---
name: vendor-overlap-review
description: Analyze vendor and counterparty-level spend through Control MCP, identify overlapping services and redundant subscriptions, and quantify savings opportunities. Use this for vendor spend, duplicate tools, SaaS sprawl, subscription audits, recurring-charge reviews, and cost-cutting questions, even when the request does not use the words vendor or overlap.
---
# Vendor overlap and savings review
## Prerequisites
A Control MCP connector must be available. If several Control connectors are connected, confirm which workspace to use
before pulling data. Verify that the first result is non-empty; an empty result usually means the wrong workspace.
## Workflow
1. Pull counterparty-level spend with `get_group_transactions_browse`, or `get_entity_transactions_browse` when the user
names specific entities. Use a trailing twelve months unless the user gives a period, set `counterpartyRole` to
`vendor`, and group by counterparty name with a summed value column.
2. Restrict the query to expense accounts. Excluding balance-sheet group accounts such as receivables, payables, bank,
and VAT is more reliable than excluding individual accounts. Exclude zero-value technical lines.
3. Merge counterparty names that represent the same real-world vendor before analyzing. Paginate with `offset` when the
result is large and state the coverage instead of truncating silently.
4. For ambiguous vendors, fetch their transactions with a counterparty filter to review descriptions, frequency, and
account coding. Categorize by what the vendor sells, not by the account the spend landed on, and flag suspected
miscoding separately.
5. Assign each vendor to a functional category, for example cloud and infrastructure, developer tooling, AI APIs,
communication, productivity, design, marketing and analytics, sales, finance tools, professional services, HR and
payroll, security and compliance, travel, meals, facilities, insurance, and telecom.
6. Within each category, flag direct overlap between vendors providing substantially the same service; recurring
low-value charges that never vary; unit-price increases without matching usage growth; billing-plan mismatches such
as monthly billing for permanent tools or per-seat licenses that may exceed headcount; and duplicate charges of the
same amount from the same vendor in one month.
7. Flag concentration and accrual gaps in lumpy professional-services spend separately from savings.
## Rules
- Mark every item high, medium, or low confidence and state what would confirm it, such as usage data, seat counts, or
contract terms. Overlap inferred from vendor names alone is a hypothesis, not a finding.
- Never invent vendors or amounts. Every figure must trace to fetched rows.
- Report spend as expense-side P&L value and ignore VAT-only lines.
- Keep the default group-currency conversion so vendors are comparable across entities, and label every figure with the
group reporting currency from `get_group_accounting_settings`. Only pass `convertToGroupCurrency: false` when the user
asks for original transaction currencies, and then report per-entity figures separately instead of summing across
currencies.
- If the requested period extends beyond available data, say so instead of extrapolating.
## Output
Return a short prose summary and one table with category, vendors, twelve-month spend, finding, estimated annual saving,
and confidence. Follow it with a ranked action list of cancel, consolidate, renegotiate, or re-code decisions, with the
largest three quantified per year. State total addressable savings as a range. Include the Control app link returned by
the tool so the user can verify the underlying transactions.
