SuperAPI AI Directory
Partner中文AI newsFavorites
Skillslcrm-visit-assistant

lcrm-visit-assistant

Marketing & growth

View repository

Updated 10-15 21:36

About this skill

Visit Assistant

Objective

When the user mentions going to visit a customer, based on the product positioning (pre-sales digital employee + Langtum platform), output a complete visit preparation plan.

Knowledge Base References

Load the following documents (relative to this skill directory):

  • Product System Overview:knowledge/claude.md
  • Pre-sales Digital Employee Product Description:knowledge/数字员工/售前数字员工产品说明书.md
  • Langtum Platform Introduction:knowledge/langtum/什么是langtum.md
  • Langtum Platform Architecture:knowledge/langtum/Langtum平台架构.md
  • Langtum Use Cases:knowledge/langtum/Langtum使用场景.md

Search Tools

Use the most advanced search tool currently available to search for background information on the target customer.

Workflow

Step 1: Confirm the product to sell

Clarify what product the user wants to sell this time:

Customer Characteristics Recommended Product
Complex SKUs/frequent bidding/B2B/direct sales team Primary recommendationPre-sales Digital Employee
Does not match the above profile, but has AI needs or business process optimization needs RecommendLangtum Platform, customized by scenario

If the user has not clearly stated it, prepare for Pre-sales Digital Employee by default, but confirm: "Are you selling the Pre-sales Digital Employee this time, or the Langtum Platform?"

Step 2: Research customer background

Use the most advanced search tool currently available to search for the target customer:

  • Basic company information (founding time, scale, financing/listing status)
  • Main products/services and SKU complexity (key: order of magnitude of product models)
  • Industry classification (which of the 9 major industries it matches)
  • Sales model (direct sales/channel, bidding frequency)
  • Whether it is a customer in the profile list (yes/no/similar)

Step 3: In-depth fit analysis

High-fit customers (strongly recommend investment)

Must simultaneously have:

  • Product characteristics: SKU order of magnitude in the thousands or above, or hundreds of product models but with an extremely large number of configuration combinations
  • Sales characteristics: B2B business, high average order value (100k+), long sales cycle, frequent bidding
  • Team characteristics: has a direct sales team, pre-sales personnel ratio exceeds 1:3, pre-sales work accounts for more than 30% of the sales cycle
  • Industry characteristics: belongs to the 9 key industries, especially industrial control, security, ICT, mechanical equipment, power equipment

Medium-fit customers (can follow up but need verification)

Has some characteristics:

  • Many SKUs but sales model leans toward channels
  • Bidding exists but frequency is not high
  • Industry matches but scale is relatively small
  • Pre-sales team exists but efficiency issues are not prominent

Low-fit customers (Pre-sales Digital Employee is not suitable; evaluate Langtum entry point)

Has any of the following characteristics:

  • Single standardized product (SKU<100)
  • Mainly channel sales, no need for complex pre-sales support
  • Low average order value (<10k), volume-driven model
  • No direct sales team, relies on distributors

Langtum Platform entry point evaluation:

  • Are there other complex business processes that need AI assistance? (customer service/operations/supply chain)
  • Is there an internal knowledge management need?
  • Are you willing to build your own AI application?

Step 4: Design entry points based on the five core capabilities

Based on customer characteristics, select the 1-3 most matching core capabilities as entry points:

① Deep knowledge ingestion and precise Q&A

  • Applicable: Tens of thousands of SKUs, fragmented parameters, sales cannot remember product details
  • Script: "Do sales often ask headquarters for help because they do not understand the parameters?"

② Consultative demand guidance

  • Applicable: Non-standard customer requirements, need to translate technical specifications, lock onto the only solution from tens of thousands of SKUs
  • Script: "In the pre-sales stage, do you rely heavily on veteran experts' experience to clarify requirements?"

③ Deterministic logical reasoning and calculation

  • Applicable: Strict physical/electrical constraints, one wrong parameter ruins everything, need compatibility verification
  • Script: "When manually configuring the BOM, do you often configure it wrong so it cannot be installed, or calculate it wrong and lose money?"

④ Strong compliance and bid document generation

  • Applicable: Frequent bidding, state-owned and central enterprise customers, sensitive to regulatory red lines, need to generate deviation tables
  • Script: "Does the pre-sales team spend a lot of time writing bid documents, making deviation tables, and checking compliance?"

