Sales Intelligence
AI-powered lead scoring — evaluate inbound leads against a weighted rubric and let only qualified opportunities reach your sales reps.
Overview
Sales Intelligence is TotalApp's AI-powered lead scoring screen inside the Sales module. It reads raw text about a lead — discovery call notes, LinkedIn profile summaries, CRM notes, email threads — and scores them against a configurable weighted rubric, eliminating guesswork from pipeline qualification decisions.
The screen is built on the MatrixEngine framework: a two-phase AI pipeline that first enforces hard non-negotiable gates, then runs a nuanced weighted scoring analysis. Results are stored per lead and displayed in the main table with score badges and Qualified / Disqualified status labels.
Two-Phase AI Pipeline
Phase 1 is a strict Linter — if a lead fails a Must-Have gate, the score is immediately zeroed and analysis stops. Phase 2 is a nuanced Critic that evaluates evidence quality across weighted categories. This design prevents a lead with glowing notes but no confirmed budget from reaching your sales reps as "Qualified".
Quick Start
- Switch to Sales mode in the header.
- Open Sales Intelligence from the left sidebar.
- Add a lead — either Add Lead and fill in company name, contact person, email, industry manually, or use the Enrich from website field inside that same modal to auto-fill from a URL.
- Click the lead row in the table to select it (row highlights, and a chevron rotates to indicate it's active).
- Click Run Analysis in the top-right of the header — the sandbox and report open in a section below the table.
- Paste or edit qualification notes in the content sandbox — discovery call summary, LinkedIn profile, CRM notes, email thread, or any relevant text.
- Click Run Analysis again inside the sandbox — the AI runs the two-phase evaluation and the result appears immediately below it.
- The lead row updates: score badge, status (Qualified / Disqualified), and analyzed timestamp.
No lead selected
Clicking Run Analysis before selecting a row shows an inline warning: "Select a row from Leads on the left to view or run AI scoring." Select a lead first, then run the analysis.
Adding a Lead — Three Ways
There's no single "correct" way to get a lead into the table — pick whichever fits the moment:
| Method | Best for | What gets filled in |
|---|---|---|
| Add Lead (manual) | A lead from a call, referral, or event — you already have the details. | Whatever you type: company, contact, email, industry, notes. |
| Enrich from website (inside Add Lead) | You only have a URL and want the form filled in before saving. | Company name, industry, email, value proposition, product list — appended into the Notes field. |
| Data Miner → Send to AI Sales Intelligence | Bulk or research-driven prospecting — you're mining a list of company sites. | Company name, contact name, email, industry, plus a note linking back to the source URL. Created directly as a pending lead, no form to fill at all. |
The one-click path
If you're already in Data Miner researching a company, the fastest route is: run the extraction, click Send to AI Sales Intelligence, then click View in AI Sales Intelligence to jump straight to the new lead. No manual data entry at all — every field the extraction found carries over automatically.
Leads Table — Selection & List Style
The Leads heading sits above the table as its own title, separate from the panel border — matching the same list-header pattern used across My Workspace screens.
Multi-select checkboxes
A checkbox appears on every row (and a select-all checkbox in the header). Checking rows shows a "N selected" counter above the table with a quick clear (×) button. Selection is independent from clicking a row to open its report — clicking the row itself opens/closes the Report section, while the checkbox marks it for bulk reference.
Table or Cards layout
The list renders as a dense table or as stacked cards depending on the List Style setting — see below.
Controlled from Settings → Appearance
Row layout (Table / Cards), corner style (Sharp / Rounded), and the multi-select mark shape (Circle / Square) are set once in Settings → Appearance → List Style and apply consistently to Sales Intelligence, Vendor Matrix, Compliance Radar, and every other My Workspace-style list in TotalApp — there's no per-screen list style toggle.
Evaluation Pipeline
Phase 1 — Linter (Gatekeeper)
The AI strictly checks every Must-Have item against the raw content. If any single Must-Have is absent or unverifiable, the result is immediate: passed_gate = false, total_score = 0. Phase 2 does not run.
Phase 2 — Critic (Scoring)
Only runs if Phase 1 passes. The AI scores the content against each weighted category from 0–100, providing reasoning text for each score.
Phase 3 — Math
The final score is calculated as: total_score = Σ (raw_score × weight / 100), capped at 100 and rounded to an integer.
Ideal Customer Profile Gates
Must-Have gates should match your ICP (Ideal Customer Profile). A lead matching every weighted category but failing a hard ICP gate — such as "B2B company" or "Annual revenue > $1M" — should never appear Qualified. The gate enforces that invariant regardless of scoring weights.
