Product Specification

WWP Customer
Intelligence Portal

Today, your contact data lives in separate systems that don't talk to each other. Pipedrive knows about deals. Kit knows about email engagement. Circle knows who's active in the Academy. Fathom knows who's been on a call. There's no way to ask "who are our most engaged prospects right now?" without manually checking each one.

This project builds a system that connects them, merges contact records into unified profiles, computes an engagement score for every person, and makes all of it available through three interfaces: a web portal, a Claude.ai connection, and automated alerts delivered through email, Slack, or whatever channel makes sense for your team.

Interface 1

Web portal

A dedicated website at portal.winwithoutpitching.com where Blair and the RTDs log in to see a lead dashboard, search contacts, view full profiles, review weekly digests, and track the pipeline. Data refreshes every 15–30 minutes.

Interface 2

Claude.ai connection

The same data is accessible inside Claude.ai through an MCP connection. You can ask "tell me everything about Sarah Chen" or "who are the hottest EMEA leads this month?" and get answers drawn from all sources. This also works in Cowork alongside your other tools.

Interface 3

Automated alerts

When a contact's engagement score crosses a configurable threshold, the system notifies the RTD responsible for that contact's region. The notification includes the contact's full profile and a specific recommended next step written by an AI agent that has reviewed the contact's complete history across all sources. Alerts can be delivered via email, Slack, or any channel that fits your workflow — the delivery mechanism is configurable.

Data sources

What we pull from each system

Each source provides different data about your contacts. Here's exactly what we pull, how often, and what limitations exist. Understanding this is important because the quality of the scoring depends on the quality of the data going in.

Pipedrive — CRM

Pipedrive is the authoritative source for contact identity. When the system needs to decide a contact's canonical name or company, Pipedrive wins. It's also where deals and pipeline stages live.

Contact recordsName, email, phone, organization name, custom fields. Full person records via the Persons API.
DealsDeal title, value, stage, assigned owner, creation date, last activity. Each deal is linked to a person.
Deal stagesThe pipeline stage each deal is in (e.g., MQL, SQL, Proposal, Won, Lost). Used for opportunity tracking.
ActivitiesLogged calls, meetings, emails, tasks associated with a contact. Each has a type, date, and note.

SYNC DETAILS

Frequency: Every 15 minutes via API polling. Also receives webhooks for new contacts, new deals, and deal stage changes — these are processed in real time.

Authentication: API key (one per account).

Rate limits: 30,000 API tokens per day. A typical sync uses 200–400 tokens. No concerns at your scale.

What we write back: Phase 2 adds the ability to create MQL-status Opportunities in Pipedrive when a lead qualifies. We never modify existing records.

Limitations: None significant. Pipedrive's API is the most complete of the connected sources.

Kit — Email Marketing (formerly ConvertKit)

Kit tells us about email engagement: who subscribes, what they click, which sequences they're in, and which tags they carry. This is one of the most frequent engagement signals.

SubscribersEmail, first/last name, subscribe date, state (active, cancelled, bounced), custom fields including company.
TagsAll tags assigned to a subscriber. Tags often encode interest signals (e.g., "Pricing Creativity Interested").
SequencesWhich email sequences a subscriber is enrolled in or has completed.
Engagement eventsLink clicks, form subscriptions, tag additions, course completions. Captured via webhooks only (see caveat below).

SYNC DETAILS

Frequency: Subscriber list syncs every 30 minutes. Engagement events arrive in real time via webhooks.

Authentication: API key (V4 API).

Rate limits: 120 requests per minute. Manageable.

Caveat: no historical engagement data

Kit's API does not let us pull past open/click history per subscriber. We can only capture engagement events going forward via webhooks. This means that for existing subscribers, the system starts with limited engagement data — their scores will improve over the first few weeks as new interactions are captured. We do pull their subscribe date, tags, and sequences, which gives us a starting signal.

Circle — Academy Community

Circle tells us about Academy engagement: who's a member, what they post, which spaces they participate in, and whether they've completed courses.

MembersName, email, join date, membership status, space memberships.
PostsPost title, author (with email), space, creation date. Each post is linked to its author.
SpacesWhich community spaces exist and which members belong to each.
Course lessonsCourse enrollment and progress data where available.