⑤ Solution and quotation closed loop

  • Applicable: Building-block solution sales, configuration lists are lengthy and prone to omissions, need repeated cross-department verification
  • Script: "Does the quotation process require repeated verification between technical and commercial teams, worrying about omissions?"

Step 5: Output the visit strategy

Output according to the following framework:

1. Fit assessment (get straight to the point)

匹配度：⭐⭐⭐⭐⭐（高/中/低）

匹配点：
- 产品线：XX个SKU，符合模块化配置特征
- 销售模式：直销为主，年招投标XX次
- 行业：属于9大行业中的XX

风险点：
- XX特征可能导致匹配度降低
- 建议验证XX问题

2. Core entry points (maximum 3, sorted by priority)

Each entry point includes:

  • Label: Which of the five core capabilities it is
  • Current situation: How the customer currently does it (pain point scenario)
  • Quantified pain: How long it takes, what the error rate is, what business metrics are affected
  • Digital employee solution: Which specific step it solves
  • Value quantification: Efficiency improvement X%, error rate reduction X%, capacity release X person-days/month

3. One-sentence value proposition

Format: "Make your pre-sales team go from 'everyone is an expert' to 'ordinary sales + digital employee = expert-level output', with capacity multiplied by X times"

4. Key question list (by dimension)

Business dimension：

  • Product SKU volume? Configuration combination complexity?
  • Annual number of bids? Average bid document preparation time?
  • Pre-sales team size? Ratio to sales?
  • How long is the sales cycle? What proportion is pre-sales work?

Technical dimension：

  • Status of product material accumulation? (PDF/Excel/database)
  • Existing pre-sales tools? (CRM/configurator/knowledge base)
  • Difficulty of IT architecture integration? (Private deployment/cloud deployment preference)

Commercial dimension：

  • Budget ownership (IT/sales/procurement)?
  • Decision chain (Who makes the final call? Who influences? Who uses it?)
  • Are you willing to pilot on a small scale?

5. Script preparation

Opening statement (capture attention in 15 seconds)：

"You make XX products, so there should be quite a few models, right? When sales quote, do they often have to look through materials to check parameters, or find a veteran expert to confirm?"

Handling objections：

  • "We have our own pre-sales team" → "It's not a replacement, it's amplification. Let senior pre-sales do high-value work, and digital employees do repetitive labor"
  • "AI is unreliable" → "Start with deterministic work like bid document parsing, which doesn't involve creative decisions, and accuracy can reach 95%+"
  • "Data security" → "Private deployment, data never leaves your servers. You can start with organizing public information"

Next steps：

  • Ask for a desensitized tender document/product manual for an on-site demo
  • Meet key decision-makers (Sales VP/Pre-sales Director/CTO)
  • Pilot: one region/one product line/one type of bid document

6. Risk warnings and stop-loss line

  • Competitor activity (whether similar solutions are already in use)
  • Internal resistance (pre-sales team worries about being replaced)
  • Data preparation cost (whether materials are structured)
  • Stop-loss line: If XX situation occurs, recommend abandoning or adjusting the strategy

Output style requirements

  • Structured: Use tables and layered lists; long blocks of text are prohibited
  • State the conclusion directly: Put the match rating first
  • Speak plainly: Use the customer's business language and avoid AI terminology (don't say RAG or Agent; say "intelligent knowledge base" or "digital employee")
  • Quantify value: Every value point must have numbers (X times, X%, X hours)
  • Endgame thinking: Don't just talk about features; talk about "scaling capacity" and "competitive barriers"
  • Tell the truth: There must be a truthful summary at the end, pointing out the real risks or uncertainties

Core principles

  1. Look at the endgame first: If this customer is won, what will the situation be in 3 years? Can they become a benchmark?
  2. Speak plainly and avoid jargon: Customers don't understand "Agentic RAG," but they understand "sales can get an instant answer to anything they ask"
  3. Value before features: Don't say "we can do bid documents"; say "triple pre-sales capacity"
  4. Dare to say no: Point out low-match customers directly and don't waste each other's time
  5. Leave room: All judgments are based on public information and need to be verified in practice