Getting Started Guide

Step 1 of 9

First question

Where are you with AI right now?

AI Getting Started Checklist Report

Bring this to your next management meeting. Work through it in plain English before starting a formal AI technology project.

Progress

Start using AI in your work, today.

  • Get AI workspaces on people's computers, including yours. ChatGPT, Claude, Gemini for basic chat. Next, download an AI workspace app like Codex, Claude Cowork or Cursor to start doing real work.
  • Ask questions, draft, summarize, compare options, review documents, explore ideas. Use AI to draft emails, review documents, explore ideas, compare options, and more. Don't know what do to next? Ask it.
  • Try it on Big Robot. Open an AI workspace and ask it about Big Robot. Our website is AI-readable, so you can have a real conversation with it.

People

Who is ready to move?

  • Find the people already trying. Somebody is probably using AI quietly. Learn what they are doing.
  • Find the people who need permission. Some people will not touch AI until leadership says it is allowed.
  • Find the people who need help. Training matters. A good AI rollout is a behavior change, not just a software install.

Visibility

What gets measured gets managed.

  • Build a shared dashboard for active management. It might be separate tools, an app, or a spreadsheet. But it should be a single view of the business with metrics you need to manage. Bonus points if it can be updated automatically.
  • List the places work shows up. Systems, spreadsheets, inboxes, documents, dashboards, chat, tickets. This the raw materials for AI to work with.
  • Notice what is still invisible. Phone calls, hallway decisions, personal notebooks, memory, side spreadsheets. This is the hidden work that AI will not see. If you spend a lot of time trying to track stuff down, this one is key.

Work

Discover opportunities in the work done today.

  • Look for slow work. Repetitive review, status chasing, document handling, retyping, reconciliation.
  • Look for knowledge-heavy work. Decisions that depend on knowing the customer, the job, the contract, or the history.
  • Pick work people already complain about. The best starting point is usually obvious to the people doing it every week.

Rules

What needs to stay under control?

  • Protect private information. Customer records, employee data, financial details, credentials, contracts.
  • Keep important actions reviewed. Anything that sends messages, changes records, approves work, or moves money needs a human gate.
  • Write down what "right" means. AI works better when the business can explain the rules it already uses.

Do Not Include

  • Do not include credentials, tokens, verification codes, private record values, signed URLs, attachments, or regulated data.
  • Describe systems and records by category unless a private approved workspace is being used.
  • Preserve uncertainty instead of inventing missing facts.
  • Mark omitted details clearly when they are unsafe to share publicly.

Executive Snapshot

A short summary of the workflow, owner, timing, and recommended next step.

  • Workflow Name the workflow to assess.
  • Business owner Name the role or team that owns it.
  • Why now Explain why this matters now.
  • Current AI stage Curious, experimenting, scattered pilots, already building, or not sure.
  • First concern Where AI fits, data readiness, safety, team motion, or not sure.
  • Recommended next step Getting Started Guide, Big Robot discovery call, or more mapping.

Workflow Candidate

The work that may be worth improving with AI and trusted data.

  • Work to assess Describe the work in plain English.
  • Who does it List the roles or teams involved.
  • Why it matters Explain the business value.
  • Current friction Describe what is slow, unclear, repeated, or risky.
  • Desired outcome Describe what should improve.

Current State

What works now and what creates friction.

  • What already works List useful current tools or habits.
  • What is slow List slow steps.
  • What is repeated List repeated work.
  • What depends on judgment List decisions requiring context.
  • What people complain about Capture common complaints.

Systems And Records

Systems, records, and source-of-truth boundaries.

  • Systems involved List systems by category or name when safe.
  • Records involved List record categories.
  • Source of truth State which system owns which fact if known.
  • Conflicts Note duplicate or conflicting records.
  • Do not share publicly List categories of details intentionally omitted.

Visibility Gaps

What the business can and cannot observe in software.

  • Visible in software List visible systems, reports, or dashboards.
  • Hidden work List work hidden in email, chat, calls, notes, or memory.
  • Reports used today List current management views.
  • Missing view Describe the view leaders need.

People And Permissions

Who is ready, blocked, responsible, or sensitive-access.

  • People already trying AI List roles or teams.
  • People who need permission List roles or teams.
  • People who need training List roles or teams.
  • Approval owners List approval roles.
  • Sensitive-access roles List roles with sensitive access.

Rules, Risks, And Human Gates

Information, actions, approvals, and audit boundaries.

  • Private information List sensitive categories only.
  • Important actions List actions needing control.
  • Required approvals List approval points.
  • Messages or record changes List changes needing review.
  • Money movement or compliance risk Describe risk categories.
  • Audit needs Describe what must be traceable.

AI Opportunities

Low-risk starts and controlled workflow assistance.

  • Low-risk starter uses List safe first uses.
  • Workflow-assist opportunities List assisted work.
  • Data or system gaps List gaps to solve first.
  • Human-reviewed work List work that should stay reviewed.

Readiness Signals

Whether this is clear enough for a discovery conversation.

  • Clear enough to discuss Yes, no, or partial.
  • Needs more mapping List what is missing.
  • Not ready yet State why if true.
  • Reason Explain the readiness judgment.

Open Questions

Known unknowns to resolve before implementation.

  • Question 1 Add an unresolved question.
  • Question 2 Add an unresolved question.
  • Question 3 Add an unresolved question.

Safe-To-Share Summary

A public-safe summary suitable for a contact form or call agenda.

  • Workflow summary Summarize without private data.
  • Systems summary Summarize system categories.
  • Risk summary Summarize risk categories.
  • Intentionally omitted List omitted sensitive categories.

Big Robot Discovery Agenda

Suggested first-call structure.

  • Confirm workflow and owner Confirm the workflow and business owner.
  • Review systems and source-of-truth boundaries Review systems and authority.
  • Identify records, rules, and work surfaces Map records, rules, and work.
  • Decide what needs human review Identify approval gates.
  • Define smallest useful next step Name the first practical implementation step.