What a Sales Performance Dashboard actually does
Surfaces weekly AI coaching summaries per sales rep and deal-risk classifications across pipeline stages from CRM activity data.
A white-label AI sales performance dashboard pulls deal notes, activity logs, and pipeline data from CRMs (HubSpot, Salesforce, Pipedrive, Close) and layers two AI outputs: per-rep weekly coaching summaries that identify patterns in rep behavior (response time, note quality, stage progression velocity) and deal-risk classifications that flag which deals are likely to slip or close based on activity signals and stage characteristics. The dashboard is designed for RevOps consultancies serving mid-market sales teams — the users are sales managers and VP Sales, not the reps themselves.
The market context is critical: HubSpot Sales Analytics, Salesforce Sales Cloud Einstein, Pipedrive Insights, and the standalone platforms (Gong at $1,000+/seat/yr, Clari at enterprise quote, Outreach Kaia at enterprise) all own this category — and none publishes a white-label SKU. The honest WL play is not to out-AI Gong on call analysis (that requires proprietary speech-to-text + call recording infrastructure) but to out-compete on the data-synthesis and reporting layer for mid-market clients who have HubSpot or Pipedrive but can't afford Gong/Clari. The AI roles are well-defined: Sonnet 4.6 for weekly narrative coaching summaries ($0.012 per summary), Haiku 4.5 for deal-risk classification ($0.0005 per deal when cached), and straight BI for everything else. The remaining engineering complexity is in the CRM API integrations — each adds 2–3 weeks of build time.
AI capabilities involved
Weekly per-rep coaching summary generation
Deal-risk classification (at risk / on track / ahead)
Deal-note summarization across CRM activity logs
Forecast commentary on pipeline shifts
Win/loss reason extraction from closed deals
Who uses this
- RevOps consultancies serving 5–20 mid-market sales teams at $399–799 ARPU per client
- B2B SaaS founders building a CRM-agnostic performance layer for companies using HubSpot, Pipedrive, or Close
- Sales-training agencies building a branded coaching-accountability tool that connects session insights to CRM pipeline data
- Fractional VP Sales and interim revenue leaders who want a branded dashboard to deliver weekly performance reports to their clients
SaaS alternatives on the market
Real products you can sign up for today — with current 2026 pricing, honest pros and cons.
HubSpot Sales Analytics
Sales teams already on HubSpot CRM who need reporting within their existing tool without a third-party dashboard
$90/seat/mo (Sales Hub Pro)
Pros
- +Native integration with HubSpot CRM — no data sync required.
- +Deal pipeline views, forecast summaries, and rep activity dashboards built in.
- +No additional integration or maintenance overhead.
Cons
- −No white-label — all reports and dashboards show HubSpot branding.
- −Only works for HubSpot CRM data — no cross-CRM view.
- −No AI weekly coaching narrative — predictive scoring only.
- −Sales Hub Pro ($90/seat/mo) adds significant per-seat cost for mid-market teams.
Gong
Mid-to-large enterprise sales teams (50+ reps) with budget for $1K+/seat/yr and a need for call recording and conversation intelligence
Quote-based (~$1,000+/seat/yr minimum estimate)
Pros
- +Industry-leading call recording and conversation intelligence — the gold standard for sales coaching.
- +AI-generated deal health scores based on actual call content and CRM activity.
- +Strong integrations with Salesforce, HubSpot, and most major CRMs.
Cons
- −No white-label capability at any tier.
- −Enterprise-only with 50+ seat minimums common in practice.
- −Call recording infrastructure adds compliance complexity (two-party consent in 11 US states).
- −RevOps consultancies cannot resell Gong access under their own brand.
Clari
Enterprise companies with $50M+ ARR that need revenue-grade forecast accuracy and executive pipeline inspection at scale
Enterprise quote only
Pros
- +Best-in-class revenue forecasting with AI-driven pipeline health scoring.
- +Real-time deal inspection across the full sales pipeline.
- +Strong executive reporting for VP Sales and CRO.
Cons
- −No white-label capability.
- −Enterprise-only with no published pricing — significant sales cycle required.