Evaluation Rules Panel
The accordion at the top of the screen controls all rules. Changes apply to the next analysis run only — existing results are not retroactively updated.
| Control | What it does |
|---|---|
| Threshold slider | Minimum total_score to count as Qualified. Default: 70. Raise for stricter pipeline standards. |
| Weight inputs | Per-category percentage. Must sum to 100 (a warning badge appears otherwise). |
| Must-Have tags | Type a gate + press Enter to add. Each tag is strictly checked in Phase 1. Examples: "Decision maker contact", "Budget confirmed", "B2B company". |
| Nice-to-Have tags | Soft bonus hints passed to the AI. Not gating — a missing Nice-to-Have lowers the score but does not block Qualified status. Examples: "Referral", "Demo requested", "Technical champion identified". |
Default Scoring Categories
| Category | Default weight | What the AI looks for |
|---|---|---|
| Bütçe/Ciro Uyumu | 40% | Annual revenue signals, stated budget, pricing tier alignment, ability to pay |
| Teknolojik Altyapı | 30% | Existing tech stack, integration readiness, digital maturity, software adoption signals |
| Karar Verici Teması | 30% | Whether the contact is a decision-maker or influencer, seniority signals, organisational authority |
Score Colours & Status Labels
Emerald — Qualified
total_score ≥ threshold. Gate passed and weighted score met the minimum. Lead is ready to advance in the pipeline.
Amber — Borderline
total_score ≥ threshold × 0.6 but below threshold. Lead shows potential but is missing key qualification signals in one or more categories.
Rose — Disqualified
total_score < threshold × 0.6 or gate failed. Lead does not meet qualification standards. Review the Missing Gates list in the Report section below the table.
| Status | Meaning |
|---|---|
| pending | Lead added, not yet analyzed. |
| qualified | Phase 1 passed + score ≥ threshold. |
| disqualified | Phase 1 failed or score < threshold. |
You can override the status by hand
The Status column in the table is a dropdown — click it on any row to switch a lead between Pending, Qualified, and Disqualified without running an AI analysis. Useful when you already know the outcome from a call and don't need the scoring breakdown, or want to correct a result manually.
AI Report Section
Selecting a lead row no longer opens a side panel — the report now renders as its own bordered section directly below the table, so the table, rules, and report all stay visible on one scroll. It contains:
Content Sandbox
Editable textarea pre-filled with the lead's saved notes, plus the Run Analysis button (report icon). Paste new information — call transcripts, emails, LinkedIn summaries — and re-run analysis. Each run overwrites the previous result.
Gate Banner
Green (Qualified) or red (Disqualified) with the total score displayed prominently. Immediate visual verdict on the lead's ICP fit.
Missing Gates
Shown only on disqualification. Lists exactly which Must-Have items were absent or unverifiable from the qualification content.
Executive Summary
One-paragraph AI verdict summarising the lead's overall qualification fitness and the key factors behind the score.
Breakdown Bars
Per-category raw score, weighted contribution, and reasoning text explaining what evidence was found (or missing) in the notes.
AI Assistant Panel
The right-side panel is now a chat-style AI Assistant, not the report — it stays collapsed by default and opens from the vertical "AI Assistant" tab on the right edge of the screen (or the sparkle icon once expanded).
Context-aware chat
The assistant sees every lead in the current list — name, contact/industry, status, and score (if analyzed) — and can answer questions like "which leads have the highest scores?" or "which leads failed a must-have gate?" without you needing to open each report individually.
Rule tuning advice
Ask it to suggest adjustments to your weights, Must-Have gates, or Nice-to-Have criteria based on the results you've gathered so far.
Fullscreen & history
Same controls as every other AI Assistant panel in TotalApp: fullscreen toggle, new chat, and per-session chat history.
Not the same as Run Analysis
The AI Assistant is a conversational helper across the whole list — it does not score a specific lead. To score a lead against your rubric, select its row and use Run Analysis in the header or the Report section below the table.
AI Engine
Both Run Analysis and the AI Assistant chat respect the Writer Engine selected in Settings → Agentic:
| Setting | Model used |
|---|---|
| API mode (default) | Anthropic claude-sonnet-4-6 |
| API mode + Cohere | command-r7b-12-2024 |
| Local CLI | claude --print fallback (no API key needed) |
| Ollama / Local LLM / In-Browser (WebGPU) | Runs entirely on your machine or in the browser — requests never reach the TotalApp server. |
Tips
- Paste the discovery call transcript for most accurate scoring. Raw call notes surface budget signals, authority level, and tech stack details far more reliably than a brief CRM summary.
- Must-Have gates should match your ICP. Common gates: "Annual revenue > $1M", "B2B company", "SaaS buyer", "Decision maker contacted". Vague gates like "good fit" are not checked reliably — be specific.
- Re-analyze after new information. When a follow-up call reveals budget confirmation or a new contact is identified, paste the updated notes and click Run Analysis again. The previous result is overwritten.
- Raw content is truncated to 4,000 characters. Keep notes focused on the most qualification-relevant signals — budget, authority, need, and timeline.
Content Limit
Raw content is truncated to 4,000 characters before being sent to the AI. For lengthy call transcripts or email chains, paste only the most qualification-relevant section — the part that addresses budget, decision authority, existing tech stack, or stated needs.
Frequently Asked Questions
defaultWeights prop on the MatrixWorkstation component. Changing category names or adding new ones requires a code change and redeployment. Contact your TotalApp administrator for custom configurations.