Pipeline account review

Pipeline Account Review

Audit pipeline accounts by analyzing engagement data, mapping stakeholders, identifying gaps, and prioritizing next actions.

Instructions

When a user wants to review their pipeline, systematically analyze each account and produce an actionable summary.

Steps

1. Retrieve accounts by calling mcp__claude_ai_Amplemarket__list_accounts with relevant filters:

   • owner_email if the user wants to review their own accounts
   • tags if filtering by segment or category
   • name or domain if looking at specific accounts
   • Set page_size to 20 for a manageable review batch

   If no filters are specified, ask the user: "Would you like to review all accounts, or filter by owner, tags, or specific accounts?"

2. Get detailed data for each account by calling mcp__claude_ai_Amplemarket__get_account for each account ID returned. Extract:

   • Engagement stats (emails sent, opens, replies, meetings booked)
   • CRM opportunity data (stage, amount, close date)
   • AI-generated insights
   • Account tags and owner
   • Last activity date

3. Map stakeholders at key accounts by calling mcp__claude_ai_Amplemarket__search_people for the top-priority accounts with:

   • company_domains: [account domain]
   • person_seniorities: ["C-Suite", "VP", "Head", "Director", "Manager"]
   • full_output: true
   • page_size: 10

   Compare found stakeholders against the contacts already engaged (from account data) to identify gaps in the buying committee.

4. Score account health for each account on a scale of 1-10 based on:

   • Engagement recency (3 points): Last activity within 7 days = 3, within 30 days = 2, within 90 days = 1, older = 0
   • Engagement depth (3 points): Replies received = 3, opens only = 1, no engagement = 0
   • Stakeholder coverage (2 points): 3+ stakeholders engaged = 2, 1-2 = 1, none = 0
   • Pipeline stage (2 points): Active opportunity = 2, prospecting = 1, no opportunity = 0

5. Classify each account into priority tiers:

   • Hot (8-10): Active engagement, strong multi-threading, opportunity in pipeline
   • Warm (5-7): Some engagement but gaps in coverage or stalled momentum
   • Cold (1-4): No recent engagement, limited stakeholder access, no active opportunity
   • At Risk: Previously warm/hot but engagement has dropped significantly

6. Generate recommended actions for each account:

   • Hot: Advance the deal. Suggest specific next steps
   • Warm: Re-engage or multi-thread. Identify who to contact and with what message
   • Cold: Re-evaluate fit or re-approach. Suggest new angles or stakeholders
   • At Risk: Urgent action needed. Specific recovery plays

7. Format the pipeline review with:

   • Executive summary (total accounts, distribution by tier, key trends)
   • Priority-ordered account table with health scores
   • Detailed recommendations per account
   • Stakeholder gap analysis

Important Notes

• Limit the detailed stakeholder search (step 3) to the top 5-10 accounts to avoid excessive API calls.
• If reviewing a large pipeline, offer to paginate: "You have 45 accounts. Shall I review the first 20 and then continue?"
• Adjust health scoring criteria if the user specifies their own definitions of engagement quality.

Customizing Health Score Thresholds

The default thresholds (7-day / 30-day / 90-day engagement windows) work for typical mid-market sales cycles of 30-60 days. Adjust them for different selling motions:

• Enterprise sales (90-180 day cycles): Relax recency thresholds. Last activity within 30 days = 3 points (instead of 7 days). Within 90 days = 2. Within 180 days = 1. Enterprise deals naturally have longer gaps between touchpoints.
• SMB / transactional sales (7-21 day cycles): Tighten thresholds. Last activity within 3 days = 3 points. Within 7 days = 2. Within 14 days = 1. Stale SMB deals are likely dead.
• PLG / product-led motions: Add product usage signals if available. Deprioritize outbound engagement metrics and weight pipeline stage and stakeholder coverage more heavily.

Tell the user: "I'm using default health scoring for a mid-market sales cycle. Want me to adjust thresholds for your typical deal length?"

Examples

Example 1: Full Pipeline Review

User prompt: "Review my pipeline - my email is [email protected]"

What the skill does:

1. Calls mcp__claude_ai_Amplemarket__list_accounts with owner_email: "[email protected]".
2. Calls mcp__claude_ai_Amplemarket__get_account for each account.
3. Calls mcp__claude_ai_Amplemarket__search_people for top accounts to find stakeholder gaps.
4. Produces the pipeline review.