- −Not designed for RevOps consultancies managing multiple client teams.
- −Overkill for mid-market companies with 5–25 rep sales teams.
Pipedrive Insights
Small-to-mid sales teams on Pipedrive who need basic pipeline reporting within their existing CRM subscription
Bundled in Pipedrive plans ($24–64/user/mo)
Pros
- +Clean visual pipeline and activity reports for Pipedrive CRM users.
- +Custom dashboard builder with drag-and-drop widgets.
- +No additional cost beyond Pipedrive subscription.
Cons
- −No white-label capability.
- −Pipedrive-only — no cross-CRM view.
- −No AI coaching narrative or deal-risk classification.
- −Limited to what Pipedrive's own data model exposes.
The AI stack
The sales dashboard stack is intentionally two-tier: Haiku 4.5 for high-frequency classification tasks (deal-risk, note summarization) and Sonnet 4.6 for weekly narrative coaching summaries. The BI layer (charts, pipeline views, forecast totals) is pure SQL — no LLM needed.
Weekly rep coaching summary
Generates a 200–400 word natural-language coaching summary per rep covering deal activity, response patterns, and progress against quota
Claude Sonnet 4.6
$3.00 / $15.00 per M tokensAll weekly coaching summaries — this is the right model for the task
GPT-5.4 mini
$0.75 / $4.50 per M tokensHigh-volume builds with 100+ reps where even $0.012/summary becomes a meaningful COGS line
Our pick: Claude Sonnet 4.6 for all weekly coaching summaries. The cost is negligible — even 200 weekly rep summaries per month costs $2.40 in AI COGS.
Deal-risk classification
Classifies each open deal as 'at risk,' 'on track,' or 'ahead' based on stage progression velocity, activity recency, and deal age
Claude Haiku 4.5
$1.00 / $5.00 per M tokensProduction deal-risk classification running nightly on all open deals across all tenants
GPT-5.4 nano
$0.20 / $1.25 per M tokensHigh-volume pipelines with 1,000+ open deals where cost-per-classification matters
Our pick: Claude Haiku 4.5 with prompt caching on the stage-criteria and deal-risk rubric. At $0.0005/deal when cached, classifying 500 deals/month per tenant costs $0.25 in AI COGS.
Deal-note summarization
Summarizes CRM activity notes for a deal into a 3–5 sentence summary for quick manager review
Claude Haiku 4.5
$1.00 / $5.00 per M tokensOn-demand note summarization triggered when a manager clicks into a deal
Our pick: Haiku 4.5 for on-demand deal-note summarization. Call only when a manager opens a deal's detail view — do not batch-summarize all deals preemptively.
Forecast commentary
Generates a natural-language weekly pipeline forecast commentary for the VP Sales view
Claude Sonnet 4.6
$3.00 / $15.00 per M tokensWeekly forecast commentary per client — fires once per week per tenant
Our pick: Claude Sonnet 4.6 for all weekly forecast commentary. At $0.022/commentary (12K-token pipeline input), cost is negligible even at 100 client tenants.
Reference architecture
The platform runs a nightly data sync from connected CRMs into a normalized Supabase schema, a daily deal-risk classification batch, and a weekly async batch that generates rep coaching summaries and forecast commentary. The hardest engineering challenge is the CRM API integration layer — each CRM (HubSpot, Salesforce, Pipedrive, Close) has a different data model for deals, activities, and contacts, requiring a normalization layer before AI processing.
Tenant RevOps consultant connects client CRM accounts
Next.js OAuth flows + Supabase encrypted token storageHubSpot uses OAuth 2.0 with the CRM Read scopes. Salesforce uses Connected Apps OAuth. Pipedrive and Close use API keys. All tokens are encrypted in Supabase. A test-connection endpoint validates each integration on setup.
Nightly CRM data sync
Inngest nightly cronFor each connected client, the cron pulls all open deals plus closed deals from the last 90 days. Data is normalized into a unified `deals` schema (id, stage, owner_id, amount, close_date, last_activity_date) and a `activities` table (id, deal_id, type, content, created_at). CRM-specific field mappings are stored per tenant.