SYNC DETAILS

Frequency: Every 30 minutes via API. Webhooks for new members and new posts.

Authentication: API key (Admin API v2).

Rate limits: 2,000 requests per 5 minutes. Very generous.

Caveat: requires Business plan

Circle's API is only available on their Business or Enterprise plan. We need to confirm that the WWP Academy's Circle plan includes API access before development begins. Also, Circle doesn't provide a built-in engagement score — we compute engagement ourselves by counting and weighting each member's posts, comments, and space participation.

Fathom — Call Recordings

Fathom captures what happens on sales and discovery calls: who attended, what was discussed, and what the AI-generated summary says. A call is one of the strongest buying signals.

Meeting listTitle, date, duration, who recorded it.
AttendeesName, email, and whether they're internal or external. The email is how we match a call attendee to a contact in our database.
AI summariesFathom's AI-generated meeting summary, including key topics discussed.
Action itemsAI-extracted action items from the call.
TranscriptsFull meeting transcripts with speaker labels and timestamps. Stored for reference but not indexed for search in Phase 1.

SYNC DETAILS

Frequency: Primarily webhook-driven — Fathom sends us a notification when a meeting finishes processing. We also poll for any missed meetings.

Authentication: API key (user-scoped).

Rate limits: 60 requests per minute.

Caveat: user-scoped API keys

Fathom API keys only see meetings recorded by the key owner, or meetings shared to their team. If multiple people at WWP record calls, we need to ensure all meetings are shared to a common team, or use an API key from someone with visibility into all relevant meetings. We need you to verify your Fathom team sharing settings.

Clay — LinkedIn Enrichment

When a new contact enters the system, we enrich their record with LinkedIn-derived data via Clay: job title, company name, company size, location, company domain, and estimated revenue range.

This is how we know that "sarah@altitudedesign.com" is a Creative Director at a 12-person firm in Sydney. It's also how we assign contacts to regions (NA, LATAM, EMEA, APAC) — the location data comes from here.

SYNC DETAILS

Frequency: Triggered automatically when a new contact is created. Not on a polling schedule — enrichment runs once per contact.

Authentication: API key (via Clay's enrichment API, following the same pattern we use in our existing Insight Engine).

Business emails enrich more reliably than personal (@gmail) addresses. Missing enrichment is treated as neutral — it doesn't penalize a contact's score.

GA4 — Website Analytics

NOT INCLUDED IN CURRENT SCOPE

GA4 could provide website behavior data (pages visited, session frequency, traffic sources), but its API only exposes aggregate metrics — it cannot tell us which specific contact visited which page. Attributing web behavior to known contacts requires a separate BigQuery integration plus adding a tracking tag to your website. The added complexity and setup burden don't justify the value at this stage, especially since the other five sources already provide strong per-contact engagement signals. If we find that website behavior data would meaningfully improve scoring accuracy, we can add it later as a standalone enhancement.

Gmail — Direct Email History

NOT INCLUDED — SECURITY CONCERNS

Connecting directly to Blair's Gmail would give the system access to historical email correspondence with leads — valuable context for understanding relationships. However, this requires granting programmatic access to a personal email account, which introduces significant security and privacy risks: broad read access to all mail, potential exposure of sensitive correspondence, and compliance concerns around storing personal email content in a third-party system. The risk doesn't justify the benefit. Fathom call summaries and Pipedrive activity notes capture much of the same relationship context through safer, scoped channels.

Data synthesis

How five sources become one contact

The same person might exist as "Sarah Chen" in Pipedrive, "sarah@altitudedesign.com" in Kit, user #4521 in Circle, and an attendee on a Fathom call. The system needs to recognize these are all the same person and merge their data into a single profile. Here's how that works.