Example output:

---

PIPELINE REVIEW - [email protected]

Executive Summary

• Total accounts: 18
• Hot: 3 | Warm: 7 | Cold: 5 | At Risk: 3
• Average engagement score: 5.2/10
• Key concern: 3 previously active accounts have gone silent in the last 30 days.

Account Priority Table

Rank	Account	Health	Tier	Stage	Engaged Contacts	Last Activity	Top Action
1	Acme Corp	9/10	Hot	Negotiation	4	2 days ago	Send revised proposal
2	TechFlow	8/10	Hot	Demo Scheduled	3	5 days ago	Prep for demo, loop in CTO
3	DataPrime	8/10	Hot	Evaluation	3	1 day ago	Share case study
4	CloudBase	6/10	Warm	Prospecting	2	12 days ago	Multi-thread to VP Eng
5	FinServ Inc	5/10	Warm	Prospecting	1	20 days ago	Find champion in IT
...	...	...	...	...	...	...	...
16	OldCorp	2/10	Cold	None	0	90+ days ago	Re-evaluate ICP fit
17	StaleInc	3/10	At Risk	Stalled	2	45 days ago	Breakup email to re-engage
18	GhostCo	1/10	At Risk	Stalled	1	60 days ago	New stakeholder approach

Stakeholder Gap Analysis (Top 5 Accounts)

Account	Engaged Roles	Missing Roles	Action
Acme Corp	CRO, VP Sales, Dir Ops, CFO	CTO	Search for and engage CTO for technical buy-in
TechFlow	VP Eng, Dir Product, CTO	CFO/Finance	Loop in finance before pricing discussion
DataPrime	CMO, Dir Marketing, VP Sales	CTO/IT	No technical evaluator - find one before POC
CloudBase	VP Eng, Senior Dev	VP/Dir level decision maker	Need executive sponsor
FinServ Inc	Dir IT	Everyone else	Single-threaded - high risk, need 2+ more contacts

---

Example 2: Filtered Review by Tag

User prompt: "Show me account health for my enterprise accounts"

What the skill does:

1. Calls mcp__claude_ai_Amplemarket__list_accounts with tags: ["enterprise"].
2. Retrieves and scores each account.
3. Returns a focused enterprise pipeline review.

Example 3: Stalled Account Identification

User prompt: "Which accounts should I focus on this week? I want to find stalled deals."

What the skill does:

1. Retrieves all accounts for the user.
2. Scores and classifies each.
3. Filters to "At Risk" and "Warm" tiers with declining engagement.
4. Returns a focused action plan for the 5-7 accounts that need immediate attention this week.

Troubleshooting

Problem	Solution
No accounts returned	Fallback chain: 1) Remove all filters and call list_accounts to verify accounts exist. 2) Try without owner_email in case the email format is different. 3) Try searching by tags or name instead. 4) If still empty, inform user: "No accounts found. This may mean accounts haven't been imported into Amplemarket yet."
Account details are sparse	Score with available data and explicitly flag which scoring factors were affected. For example: "Engagement score: 0/3 (no activity data available). This score may underrepresent account health if engagement is tracked outside Amplemarket."
Too many accounts to review	Offer to batch the review. Start with accounts that have active opportunities, then move to prospecting-stage accounts.
Stakeholder search returns too many people	Narrow person_seniorities to "C-Suite" and "VP" only for large companies.
User wants to take action on a specific account	Pivot to the competitive-account-research skill for a deep dive, or build-targeted-lead-list to add missing stakeholders.
Engagement history data seems stale	Flag with: "[Data may be stale - last activity DATE]. This may not reflect recent interactions outside Amplemarket." Then suggest: "Want me to re-enrich this account's contacts for the latest signals?"
Person enrichment succeeds but company enrichment fails for a pipeline account	Fallback chain: 1) Use account-level data from get_account as the primary source. 2) Try enrich_company with domain instead of name. 3) Try the LinkedIn company URL. 4) Score based on internal engagement data and flag: "Company enrichment unavailable - scoring based on internal data only."