Deal-risk classification batch
Haiku 4.5 Edge Function batchRuns nightly after the CRM sync. For each open deal, Haiku 4.5 (with cached risk-rubric prompt) receives: deal stage, days in stage, last activity date, activity frequency over last 14 days, deal amount. Returns 'at_risk' / 'on_track' / 'ahead' plus a one-sentence rationale. Results stored in `deal_risk_scores` table.
Weekly rep coaching summary generation
Inngest Monday cron + Sonnet 4.6Every Monday morning, Inngest generates one coaching summary per sales rep for each client. The Edge Function assembles: rep name, 7 days of deal activity (stages moved, activities logged, deals created/closed), quota progress, and deal-risk distribution across their pipeline. Sonnet 4.6 returns a 300-word coaching narrative. Stored in `weekly_coaching_summaries`.
Weekly forecast commentary generation
Inngest Monday cron + Sonnet 4.6The forecast cron assembles last week's pipeline snapshot vs. this week's (deals added, deals closed, stage changes, amount changes) and passes the delta to Sonnet 4.6. Returns a CFO-style 'what changed in the pipeline and why it matters' narrative for the VP Sales view.
Manager reviews dashboard and coaching summaries
Next.js dashboardThe manager dashboard shows: pipeline view by stage with deal-risk color coding (red/yellow/green), rep performance table with last coaching summary snippet, forecast commentary card, and win/loss analysis for closed deals. Clicking a deal triggers on-demand Haiku 4.5 note summarization.
Tenant administrator exports branded weekly reports
Next.js + Puppeteer PDF exportRevOps consultant exports a white-labeled weekly performance report (logo, colors, client name) as a PDF. The report includes the forecast commentary, rep coaching summaries, and pipeline health metrics. Delivered via Resend email to the sales manager.
Estimated cost per request
~$0.012 per weekly rep coaching summary (Sonnet 4.6 on 8K-token rep activity input); ~$0.0005 per deal-risk classification (Haiku 4.5 cached); ~$0.022 per weekly forecast commentary (Sonnet 4.6 on 12K-token pipeline delta)
Cost calculator
Drag the sliders to model your actual usage. The numbers update in real time so you can stress-test economics before writing a single line of code.
Model assumes a RevOps consultancy with multiple client tenants. Each tenant has a sales team with reps and open deals. AI costs are minimal — the dominant costs are infrastructure and CRM integration maintenance.
Estimated monthly cost
$78.33
≈ $940 per year
Calculator notes
- Weekly coaching summaries at $0.012 each × 4 weeks = $0.048/rep/month. 8 clients × 10 reps = $3.84/month in coaching summary COGS against $3,192/month in revenue at $399 ARPU.
- Deal-risk classification cost assumes Haiku 4.5 cached prompt. Without caching, multiply by ~5× — still under $1/client/month.
- Forecast commentary at $0.022 × 4 per client/month = $0.088/client/month. For 8 clients: $0.70/month total.
- CRM API calls are free (HubSpot, Pipedrive, Close all provide free API access). Salesforce charges for API calls above the daily org limit — verify with client before connecting.
Build it yourself with vibe-coding tools
By Sunday night you'll have a working CRM-export dashboard: sales managers paste a CSV export from their CRM, and within 60 seconds get AI-generated rep coaching summaries and deal-risk flags — good enough to demo to 2 RevOps clients.
