{"name":"Customer Intelligence Agent","description":"Tells your sales team which customers matter most, which are slipping away, and who to call this week. Send your sales history (customer, date, amount) as a file or a paste — from Excel, Salesforce, HubSpot, QuickBooks, NetSuite, SAP and others — and get A/B/C/D customer tiers, the money at risk if quiet customers are not contacted, and a ranked call list with a reason for each account. The same data always gives the same answer, so results can be trusted and checked.\n\nFor developers: scoring uses published statistical models, not a language model (Pareto/ABC, RFM, BG/NBD 2005, Gamma-Gamma 2005, Markov tier migration, bootstrap ranges with a fixed seed; validated on the CDNOW benchmark), dated to the last date in the data, so identical input gives identical output on every call. Gemini only words replies to typed questions. Results are available to other agents as JSON (result/v1 extension); sign-in via OAuth2 for Google Cloud Marketplace customers.","url":"https://customer-tiering-agent-gcp.saifs.ai/a2a","protocolVersion":"0.3.0","preferredTransport":"JSONRPC","version":"1.3.2","provider":{"organization":"Saifee Software Solutions","url":"https://saifs.ai"},"documentationUrl":"https://saifs.ai","supportedInterfaces":[{"url":"https://customer-tiering-agent-gcp.saifs.ai/a2a","protocolBinding":"JSONRPC","protocolVersion":"1.0"}],"capabilities":{"streaming":true,"pushNotifications":false,"extensions":[{"uri":"https://a2ui.org/a2a-extension/a2ui/v0.9.1","description":"Renders results as A2UI v0.9.1: book summary, money at risk, customer tiers, the weekly call list and top accounts.","required":false,"params":{"supportedCatalogIds":["https://www.gstatic.com/vertexaisearch/a2ui/v0_9/gemini_enterprise_composite_catalog.json","https://a2ui.org/specification/v0_9/catalogs/basic/catalog.json"]}},{"uri":"https://a2ui.org/a2a-extension/a2ui/v0.9","description":"Renders results as A2UI v0.9 (Gemini Enterprise).","required":false,"params":{"supportedCatalogIds":["https://www.gstatic.com/vertexaisearch/a2ui/v0_9/gemini_enterprise_composite_catalog.json","https://a2ui.org/specification/v0_9/catalogs/basic/catalog.json"]}},{"uri":"https://customer-tiering-agent-gcp.saifs.ai/demo/extensions/result/v1","description":"Machine-readable result for other agents. Activate it (X-A2A-Extensions header, message.extensions, or message.metadata['customerIntelligence/result']=true) and the reply carries a JSON DataPart (metadata.mimeType 'application/json') with tier counts and members, every customer's tier, status, chance still buying, 12-month value and money at risk, this week's call list with actions and reasons, and the book's money at risk. The URI serves the JSON Schema. Without it, the reply is the sentence and A2UI screens Gemini Enterprise shows.","required":false,"params":{"mimeType":"application/json","schemaUrl":"https://customer-tiering-agent-gcp.saifs.ai/demo/extensions/result/v1"}},{"uri":"https://cloud.google.com/marketplace/docs/partners/ai-agents/setup-dcr","description":"Dynamic Client Registration for Gemini Enterprise customers.","required":false,"params":{"target_url":"https://multiplewords.in/gcp/agent/dcr"}}]},"defaultInputModes":["text/plain","text/csv","application/json","application/vnd.openxmlformats-officedocument.spreadsheetml.sheet","application/vnd.ms-excel","application/vnd.oasis.opendocument.spreadsheet","application/xml","text/xml"],"defaultOutputModes":["text/plain","application/json"],"skills":[{"id":"customer_segmentation","name":"Customer segmentation","description":"Segments every customer into A/B/C/D tiers from purchase history, merging spelling variants of the same customer automatically, and shows how revenue is concentrated across the book. Send transaction rows as CSV, a table, JSON or an attached file with three fields: customer, date and amount. Common column names such as 'Party Name', 'Invoice Date' and 'Grand Total' are recognised. Rows sent this way are analysed exactly as received, at any size, without being retyped by an AI model. Deterministic: the same data gives the same answer on every call.","tags":["customer segmentation","ABC analysis","customer intelligence","sales"],"examples":["Segment my customers from this order export","Who are my A accounts?","Are we too dependent on a few customers?"],"inputModes":["text/plain","text/csv","application/json","application/vnd.openxmlformats-officedocument.spreadsheetml.sheet","application/vnd.ms-excel","application/vnd.oasis.opendocument.spreadsheet","application/xml","text/xml"],"outputModes":["text/plain","application/json"]},{"id":"churn_and_risk","name":"Churn