Section 4B: Skills, Custom Workflows & Python Analysis

Content Type: How-To

4.3 Use built-in skills

Skills are pre-built, multi-step workflows that run inside the Lightfield agent. Trigger them with a prompt and the agent executes the full workflow — researching, drafting, updating, or reporting as needed.

How to run a skill

"Run the cold opportunity reactivation skill""Run the active opportunities report""Run the account research skill for Thoughtium"

You can also browse and launch skills from the Skills section in the left sidebar.

Built-in skills

Cold opportunity reactivation — Scans all Churned, Closed Lost, Canceled, and Later-stage deals. For each, checks whether anything has changed (new funding, filled feature gap, contact changed companies). Drafts personalized re-engagement emails for the best candidates.

Active opportunities report — Generates a visual pipeline breakdown by stage, owner, and deal value. Good for pipeline reviews and weekly syncs.

Account research — Researches a company before a call. Pulls public info, recent news, funding history, and CRM context into a prep briefing.

Daily briefing — Summarizes today's calendar, yesterday's deal activity, and open tasks. Run it first thing in the morning.

Draft quote / order form — Generates a formatted pricing proposal for a specific opportunity, pulled from deal details and your pricing knowledge document.

Reschedule no-show — Drafts a friendly rescheduling email when a prospect misses a meeting.

Handoff contact to sales — Drafts a warm intro/follow-up from a colleague to a contact, using context from the most recent meeting.

4.4 Build a custom skill

If the built-in skills don't cover your use case, build your own. Skills are markdown files stored in your workspace and triggered by name in the agent.

Create via the agent

"Create a skill called 'Monthly churn review' that looks at all deals moved to Churned in the past 30 days, extracts the primary reason, and writes a summary report."

The agent creates the skill file. Trigger it any time: "Run the monthly churn review skill."

Create via Settings

Skills (left sidebar) → + New skill → name it, write the prompt → Save.

What makes a good skill prompt

• Be specific about inputs and outputs — what data it should look at, and what it should produce

• Reference knowledge documents: "Use the pricing document to calculate the proposal amount"

• Skills can chain actions: research → draft → update CRM → send email

4.5 Run Python analysis on your CRM data

The agent can write and execute Python scripts against your CRM data — useful for custom analysis, bulk operations, or charts not available in the standard UI.

How it works

Open the agent → describe the analysis → the agent writes the Python code, runs it, and returns the results. No coding experience needed.

Example prompts

"Analyze my pipeline: show average days in each stage, and flag any deal stuck in the same stage for more than 30 days."

"Analyze all churned deals from the past 6 months. What are the most common close reasons? Show a bar chart."

"Compare win rate and average deal size for each member of my sales team."

Starting with an export

For larger analyses, export first: "Export all my opportunities to a CSV." Then reference that file in your analysis prompt.

What you can do with results

• View tables and charts directly in the chat

• Save results as a file in your workspace

• Act on the insights: "Based on this analysis, update the tier field for all Gold-tier accounts"