Time to MVP
12–16 hours (1 weekend)
Total cost to MVP
$25 Lovable Pro + ~$30 API credits (Anthropic)
You'll need
Starter prompt
Build a white-label AI sales performance dashboard called RepIQ. Use Vite + React + TypeScript + Tailwind + Supabase Auth. Two roles: RevOps Consultant (admin who manages multiple clients) and Sales Manager (reads their own team's data). Database schema: - `clients` table: id, consultant_id, company_name, logo_url, brand_color, crm_type (hubspot/pipedrive/salesforce/csv) - `reps` table: id, client_id, name, email, quota_amount - `deals` table: id, client_id, rep_id, name, stage VARCHAR, amount NUMERIC, close_date DATE, days_in_stage INTEGER, last_activity_date TIMESTAMPTZ - `activities` table: id, deal_id, type (email/call/meeting/note), content TEXT, created_at TIMESTAMPTZ - `deal_risk_scores` table: id, deal_id, score_date DATE, risk_level (at_risk/on_track/ahead), rationale TEXT - `weekly_coaching_summaries` table: id, rep_id, week_start DATE, summary_md TEXT - `weekly_forecasts` table: id, client_id, week_start DATE, commentary_md TEXT RLS: clients can only read their own client row and its related data. Consultants can read all their clients. Pages: 1. /dashboard — RevOps consultant view: list of clients with status cards (pipeline health, risk deals count, report generated date) 2. /clients/:id — Client dashboard: pipeline by stage (Kanban-style with risk badges), rep performance table, this week's forecast commentary 3. /clients/:id/reps/:rep_id — Rep detail: last 5 coaching summaries, deal list, activity timeline 4. /upload — CSV upload form (deals CSV + activities CSV) for clients not yet on live CRM API 5. /reports — List of generated weekly reports with PDF download Edge Functions: - classify-deal-risk: takes array of deal objects, calls Haiku 4.5 with cached risk-rubric prompt, returns array with risk_level + rationale per deal - generate-coaching-summary: takes rep_id + week_start, assembles deal + activity data for the rep, calls Sonnet 4.6, returns coaching markdown - generate-forecast-commentary: takes client_id + week_start, assembles pipeline delta (this week vs last week), calls Sonnet 4.6, returns commentary markdown Haiku 4.5 system prompt for deal-risk MUST say: 'Classify this deal as at_risk, on_track, or ahead based on: days in current stage vs. average close time, days since last activity, activity frequency in last 14 days, and deal age vs. expected close date. Return JSON: {risk_level: string, rationale: string (max 50 words)}' Add CSV import: when a deals.csv and activities.csv are uploaded, parse and insert into deals/activities tables, then auto-trigger classify-deal-risk for all uploaded deals. Add Stripe $399/mo per client slot billing. Block classify and generate functions behind active subscription check.
Paste this into Lovable
Follow-up prompts (run in order)
- 1
Add HubSpot integration. Create an OAuth flow in /clients/:id/settings to connect a HubSpot account. Once connected, add a 'Sync from HubSpot' button that calls the HubSpot CRM API (deals endpoint with associations for contacts and activities) and upserts into Supabase. Map HubSpot lifecycle stages to a normalized stage vocabulary.
- 2
Wire classify-deal-risk to an Inngest nightly cron that runs at 2am for all deals updated in the last 7 days. Show a 'Last classified' timestamp on the pipeline view. Add a manual 'Re-classify' button per deal for on-demand refresh.
- 3
Wire generate-coaching-summary to an Inngest Monday 6am cron for all reps across all active clients. Send a digest email to the RevOps consultant (not the sales manager) via Resend with links to each rep summary. Add a 'Share with Manager' toggle per summary.
- 4
Add Pipedrive integration. Use the Pipedrive REST API (OAuth) to pull deals from the /deals endpoint with activities and notes. Map Pipedrive pipeline stages to the normalized stage vocabulary. Show a 'Connected CRM: Pipedrive' badge on the client card.
- 5
Add a weekly PDF report export. Wire a 'Generate Weekly Report' button that assembles: forecast commentary, rep coaching summaries (top 5 reps by deal count), pipeline health metrics, and risk deal list into a branded PDF via Puppeteer. Store in Supabase Storage. Email to the consultant and optionally to the sales manager.
Expected output
A working CRM-export sales dashboard with deal-risk badges, AI rep coaching summaries, and weekly forecast commentary — ready to demonstrate to 2 RevOps clients within the first weekend.
Known gotchas
- !HubSpot OAuth requires a developer app and app review before going live with more than 10 connected accounts. Start the review process immediately — it takes 5–15 business days and requires specifying exactly which CRM scopes you'll access.
- !Salesforce's REST API rate limits vary by edition — Developer Edition (free) is severely limited (5 API calls/sec); production orgs have higher limits. Test with a production sandbox, not a Developer Edition.
