AI agent specialist with proven, reusable agents for corporate sales, business development

We help corporate companies in Asia-Pacific put AI agents to work in their sales, business development and operations teams. Our clients include insurers, telcos and financial-services companies with large partner and account portfolios. They want agents that save their teams time on account preparation, CRM work, follow-ups, operations and reporting. We are not looking for someone to build agents from scratch. We are looking for a specialist who has already built and deployed sales, BD and operations agents at several corporate companies, and who can adapt those proven agents to a new client quickly, safely and at a predictable price. Agents we want to offer (show us which you already have in production) Sales, business development and account management Account and meeting prep brief: a one-page brief before a client or partner meeting, from CRM history, emails, open issues, KPIs and public news. Meeting-to-CRM capture: after a meeting, drafts the CRM update, action list and follow-up email for approval. Follow-up and outreach drafting: personalised emails and sequences from CRM and account context. Proposal / RFP first draft from approved past answers and product documents. Product, pricing and contract Q&A with citations to the source document. Pipeline and CRM hygiene: flags stale deals, missing data and next steps. Account or partner performance pack / QBR draft from CRM and data-warehouse figures. Operations and reporting Inbox and ticket triage: summarises threads and drafts replies. Meeting minutes and action tracking. Recurring report automation: weekly and monthly reports with drafted commentary from approved figures. SOP and process Q&A for operations staff, with citations. Partner file intake and reconciliation support: checks incoming partner files, investigates mismatches and drafts evidence. It never decides a match or an amount. Requirements Agents in production at at least 3 corporate companies (500+ employees), with real users. Not demos or tutorials. At least one sales or BD agent and one operations or reporting agent among them. A reusable library or templates you can deploy at a new client in days or weeks, not months. Experience building inside the client’s environment: their cloud tenant, single sign-on, security review and data-protection rules. Human-in-the-loop design: agents read and draft; humans approve anything that is sent or changes records. Logging and an audit trail for every action. Evaluation: test sets, acceptance thresholds and monitoring (e.g. LangSmith, Langfuse or the cloud-native equivalent). Clear stakeholder work: running a discovery workshop with sales and operations teams, writing documentation and training users. Technical skills Microsoft 365 Copilot / Copilot Studio, Salesforce Agentforce, HubSpot or Google Gemini Enterprise. LLM APIs (OpenAI, Anthropic, Gemini, Azure OpenAI) and an agent framework (LangGraph, OpenAI Agents SDK, MCP) or n8n/Make. RAG over company documents, with vector databases. Integrations: CRM APIs (Salesforce, Dynamics, HubSpot), Microsoft Graph or Google Workspace, data warehouses. Python or TypeScript; deployment on AWS, Azure or GCP. Nice to have Agents built for insurers, insurtechs, banks or telcos, ideally in Asia-Pacific. Knowledge of AI governance rules such as the Singapore MAS AI guidelines, PDPA (Singapore, Malaysia, Thailand) or the EU AI Act. Multilingual agents (Thai, Bahasa, Vietnamese, Chinese, Japanese). A small team that can take on several client projects at the same time. What we want to agree with you Phase 1: a demo of your agent library (recorded or live, with client data hidden), plus a written overview per agent: what it does, stack, typical setup time, typical results, running cost per user per month. Phase 2: adapt and deploy two agents for a first client as a fixed-price pilot, with agreed acceptance thresholds. After that: ongoing work per client, plus maintenance. Commercial model: please propose a setup fee per agent per client, a monthly fee if applicable, and your hourly rate for custom work. Reuse rights: please state clearly what we and our clients may reuse, rebrand and own (code, prompts, workflows, test sets), and what stays yours. An NDA and data processing agreement are signed before any client data is shared. Screening questions List the corporate companies (or their industry and size, if confidential) where your sales/BD and operations agents are in production. For each: users, how long it has run, and the measured result. Which agents in your library can you redeploy at a new client within 2–4 weeks? What has to be customised each time? Describe one agent that failed or was cancelled at a client. What did you learn? How do you give each agent its own identity and least-privilege access in a client’s tenant, and how do you prevent data leakage and prompt injection? How do you evaluate a meeting-prep agent and a contract Q&A agent before go-live? Which metrics, which tool, which pass threshold? When do you choose Copilot Studio or Agentforce over a custom agent on a cloud platform? What does a typical agent cost a client per user per month to run? Can you give two references from corporate clients we can speak with? Red flags (we will not proceed if…) Only demos, tutorials or one-person startups; no corporate deployments. Wants to host the agents or client data in your own SaaS or account. No evaluation or monitoring approach, or “users liked it” as the measure. Agents that send messages or change records without human approval. No answer on running costs, data protection or reuse rights.