Saizen GPT Suite

Overview

I architected a suite of specialised custom GPTs for Saizen Acuity, designed to turn repeatable business tasks into structured deliverables. Rather than relying on one general-purpose chatbot, each engine has a defined scope, required inputs, processing procedure and predictable output format.

What It Delivers

• 14 configured custom GPT engines and 30 defined workflow profiles.
• Structured outputs suitable for direct use in documents, spreadsheets and CRM processes.
• A ChatGPT-native implementation requiring no separately hosted servers or API orchestration layer.
• A consistent framework for reviewing and transferring work between specialised assistants.

What I Did

• Designed 14 specialist engines spanning research, CRM, sales, proposals, email, competitive analysis, advertising, content and operations.
• Defined a seven-part instruction architecture: role, intake, procedure, output contract, guardrails, self-check and handoff.
• Created fixed schemas for deliverables, including required fields, tables, assumptions, risks and acceptance criteria.
• Designed 30 mission profiles that combine specialist engines into multi-step business workflows.
• Standardised compact key-value handoff blocks so outputs could be transferred between engines without reformatting entire conversations.
• Added input-completeness checks, scope routing, human-review gates and claim-verification rules.
• Versioned instruction sets and managed context budgets to improve repeatability across longer workflows.

Outcomes

• Prompt instructions treated as repeatable operational specifications, with defined outputs and review controls.

Proven Impact

• 14 specialised engines, 30 mission profiles and one output contract per engine.

Why It Matters

• The system treats prompt instructions as repeatable operational specifications, with defined outputs and review controls rather than unstructured conversational responses.