Unified contact graph — Sarah Chen

Sarah Chen
sarah@altitudedesign.com
92
Score
97
Percentile
4
Sources
APAC · Creative Director · Altitude Design Group
Pipedrive CRM
Name: Sarah Chen
Org: Altitude Design Group
Deal: Masterclass Q3 ($4,500)
Stage: Open
Person ID: pd_4521 · synced 3m ago
Kit Email
Subscribed: Apr 15, 2026
Via: Pricing Creativity page
Tags: masterclass-interest
Clicks: 5 in last 30 days
Subscriber ID: kit_8923 · webhook
Circle Academy
Joined: Apr 20, 2026
Posts: 3 (Positioning space)
Comments: 2
Courses: Positioning (completed)
Member ID: circle_301 · synced 12m ago
Fathom Calls
Calls: 2 (Apr 28, May 9)
Last summary: "Ready to commit if timing works for her team"
Action items: Send Q3 dates
Matched by: sarah@altitudedesign.com
Clay LinkedIn
Title: Creative Director
Location: Sydney, Australia
Company size: 12 employees
Revenue: $1M – $5M
Enriched via: sarah@altitudedesign.com
Feeds into scoring engine ↓

Identity resolution: matching records across sources

The primary matching key is email address. When a subscriber in Kit has the same email as a person in Pipedrive, they're merged into one contact. This works for the vast majority of cases because email is consistently captured across all sources.

HOW EACH FIELD IS RESOLVED WHEN SOURCES DISAGREE

Name Pipedrive → Pipedrive is the authoritative source. If Kit says "S. Chen" and Pipedrive says "Sarah Chen", the profile shows "Sarah Chen".
Company Pipedrive Clay → Pipedrive org name is primary. Clay supplements with domain, revenue, and size data.
Title & Role Clay → Comes exclusively from Clay's LinkedIn enrichment. Not available in any other source.
Location Clay → Used for region assignment (NA/LATAM/EMEA/APAC). Falls back to email domain TLD if Clay enrichment is unavailable.
Engagement Kit Circle Fathom Pipedrive → All sources contribute activities. Every email click, Academy post, call, and deal action is stored as an activity record linked to the unified contact.

The contact graph

The result of this synthesis is a database we call the contact graph. Each contact has:

One identity — canonical name, email, company, title, location, region. Resolved from the sources above.

Source links — references back to their record in each original system (Pipedrive person ID, Kit subscriber ID, Circle member ID, etc.).

Enrichment data — LinkedIn profile information from Clay: headline, company size, revenue range, company domain.

Activity timeline — a chronological log of every engagement event across all sources. Each event has a source, type, date, and metadata (e.g., which email was clicked, what the call was about).

Score — a computed number reflecting their total weighted engagement, with history over time so you can see whether they're trending up or down.

Opportunities — any Pipedrive deals associated with this contact.

Lead scoring

How engagement becomes a number

The scoring engine takes every activity in a contact's timeline, applies a weight based on the activity type, applies a time decay based on how recent it is, and sums the result. The output is a single number that represents how engaged this person is right now. Here's the detail on each component.

Signal weights

Each activity type has a weight reflecting how strongly it signals genuine interest. These are the starting defaults — they'll be tuned based on what actually predicts conversions for WWP.

Attended a call (Fathom)2.0x
Deal created (Pipedrive)2.0x
Deal stage changed (Pipedrive)1.5x
Completed a course (Circle)1.5x
Posted in Academy (Circle)1.0x
Clicked an email link (Kit)1.0x
Subscribed via form (Kit)1.0x
Commented in Academy (Circle)0.75x
Joined Academy (Circle)0.5x
Tagged (Kit)0.5x

Recency decay

A signal's contribution to the score decreases over time. We use a 30-day half-life: an activity from today counts at full value, one from 30 days ago counts at half, one from 60 days ago counts at a quarter, and so on. This prevents leads who were active six months ago from sitting permanently at the top of the list.

SIGNAL VALUE OVER TIME
Today 30 days 60d 100% 50% half-life

Concrete example

Sarah Chen's score of 92 breaks down like this:

3 Academy posts this week (3 × 1.0 × ~100% recency) = 3.0
2 Fathom calls in last 2 weeks (2 × 2.0 × ~90% recency) = 3.6
1 deal created 7 days ago (1 × 2.0 × ~85% recency) = 1.7
5 email clicks over 3 weeks (5 × 1.0 × varying decay) = 4.1
Plus older activity with heavier decay...
Total: 92 (after accumulation of all historical activity)

Percentiles and thresholds

Raw scores aren't that useful on their own — "is 72 good?" depends on the distribution. So we also compute percentiles: a score at the 90th percentile means the contact is more engaged than 90% of your database.