- !The nightly deal-risk classification batch will fire for all tenants simultaneously at 2am. With 50+ clients and 20 reps each, that's 1,000+ Haiku 4.5 calls at once. Add an Inngest fan-out with a concurrency limit of 20 to avoid Anthropic rate-limit errors.
- !CRM deal stages are not standardized — HubSpot uses 'presentationscheduled' while Pipedrive uses 'Demo Scheduled.' Build a per-client stage mapping table in Supabase that normalizes to your standard vocabulary (Qualified / Demo / Proposal / Negotiation / Closed).
- !Sonnet 4.6 coaching summaries require meaningful CRM note data — reps who log fewer than 3 activities per week produce summaries that say 'insufficient activity data.' Add a minimum-activity threshold check before generating the summary and display a 'Needs more CRM activity' badge instead.
- !Supabase RLS for multi-tenant requires explicit policies that prevent consultants from accessing other consultants' client data. Test with two consultant accounts before launch — Lovable's default policies may not isolate at the consultant level.
Compliance & risk reality check
A sales performance dashboard processes individual rep-level performance data and deal information, which raises GDPR, automated decision-making, and data protection obligations at the employer level.
GDPR Art. 22 on automated performance scoring
Deal-risk classifications and rep coaching summaries are AI-generated assessments of individual employee performance. If these scores are used as the sole basis for employment decisions (performance improvement plans, termination), they may constitute automated decision-making under GDPR Art. 22, which requires human review and the right to explanation.
Mitigation: Frame all AI outputs explicitly as 'insights for manager review' — not automated decisions. Add a human-review gate in the UI (manager must mark summaries as 'reviewed' before they appear in formal performance records). Include a standard disclaimer in client contracts that AI outputs require manager judgment before employment action.
SOC 2 for mid-market sales data
RevOps consultancies serving mid-market B2B companies will increasingly face SOC 2 Type II requirements from their clients' procurement and security teams. The platform handles deal values, customer contact names, and sales activity data — all of which are commercially sensitive.
Mitigation: Use Supabase's existing SOC 2 Type II certification and Anthropic's enterprise data handling in initial security questionnaires. Pursue your own SOC 2 when you have 5+ clients with formal procurement requirements. In the interim, provide evidence of: encryption at rest and in transit, access logging, and data retention policies.
DPA required per client tenant
Handling CRM data (deal notes, contact names, email interactions) for EU-based client companies makes you a data processor under GDPR. A Data Processing Agreement is legally required per client before processing their data.
Mitigation: Prepare a standard DPA template covering: purpose and scope of processing, data categories, retention period, sub-processors (Supabase, Anthropic), and deletion procedures. Include in all client onboarding contracts before connecting their CRM.
Call recording compliance (if Gong-like features are added later)
If call recording or call analytics are added to the dashboard in a future sprint (to compete with Gong), two-party consent is required in 11 US states (CA, CT, FL, IL, MD, MA, MI, MT, NH, OR, PA, WA) and under EU GDPR for all call participants. This changes the compliance profile of the product significantly.
Mitigation: Do not add call recording features without first implementing a compliant consent-collection flow. This is a separate sprint with significant legal review. Keep the MVP scope to CRM activity data (notes, stage changes, emails logged) — not call recordings.
Build vs buy: the real math
10–14 weeks
Custom build time
$22,000–$38,000
One-time investment
6–8 months
Breakeven vs buying
At $399 ARPU per RevOps client and 8 clients, monthly revenue is $3,192 against $77/mo in infrastructure and $5/mo in AI COGS. The $30K mid-band build cost pays for itself in roughly 9 months at 8 clients, or 6 months at 12 clients. The comparison to the vendor path is stark: Gong at $1,000/seat/yr for a 10-rep team costs $10,000/yr per client — a RevOps consultant who builds a custom dashboard can charge $399/mo ($4,788/yr) and still deliver more white-label value than Gong provides at any price. The decisive metric: Sonnet 4.6 weekly coaching summaries at $0.012 each mean the AI COGS for 10 reps per month is $0.48 — the value per coaching summary to the sales manager is $50–200+ in avoided time. That's the margin the custom build captures vs. paying Clari $50–80K/yr for an enterprise contract that doesn't white-label.
