Skip to content

Solution Library

Everything AI at SMG today — production CoE-built solutions, plus self-serve patterns associates have built themselves and shared. New entries land every fortnight.

CoE-built solutions

Production-grade, governed, owned by the AI CoE with embedded AI Solution Leads.

Solution Function Status
SLCP — Scotts Lawn Care Program Consumer / Digital · R&D Live
MyLawn Brain Consumer / Digital In Build
Bloom — SMG Agent Fleet platform Cross-functional · every department Pilot
PDP Compliance Checker — co-built with Content + Legal Marketing · Content · Legal Pilot

Gemini Tools & Ecosystem Guide

Core Gemini engines, Workspace integrations, and specialized productivity features available across Scotts — with key settings and departmental use cases.

Engine Modes: Flash, Thinking, & Pro

Before starting a task, you can adjust the reasoning depth and context capacity of the underlying model:

Mode Core Function Best Used For
Fast / Flash Low-latency engine optimized for rapid responses without deep logical verification. High-volume routine queries, quick email drafts, simple formatting, and rapid Q&A.
Thinking Executes a step-by-step reasoning loop to verify logic before generating output. Complex math, auditing financial formulas, multi-step business logic, and technical troubleshooting.
Pro Employs massive context windows (1M+ tokens) to analyze large volumes of data simultaneously. Processing 100+ page PDFs, analyzing company-wide strategy documents, or evaluating large datasets.

Tools

  • Gemini App: The primary conversational assistant for Q&A, brainstorming, writing, summarizing, and analyzing uploaded files (PDFs, spreadsheets, text documents).
  • Gemini Canvas: An interactive split-screen workspace designed to draft content side-by-side with AI, build interactive quizzes, and edit text inline.
  • Deep Research: An autonomous research assistant that executes multi-step web queries, synthesizes findings across dozens of sources, and delivers comprehensive, cited reports.
  • Gems: Pre-configured, customized versions of Gemini tailored with specific system instructions to automate repeatable organizational workflows.
  • NotebookLM: A dedicated research notebook anchored strictly to uploaded sources (PDFs, Docs, Slides) for zero-hallucination analysis, inline quote citations, and audio podcast overviews.

Google Workspace Integration

Access Gemini directly inside your daily Workspace applications without switching tabs:

  • Gmail & Google Docs: Context-aware drafting, tone adjustment, summaries, and automated proofreading.
  • Google Sheets: Features "Fill with Gemini" to automatically analyze, classify, and categorize unstructured text data, build formulas, and format charts without manual coding.
  • Google Vids: An AI-powered video creation and editing assistant for producing quick video briefs, onboarding materials, and internal updates.
  • Google Workspace Studio: A no-code AI automation tool to build multi-step workflows across Workspace apps (Gmail, Drive, Sheets, Chat, Docs) using natural language prompts.

Use Case Matrix

Department / Role Operational Scenario Recommended Tools Key Value
Operations & Supply Chain Automating inventory alert workflows, routing urgent supplier emails, or solving shipping logic. Workspace Studio + Thinking Mode Automates multi-step cross-app triggers without code while verifying complex logic.
R&D & Compliance Reviewing dense regulatory frameworks, safety datasheets, or agronomic research papers. NotebookLM or Deep Research Ensures answers are grounded in source docs with inline citations.
Marketing & Brand Drafting campaign collateral, messaging frameworks, and internal video storyboards. Gemini Canvas + Google Vids Collaborative split-screen writing paired with AI-assisted video production.
Sales & Business Dev Categorizing thousands of customer feedback entries or updating CRM records in Sheets. Google Sheets ("Fill with Gemini") Automatically tags and categorizes unstructured text inputs without writing formulas.

Data Governance Note: Workspace data and enterprise prompts processed through Scotts enterprise Gemini deployment remain confidential and are protected from public model training.

Self-serve patterns

Associate-built solutions using Gemini, NotebookLM, or Cowork — replicable today by anyone in the function. The full submission lives on the Ideas page; reach out to the submitter through your function's AI Solution Lead.

Solution Function Built with Status
Production Planning Tool — Andrew Pitts Supply Chain Gemini API + Workspace Self-serve
Tri-Engine AI Operations Framework — Terri Weaver Marketing Gemini Workspace + API + NotebookLM Self-serve
Creative Brief writing — Caitlin Dezso Marketing Gemini API Self-serve
Streamlining Product Knowledge (GEMs) — Tish Sherrick Research & Development Gemini GEMs + NotebookLM Self-serve

Status legend

Live — In production. Owned, governed, measured. Pilot — Running with a small user group; success criteria defined. In Build — Scoped, resourced, in development with an AI Solution Lead. Triage — In intake; CoE evaluating value, feasibility, foundation-readiness. Self-serve — Doable today with Cowork or Gemini; no CoE engineering required. Parked — Not now. Reason noted (e.g., foundation not ready, value insufficient, regulatory blocker).


What an entry looks like

Every Solution Library entry uses this template so they're comparable, scannable, and linkable. Sections include:

  • Problem — the business pain, in one paragraph
  • Function & owner — who it's for, who owns it
  • Where it runs — Bloom / Cowork / Gemini / standalone
  • Data it touches — and the classification level
  • Status — Live / Pilot / In Build / Triage / Self-serve / Parked
  • Demo — link, screenshot, or video
  • Source — GitLab link if applicable
  • Governance — review status and any HITL requirements
  • Patterns reused — links to predecessor entries (this is what makes the library a graph, not a list)
  • What's next — known roadmap

Want to add an entry?

If you've built something with Cowork or Gemini that solved a real problem — even a small one — the CoE wants to know. Submit it as an idea with "[for library]" in the title, or talk to your function's champion (once the network is live in FY27).