The alert threshold is set at a percentile (default: 90th). Contacts above this threshold are considered "hot" and trigger notifications to RTDs. This means the threshold automatically adjusts as your database grows — it's always relative to your actual engagement distribution.

The MQL threshold (for automatic opportunity creation in Pipedrive) requires both: score above the alert threshold AND engagement from at least 2 different sources in the last 30 days. The multi-source requirement prevents a single burst of activity in one channel from creating a premature opportunity.

What we need your input on

The system launches with the defaults above, and they'll work well enough to start. But the weights and thresholds are the most important tuning decisions in the system — they determine who you spend time on. After 2–4 weeks of real data, we'll schedule a working session to review what's actually happening:

The portal

What you see when you log in

The portal is a web application that provides four main views: a dashboard overview, a searchable lead list, individual contact profiles, and weekly digests. Here are representative mockups of each. The data shown is fictional but the structure reflects the actual system.

portal.winwithoutpitching.com
  • Dashboard
  • Leads
  • Opportunities
  • Digests
Lead Intelligence
Last updated 3 minutes ago · 5 sources connected
Total contacts
2,847
+124 this month
Avg score
34.2
+2.1 vs last week
Above threshold
31
+8 since Monday
Open opportunities
14
-2 closed this week
Top Leads This Week View all →
ContactScoreChangeRegionTop Signal
Sarah Chen
Altitude Design Group		92+28APAC3 Academy posts, 2 calls
Marcus Webb
Webb & Partners		87+15NAPricing page 4x, email clicks
Ana Oliveira
Criativa Branding		84+31LATAMCompleted course, deal created
James Thornton
Thornton Creative		61+9EMEA5 email opens this week
Priya Sharma
Kaleidoscope Agency		58-4APACSubscribed to newsletter
By Region
NA
1,241
EMEA
764
APAC
512
LATAM
330
Source Health
Pipedrive3 min ago
Kit8 min ago
Circle12 min ago
Fathomvia webhook
Clayon demand
portal.winwithoutpitching.com/contacts/sarah-chen
  • Dashboard
  • Leads
  • Opportunities
  • Digests
Sarah Chen
Creative Director · Altitude Design Group · Sydney, Australia
APAC 4 sources
92
Lead score · 97th percentile
LinkedIn Profile (via Clay)
TitleCreative Director
CompanyAltitude Design Group
LocationSydney, Australia
Company Size12 employees
Revenue Range$1M – $5M
Score Breakdown
Circle (Academy)34.2
Fathom (Calls)28.0
Kit (Email)18.5
Pipedrive (CRM)11.3
Recent Activity
Posted "My positioning journey" in Academy
May 10, 2026
Discovery call — discussed Masterclass Q3
May 9, 2026
Commented on "Pricing for value" thread
May 8, 2026
Clicked "Masterclass details" in May newsletter
May 7, 2026
Posted "Value conversation frameworks" in Academy
May 5, 2026
Deal created: Altitude — Masterclass Q3
May 4, 2026
Intro call — positioning challenges
Apr 28, 2026
Subscribed via "Pricing Creativity" landing page
Apr 15, 2026
portal.winwithoutpitching.com/digests/2026-05-11
  • Dashboard
  • Leads
  • Opportunities
  • Digests
Weekly Digest
May 5 – May 11, 2026 · Generated Monday 9:00 AM

This week's highlight: APAC saw the strongest engagement growth, driven by Sarah Chen's deep Academy activity and two discovery calls. Ana Oliveira in LATAM completed the positioning course and immediately triggered a deal creation — a strong MQL candidate. Overall, 8 new contacts crossed the alert threshold, up from 3 last week.

APAC · 2 hot leads
Sarah Chen 92
Schedule a call to discuss the Q3 Masterclass — she attended 3 Academy sessions on positioning and her firm just grew to 12 people.
Priya Sharma 58
Send the pricing workshop invite — she subscribed last week and has been opening every email.
LATAM · 1 hot lead
Ana Oliveira 84
Follow up on the deal — she completed the positioning course and created a deal in Pipedrive. Strong MQL.
NA · 1 hot lead
Marcus Webb 87
He's visited the pricing page 4 times this week and clicked through the last 3 emails. Reach out with a Masterclass invite.
EMEA · 1 warm lead
James Thornton 61
Warming up — 5 email opens this week. Not ready for a call yet but worth keeping on the radar.
Claude.ai connection

Conversational access to the same data

The portal is for structured browsing. Claude.ai is for asking questions. We connect the intelligence portal to Claude.ai as an MCP server — a tool connection that lets Claude query your contact database, pull profiles, and run searches. You ask a question in natural language; Claude calls the appropriate tool behind the scenes and synthesizes the answer.

