Recommendations, Upsell, and Rules: From Algorithms and Guardrails to the Launch of UCP
Most ERPs are storage engines, not sales engines. Sales agents forget promotions and management's sales strategy. Management loses control over price ceilings or upsell opportunities. Learn how artificial intelligence can increase B2B sales using recommendation algorithms and essential business rules (guardrails).
Foreword
Economic Impact
Companies that excel at personalization (McKinsey).
Business customers wanting proactive offers.
Sales growth for assisted junior agents.
See AI performance indicators in sales
Explore the guide
How does AI + rules work?
COMPLETE PRODUCT CATALOG (e.g. 1 million)
│
▼
STEP 1: RETRIEVAL/GENERATION (High speed, wide coverage)
"Find 500 similar products"
(Vector search + lexical filters)
│
▼
CANDIDATES (e.g. 500 products)
│
▼
STEP 2.1: SCORING / RANKING (Medium speed)
"Top items by purchase probability (pCVR)"
(Deep Learning neural network)
│
▼
STEP 2.2: RE-RANKING AND RULES (Fast)
"Remove zero stock and low margin"
(Post-processing via guardrails)
│
▼
FINAL RESULT (e.g. 3 recommended products)
See the Tech: Deep Learning
What does a hybrid architecture look like?
INTERNAL DATA ZONE MIDDLEWARE ZONE AI ZONE
┌─────────────────────────────┐ ┌─────────────────────────────┐ ┌─────────────────────────────┐
│ 1. DATA SOURCES │ │ 2. DATA HUB & SYNC │ │ 3. Vertex AI (Search + │
│ │ │ │ │ Search for commerce, │
│ ┌───────────┐ ┌───────────┐ │ │ ┌─────────────────────────┐ │ │ Model Garden) │
│ │ ERP │ │ FILE SRV │ │ │ │ Fast database │ │ │ ┌─────────────────────────┐ │
│ │ (Master) │ │ (PDF/Docs)│ │ │ │ (Cache) │ │ │ │ Vectors │ |
│ └─────┬─────┘ └─────┬─────┘ │ │ └────────────┬────────────┘ │ │ │ Index │ |
│ ▼ ▼ │ │ ▼ │ │ └────────────▲────────────┘ │
│ [Secure unidirect. Sync] │────>│ [Filtering & sanitization] │ │ │ │
└─────────────────────────────┘ │ │ │────>│ [Ingestion (public only)] │
│ ▼ │ │ │
│ ┌─────────────────────────┐ │ │ ┌─────────────────────────┐ │
│ │ API Gateway │ │<────│ | Model / LLM | │
│ │ (Combine Internal + ERP)│ │ │ | (RAG & Processing) | │
│ └────────────▲────────────┘ │ │ └─────────────────────────┘ │
└──────────────┼──────────────┘ └─────────────────────────────┘
|
▼
┌──────────────────────────┐
│ 4. Applications │
│ (Quoting / Chatbot) │
└──────────────────────────┘
See how it Combines AI intelligence with On-prem Data
What does a concrete application look like?
USER APPLICATION STOCK CHECK (ERP) AI VECTOR SEARCH
┌─────────────────────────┐ ┌───────────────────────────┐ ┌───────────────────────────┐
│ Search SKU: A-100 │ ─────>│ Stock query: │ ─────>│ IF stock == 0 THEN │
└─────────────────────────┘ │ "A-100" = 0 pcs │ │ Search neighbors: │
└───────────────────────────┘ │ "Close matches found: │
│ B-200, B-30, B-13" │
└─────────────┬─────────────┘
│
▼
┌───────────────────────────┐ ┌───────────────────────────┐
│ FINAL ENRICHMENT: │ │ DETERMINISTIC RULE: │
│ "A-100 unavailable │ <──── │ Only B-30 meets spec. │
│ Replacement: B-30 │ └───────────────────────────┘
│ (80 pcs)" │
└───────────────────────────┘
See Ten Deliverable Projects
About this guide
This guide is designed to be read selectively, depending on your role in the company. The table of contents for each chapter indicates our recommendation per section: Business Sales Tech Product. A technical reader can skip directly to Chapters 2 (How?) and 4 (Cookbook), while a sales director can start with Chapter 1 (Why?) and then skip to Chapter 3 (What?). The financial director can go directly to Chapter 5 (When?).
OPTI Software is a Google Cloud and HubSpot partner. This guide represents the opinion and methodology of OPTI Software and has not been supported or approved by Google.
The guide is part of the AI in B2B 2026 series, created by OPTI Software. First published date: 21st of January 2026. All rights reserved.