Skip the DIY — RapidDev builds the production version
A Lovable MVP gets you a demo. Production needs auth that doesn't leak data, AI calls that don't bankrupt you, observability when models drift, and code you can audit. That's what we ship.
Discovery call (free)
30 minWe map your exact Sales Performance Dashboard use case: who uses it, target volume, AI model choice, integrations, compliance scope. You get a detailed scope document and fixed-price quote within 48 hours.
AI-accelerated build
10–14 weeksOur engineers use Claude Code, Lovable, and custom tooling to ship 3–5x faster than agencies. You see weekly progress in a staging environment — not a black box.
Launch + handoff
1 weekWe deploy to your infrastructure, transfer the GitHub repo, set up CI/CD and monitoring, and train your team. You own 100% of the source code, prompts, and model configurations.
What you get
Timeline
10–14 weeks
Investment
$22,000–$38,000
vs SaaS
ROI in 6–8 months
30-min call. Fixed-price quote within 48 hours. No commitment.
Frequently asked questions
How much does it cost to build a Sales Performance Dashboard?
A custom build with RapidDev runs $22,000–$38,000 — the extended band reflects the CRM API integrations (HubSpot + Salesforce + Pipedrive are the 80% use case, each adding 2–3 weeks) and the weekly async batch pipeline. The Lovable DIY path is $25 Lovable Pro plus $30 in Anthropic API credits for the first month. Infrastructure (Supabase, Vercel, Inngest, Resend) adds $77/mo at baseline.
How long does it take to ship this?
A Lovable weekend MVP with CSV import, deal-risk classification, and rep coaching summaries takes 12–16 hours. A production RapidDev build with HubSpot + Salesforce + Pipedrive integrations, nightly data sync, weekly batch processing, and branded PDF reports takes 10–14 weeks.
Can RapidDev build this for my company?
Yes — RapidDev has shipped 600+ applications including RevOps tools, CRM integrations, and AI-powered sales analytics platforms. Book a free 30-minute consultation at rapidevelopers.com to discuss your target CRM integrations, client count, and feature requirements.
What's the difference between this and HubSpot's built-in analytics?
HubSpot's built-in analytics are HubSpot-branded, HubSpot-only, and produce charts and tables — not AI coaching narratives. This platform is CRM-agnostic (connects to HubSpot, Salesforce, Pipedrive, or Close in one dashboard), produces white-labeled branded reports under your RevOps consultancy brand, and generates natural-language weekly coaching summaries per rep using Sonnet 4.6 — something HubSpot's analytics layer doesn't do at any price.
What AI models power the coaching summaries and deal-risk?
Weekly rep coaching summaries use Claude Sonnet 4.6 ($3/$15 per M tokens) — $0.012 per summary. Deal-risk classification uses Claude Haiku 4.5 ($1/$5 per M tokens) with prompt caching at $0.0005 per deal. Forecast commentary uses Sonnet 4.6 at $0.022 per weekly commentary. Total AI COGS for 8 clients with 10 reps each: under $5/month against $3,192/month in revenue.
What happens when the rep logs no activities in their CRM?
Sonnet 4.6 will generate a placeholder coaching summary stating 'Insufficient CRM activity data to generate a coaching insight this week. Rep logged fewer than 3 activities.' Add a minimum-activity threshold check (at least 3 logged activities in 7 days) before generating the summary. Surface a 'Needs more CRM activity' badge on the rep card in the manager dashboard instead of a placeholder summary — this itself is actionable coaching intelligence.
Does this replace Gong?
No — and that's the right positioning. Gong's core value is call recording, transcription, and conversation intelligence. This platform does not record calls. The honest positioning: this is a CRM-activity intelligence and coaching layer for mid-market sales teams that have $10K/rep/yr in Gong budget but need a $40K/yr RevOps consulting product instead. The two products serve adjacent needs at different price points.
Want the production version?
- Delivered in 10–14 weeks
- You own 100% of the code
- AI cost monitoring built in
30-min call. No commitment.