This works in claude.ai, in the Claude desktop app, and in Cowork mode where you can combine it with other tools. The four tools available are: query_contact (full profile by name or email), search_leads (filter by score, region, source), get_top_leads (ranked list), and get_opportunities (pipeline view).

claude.ai/chat/wwp-sarah-chen-profile
WWP Intelligence Portal / Sarah Chen — full profile
Share
Tell me everything about Sarah Chen at Altitude Design Group
Queried contact profile across Pipedrive, Kit, Circle, Fathom, and Clay
📌 Sarah Chen — Creative Director, Altitude Design Group

Sydney, Australia. Lead score: 92 (97th percentile, +28 this week). Engagement is concentrated across three sources, with a clear acceleration in the past 10 days.

📊 Score composition
Circle (Academy) ████████████░░░░ 34.2 ← highest Fathom (Calls) ████████░░░░░░░░ 24.8 Kit (Email) ██████░░░░░░░░░░ 18.6 Pipedrive (CRM) █████░░░░░░░░░░░ 14.4 Clay (Firmographic) ──── (context only, not scored)
🎯 Signals by source
  • Academy (Circle) — 3 posts this week on positioning. Completed positioning course in April.
  • Calls (Fathom) — 2 calls. May 9 summary: "ready to commit if the timing works for her team"
  • Email (Kit) — Subscribed via Pricing Creativity, Apr 15. Clicked Masterclass link in May newsletter.
  • CRM (Pipedrive) — Open deal: "Altitude — Masterclass Q3," created May 4.
  • Firmographics (Clay) — 12-person firm, est. $1M–$5M revenue, APAC region.
🟢 Recommended next step

Given the Fathom summary signaling timing readiness and the open Masterclass Q3 deal, the highest-leverage action is confirming Q3 cohort dates and sending a direct enrollment link. The Academy activity suggests she's already self-educating on the positioning curriculum — no warm-up needed.

Profile assembled from 5 connected sources. Score reflects weighted engagement with 30-day recency decay. Last refreshed 4 minutes ago.

Show her full activity timeline
Draft the enrollment email
Who else at Altitude is in the system?
Reply to Claude…
+
Opus 4.7 Adaptive ▾

Example queries

"Who are the top 10 leads in EMEA right now?"

"Show me everyone who attended a Fathom call this month but doesn't have a deal in Pipedrive yet"

"What's the pipeline looking like — how many open deals by stage?"

"Which leads went cold in the last 30 days? Whose score dropped the most?"

"Summarize Ana Oliveira's journey from first touch to deal creation"

"Compare engagement trends in NA vs APAC this quarter"

Alert examples

What notifications actually look like

Alerts and digests arrive as formatted emails (or Slack messages, depending on configuration). Each one is written by an AI agent that has read the contact's full profile across all sources. Here are examples of each type.

mail.google.com — WWP Lead Alert: Sarah Chen (score: 92)
WWP Lead Alert: Sarah Chen (score: 92)
WWP Intelligence <alerts@winwithoutpitching.com>
to me
May 10, 2026, 2:34 PM
LEAD ALERT · APAC
Sarah Chen
Creative Director · Altitude Design Group · Sydney, Australia
92
Score
+28
This week
97th
Percentile
Recommended next step

Schedule a call to discuss Q3 Masterclass enrollment. Sarah attended 3 Academy sessions on positioning this week, clicked through the Masterclass details in the May newsletter, and her Fathom call summary from May 9 notes she's "ready to commit if timing works for her team." Her firm is 12 people and growing — the Masterclass is a natural fit. Send Q3 dates and a direct enrollment link.

