Sales Assistant Agent Playbook: Lead Follow-Up to Quote Drafts

A sales assistant Agent helps B2B/B2C sales teams: score and follow up leads, retrieve product/price knowledge, draft follow-up and quote copy, and—within policy—read CRM or write tasks/notes; high-value haggling, contracts, and discount approvals hand off to humans. Unlike “give every rep ChatGPT,” bespoke delivery follows your CRM, price entitlements, and brand voice. GeonAI designs enterprise Agents; /agents entries are capability references only—not a public trial of all 363+ presets. Email [email protected], subject “Enterprise Agent inquiry”.

Map the funnel: what is worth automating?

Principle: automate high-frequency, retrievable, auditable work first—not “close every deal.” Selection: /blog/ai-agent-selection-checklist. Demo risks: /blog/ai-agent-project-red-flags.

Scenario 1: Lead follow-up—cadence, not spam

Good fit: segment by source/intent → personalized first-touch drafts → cool-down reminders or auto-send with explicit auth. Never invent scarcity or unapproved discounts. B2B needs compliant frequency and consistent persona.

1. Define ~8–12 sales intents: pricing, samples, lead time, certifications, competitive compare
2. Ask for missing slots (volume, region, budget) before pitching
3. Log touches in CRM to prevent multi-rep spam
4. Opt-out/complaint keywords → stop outreach and hand off

Scenario 2: Product & quote copy—RAG + price entitlements

Talk tracks must come from a governed source (effective price lists, specs, certifications) with citations. Quote drafts stay inside entitled bands/templates; over-threshold discounts force approval. KB guardrails: /blog/enterprise-rag-knowledge-base-agent.

• Price lists with effective dates and customer tiers—no hard-coded “one price” in prompts
• Competitive compare only from approved sheets—no free-form attacks
• Lead time/capacity from ERP or fixed promise fields—no guessing
• Default “pending rep send”; one-click send by role if needed

Scenario 3: CRM—read first, write carefully

Writing opportunity amount/stage is riskier than notes—most pilots allow tasks/notes only. Channel patterns: /blog/custom-customer-service-agent.

Suggested pilot (6–10 weeks)

1. One team + one CRM: lead scoring + follow-up drafts (human send)
2. Add price/spec RAG with citations; no invented discounts
3. Write-back limited to tasks/notes; discounts/contracts stay human
4. KPIs: follow-up SLA, draft adoption, wrong-promise rate, rep CSAT

Metrics: conversion only if auditable

• Follow-up SLA (lead in → first meaningful touch)
• Draft adoption (edited-then-sent share)
• Wrong-promise rate (price, lead time, certs) → ~0
• Pipeline movement vs control group (avoid attribution theater)
• Handoff win/loss fields complete for review

GeonAI delivery and what to prepare

Bring: CRM + roles, top 30 sales Q&A/objections, price/spec samples, follow-up SOP, discount approval rules, target KPIs. Delivery follows our 4-step playbook; pricing factors /blog/custom-ai-agent-pricing-factors. Contact [email protected], /pricing, or Live chat.