounts — extend list and re-score.

Re-execute step 5 only when account list changes. For weight / threshold / ICP changes → recompute from cached axis scores.

Terminate when user accepts, saves, or hands off.

Anti-patterns — fail-fast checklist

1. Black-box composite — single number without component breakdown.
2. "Why now" = composite restated.
3. Silent identifier resolution on ambiguous names.
4. Fixed weights not exposed.
5. Tier without action.
6. Stale signal padding — events >90d contributing.
7. Generic "why now" — "good fit" applies to every account.
8. Ignoring missing axes — pretending engagement exists when null.
9. Auto-accepting ambiguous matches.
10. No iteration affordance.

Fallback rules

• get_gtm_context empty → use user-supplied ICP override; surface gap.
• No intent returned, or none matching the relevance set → intent score = 0 (real signal, not a gap).
• No news or scoops returned → trigger = 0; flag.
• Engagement unavailable → weight = 0; redistribute proportionally.
• All axes thin → tier C "monitor only"; honest.
• Resolution failure → list separately; never silently drop.

Never block ranking on a single missing axis. Never invent data.

Output Format

TL;DR — Account Scoring · N accounts · Pass [M]

Use case: [restate]. Weights · Thresholds A≥[X] · B[Y–Z].

Resolution: [R resolved · A ambiguous · F failed]. [If A>0: "User confirmation required."] Tier distribution: A: x · B: y · C: z.

Top 3:

1. [Account] (tier · composite) — [why now]
2. ...

Resolution Summary

Status legend: ✅ Auto-resolved · 🔍 Verified · ⚠️ Ambiguous · ❌ Failed.

Ambiguous matches — please confirm: [list top 5 candidates per ambiguous input with attributes].

Ranked Accounts

Sorted by composite descending. Engagement column – when redistributed.

| # | Account | Tag | Tier | Composite | Fit | Intent | Trigger | Eng | Why now | ZI ID |

(Tier-A rows also carry an "↳ Engagement: ..." sub-line from §7.5.)

Weights & Axes Used

fit:        [%]intent:     [%]trigger:    [%]engagement: [%]   (redistributed if axis unavailable)

Axes missing this run: [list, or "none"].

Recommended Actions per Tier

• Tier A — Route to AE for 1:1 outreach within 24h. Chain to personalize-email.
• Tier B — SDR sequence; ABM retargeting; cadence with the why-now as opener.
• Tier C — Monitor; re-score weekly.

Iteration Options

1. Accept ranking; save filter+weight set.
2. Adjust weights.
3. Tighten thresholds.
4. Refilter.
5. Swap ICP.
6. Drill into one account.
7. Add accounts.

Caveats (when relevant)

• Ambiguous pending — N accounts not yet scored.
• Failed resolutions — N inputs had no match.
• Engagement axis unavailable — surface per-account (no CRM signal — consider cross-check) for each tier-A row.
• Signal depth — enrich_company_signals returns the most recent signals per type (server-capped), so for very active accounts the intent/trigger axes reflect the most recent window rather than an exhaustive history.
• Intent thin — <3 topics resolved; intent directional.
• Stale-signal cliff — N accounts' best trigger >60d old.
• Edge-of-recency — N trigger events 80–90d.
• GTM-context gap — icp sparse; fit-axis precision reduced.

Final Filter + Weight Set (on accept)

{  "icp": { /* GTM ICP or user override */ },  "weights": {"fit": 45, "intent": 25, "trigger": 25, "engagement": 5},  "tier_thresholds": {"A": 75, "B": 50},  "intent_topics": ["..."],  "use_case": "prospecting",  "_meta": {"account_count": ..., "tier_distribution": {...}, "axes_missing": [...], "pass_count": ...}}

Chain Targets

• personalize-email per tier-A contact → grounded in the same "why now" signal.
• build-list to extend the universe.
• find-similar on a tier-A seed.
• tam-sizer with this filter set to confirm universe size.