Recent activity
Circle — Posted "My positioning journey" May 10
Fathom — Discovery call: discussed Masterclass Q3 May 9
Kit — Clicked "Masterclass details" in May newsletter May 7
Pipedrive — Deal created: Altitude — Masterclass Q3 May 4
LinkedIn (via Clay): Creative Director, Altitude Design Group (12 employees, $1M–$5M revenue), Sydney, Australia
WWP Customer Intelligence Portal · This alert was triggered because Sarah Chen crossed the 90th percentile engagement threshold.
mail.google.com — WWP Weekly Digest: APAC (May 5–11)
Weekly Lead Digest — APAC (May 5–11, 2026)
WWP Intelligence <digests@winwithoutpitching.com>
to me
May 12, 2026, 9:00 AM
WEEKLY DIGEST · APAC
May 5 – 11, 2026
Strong week for APAC. Sarah Chen is your top lead globally right now — her Academy engagement accelerated and the second Fathom call shows real purchase intent. Priya Sharma is earlier in the funnel but opening every email. Two contacts above threshold, up from zero last week.
Sarah Chen
92 ↑+28
Creative Director, Altitude Design Group · Sydney
Schedule the Q3 Masterclass enrollment call. She posted 3 times in the Academy this week, had a follow-up call where Fathom captured "ready to commit if timing works," and clicked through the Masterclass link in the newsletter. Momentum is high.
Circle: 3 posts Fathom: 2 calls Kit: 5 clicks Pipedrive: deal open
Priya Sharma
58 ↓-4
Founder, Kaleidoscope Agency · Mumbai
Warming but not ready for a call. She subscribed last week and has opened every email since, but no Academy activity or calls yet. Send the pricing workshop invite to deepen engagement before reaching out directly.
Kit: subscribed, all opens
APAC this week:
512 total contacts · 2 above threshold · avg score 31.4 (+1.8)
WWP Customer Intelligence Portal · Weekly digest for APAC region · Delivered Monday 9:00 AM
mail.google.com — MQL Created: Ana Oliveira — Criativa Branding
MQL Created: Ana Oliveira — Criativa Branding
WWP Intelligence <alerts@winwithoutpitching.com>
to Blair
May 11, 2026, 11:15 AM
MQL OPPORTUNITY CREATED · LATAM
Ana Oliveira
Brand Strategist · Criativa Branding · São Paulo, Brazil
An MQL-status Opportunity has been created in Pipedrive
Deal: Criativa Branding — MQL
Stage: MQL (pending your review)
Assigned to: Blair (for vetting before RTD distribution)
Why this contact qualified

Ana met both MQL criteria: her engagement score is 84 (above the 90th percentile threshold) and she has activity from 3 different sources in the last 30 days (Circle, Kit, and Pipedrive). The multi-source engagement pattern suggests genuine, sustained interest rather than a single-channel spike.

Qualifying activity (last 30 days)
Circle — Completed positioning course May 8
Kit — Clicked 4 email links, completed sequence Apr–May
Pipedrive — Deal created by system May 11
LinkedIn (via Clay): Brand Strategist, Criativa Branding (8 employees, <$1M revenue), São Paulo, Brazil
Action needed: Review this MQL in Pipedrive. If the lead looks qualified, assign to the LATAM RTD. If not, close the opportunity with a reason so the scoring model can learn from it.
WWP Customer Intelligence Portal · MQL created automatically when score > 90th percentile + activity from 2+ sources in 30 days
Automations

What happens without you checking in

The portal and Claude.ai are pull interfaces — you go to them when you want information. The automations are push: the system monitors the data and reaches out when something needs attention. Here's the pipeline step by step.

Step 1

Data syncs

Pipedrive, Kit, Circle poll every 15–30 min. Fathom sends webhooks when calls finish. Kit sends webhooks for clicks, subscribes, tags.

Step 2

Scores recalculate

After each sync cycle, every contact's score is recomputed. New activities get weighted and decayed. Percentiles update.

Step 3

Threshold check

The system identifies contacts who crossed above the 90th percentile since the last check. These are the new "hot" leads.

Step 4

AI reviews profile

For each hot lead, an AI agent reads their full profile across all sources and writes a 2–3 sentence recommendation for the specific next step.

Step 5

RTD is notified

The contact is routed to their region's RTD via email, Slack, or another configured channel with the full profile and recommendation.