and revenue risk","description":"Flags customers who are slipping, going quiet or likely to drop a tier, and estimates the revenue at risk across the book. Send transaction rows as CSV, a table, JSON or an attached file with three fields: customer, date and amount. Common column names such as 'Party Name', 'Invoice Date' and 'Grand Total' are recognised. Rows sent this way are analysed exactly as received, at any size, without being retyped by an AI model.","tags":["churn","retention","revenue at risk","customer health"],"examples":["Which customers are at risk of churning?","How much revenue is at risk?","Who hasn't ordered in a while?"],"inputModes":["text/plain","text/csv","application/json","application/vnd.openxmlformats-officedocument.spreadsheetml.sheet","application/vnd.ms-excel","application/vnd.oasis.opendocument.spreadsheet","application/xml","text/xml"],"outputModes":["text/plain","application/json"]},{"id":"customer_value_growth","name":"Customer value and growth","description":"Estimates each customer's predicted value and identifies accounts with upside — B customers that can become A, and where rep time earns the most. Send transaction rows as CSV, a table, JSON or an attached file with three fields: customer, date and amount. Common column names such as 'Party Name', 'Invoice Date' and 'Grand Total' are recognised. Rows sent this way are analysed exactly as received, at any size, without being retyped by an AI model.","tags":["customer lifetime value","upsell","growth","account management"],"examples":["Which B accounts can become A?","What is this customer worth?","Where should reps spend their time?"],"inputModes":["text/plain","text/csv","application/json","application/vnd.openxmlformats-officedocument.spreadsheetml.sheet","application/vnd.ms-excel","application/vnd.oasis.opendocument.spreadsheet","application/xml","text/xml"],"outputModes":["text/plain","application/json"]},{"id":"rep_call_queue","name":"Rep call queue","description":"Builds a prioritised list of who to contact this week, grouped into save, grow and verify actions, with the reason for each account. Send transaction rows as CSV, a table, JSON or an attached file with three fields: customer, date and amount. Common column names such as 'Party Name', 'Invoice Date' and 'Grand Total' are recognised. Rows sent this way are analysed exactly as received, at any size, without being retyped by an AI model.","tags":["sales","prioritisation","outreach"],"examples":["Who should my reps call this week?","Build a save / grow / verify list"],"inputModes":["text/plain","text/csv","application/json","application/vnd.openxmlformats-officedocument.spreadsheetml.sheet","application/vnd.ms-excel","application/vnd.oasis.opendocument.spreadsheet","application/xml","text/xml"],"outputModes":["text/plain","application/json"]},{"id":"agent_to_agent_results","name":"Results for other agents","description":"For agent-to-agent pipelines: activate the result/v1 extension and the reply carries the complete result as JSON (tiers and members, every customer's tier, status, chance still buying, 12-month value and money at risk, the weekly call list with actions and reasons, and how the input was read), validated against a published JSON Schema. Several books per request are scored separately and in parallel. Results are deterministic, so pipelines can compare runs and cache by input.","tags":["agent-to-agent","A2A","JSON","automation","pipeline"],"examples":["Score this CRM export and return the call list as JSON","Score each branch file and return one result per book"],"inputModes":["text/plain","text/csv","application/json","application/vnd.openxmlformats-officedocument.spreadsheetml.sheet","application/vnd.ms-excel","application/vnd.oasis.opendocument.spreadsheet","application/xml","text/xml"],"outputModes":["text/plain","application/json"]}],"iconUrl":"https://multiplewords.nyc3.cdn.digitaloceanspaces.com/mw_assets/saifs_logo.svg","securitySchemes":{"oauth2":{"type":"oauth2","description":"Sign-in for Google Cloud Marketplace customers.","flows":{"authorizationCode":{"authorizationUrl":"https://multiplewords.in/gcp/agent/oauth/authorize","tokenUrl":"https://multiplewords.in/gcp/agent/oauth/token","refreshUrl":"https://multiplewords.in/gcp/agent/oauth/token","scopes":{"agent":"Use Customer Intelligence Agent"}}}}},"security":[{"oauth2":["agent"]}]}