Real-time alerts

The recommendation is not a template. The AI agent reads the contact's actual history — their Fathom call summaries, their Academy posts, their email click patterns, their Pipedrive deal status — and writes a recommendation specific to that person. If Sarah Chen attended a call where she said she's "ready to commit if timing works," the recommendation will reference that.

The alert is routed to the RTD responsible for the contact's region based on their Clay-enriched location data. Delivery channel is configurable — email, Slack, or both.

Weekly digest (Monday 9 AM)

Every Monday, each RTD receives a region-filtered summary of the week's most engaged leads. Blair gets the full view across all four regions. The digest includes the top leads ranked by score, their score change over the week, and an AI-generated recommendation per lead. This is also viewable in the portal's Digests tab.

The digest is generated by the same AI agent that handles alerts — it queries the MCP tools, analyzes the data, and writes the narrative.

Automatic MQL creation in Pipedrive

When a lead meets two criteria — (1) score above the alert threshold AND (2) engagement from at least 2 different source systems in the last 30 days — the system creates an MQL-status Opportunity in Pipedrive. Blair reviews and vets these before they're distributed to RTDs. The multi-source requirement prevents false positives from single-channel bursts.

Google Sheets ↔ Pipedrive sync (Phase 3)

Each RTD gets a Google Sheet connected bidirectionally to Pipedrive. New Pipedrive deals appear in the sheet; updates in the sheet push back to Pipedrive. This lets RTDs manage their opportunities without needing a Pipedrive seat, saving the cost of four additional licenses.

Delivery phases

What gets built when

The system is delivered in three phases. Each phase produces a working, usable system — Phase 1 is not a prototype. The phasing reflects dependencies (you need data flowing before you can score, you need scoring before you can alert) and allows us to tune scoring with real data before automating decisions based on it.

01

Core Platform

Data flows in, contacts are unified, scores are calculated, and you can browse and query everything.

  • Connect Pipedrive, Kit, Circle, Fathom, Clay
  • GA4 excluded from initial scope (see Data Sources section)
  • Identity resolution merges records by email
  • Scoring engine running with default weights
  • Portal: dashboard, searchable lead list, contact detail pages
  • Claude.ai MCP connection (query_contact, search_leads, get_top_leads, get_opportunities)
  • Score tuning workshop after 2–4 weeks of data
02

Automations

The system starts acting on the data — alerts, digests, and opportunity creation in Pipedrive.

  • AI-powered lead alerts to RTDs when contacts cross threshold
  • Weekly digests (Monday 9 AM, region-filtered per RTD, via email/Slack)
  • Automatic MQL opportunity creation in Pipedrive
  • RTD notification routing by region
  • Digest view added to portal
03

Extended

Deeper integrations and feedback loops.

  • Portal-native opportunity management for RTDs (update stages, add notes, push to Pipedrive — replaces the need for Google Sheets or extra Pipedrive seats)
  • GA4 integration (if website behavior data proves valuable)
  • Score model refinement using conversion feedback (which scored leads actually bought?)
Next steps

What we need from you to begin

  1. Confirm Circle plan level
    The Academy's Circle plan must be Business or Enterprise tier for API access. If it's on a lower plan, we can't pull member or post data. Please confirm your current plan.
  2. Verify Fathom team sharing
    Fathom API keys are scoped to the key owner's account. If multiple people record calls, we need all meetings shared to one team. Designate who should own the API key and confirm sharing settings.
  3. Provide RTD region mapping
    We need the name and email address for each Regional Training Director: NA, LATAM, EMEA, and APAC. This determines who receives alerts and digests for each region.
  4. Provide API keys
    Pipedrive, Kit, Circle, and Fathom API keys. We store these encrypted in AWS Secrets Manager (not in code). We can walk you through generating each one.
  5. GA4 access (future, if added)
    GA4 is excluded from the initial scope. If we decide to add website behavior data later, we'll need read access to your GA4 property via a Google service account.
  6. Schedule scoring workshop
    After Phase 1 launches and 2–4 weeks of real data accumulates, we schedule a working session to review actual engagement patterns and tune the scoring weights. This is the most important step for making the system accurate.
Prepared by Newfangled · May 2026