Templates
Free checklists and worksheets you can use right away. Each one is printed in full below, so you can copy it into a doc or print the page. They’re useful on their own, and they’ll also make a Free AI evaluation more productive if you book one.
Everything to decide before you test AI on a real task: one workflow, a baseline, and the point where you’d stop.
Name the workflow
- The workflow’s name, in the words your team already uses (quoting, intake, invoice follow-up).
- Who handles it today, and which system it lives in.
- What “done” looks like when a person finishes it without AI.
Pain and risk
- How often it causes trouble, and who feels it.
- What happens if the AI gets it wrong: internal only, customer-facing, money, or safety.
- Stick to this one workflow, even if a second one looks easy to add.
Baseline
- How long a typical instance takes, written down before the pilot starts.
- What mistakes or rework look like today, with real examples.
- Where the finished work is stored now.
Window, success, and stop
- Pilot window: one to two weeks, running alongside the current process.
- Success you can observe, like drafts that are usable and a review step that holds up.
- Stop criteria: what would make you switch it off, agreed in advance.
- The one person who owns the output.
- Data notes: what information can go into the tool, and what can’t.
Book a Free AI evaluation
A month to look around, pick one workflow, test it, and decide. It’s one decision, not a company-wide overhaul.
Week 1: Take stock
- List the tools you already pay for and who uses them.
- Name the three workflows that take up the most time each week.
- Write down the one person responsible for AI today. A team or “everyone” doesn’t count.
Week 2: Pick and baseline
- Choose one workflow with high pain and low damage if it goes wrong.
- Record how it goes today using real examples, not a guessed average.
- Tell the people who do the work what you’ll be comparing.
Weeks 3 and 4: Shadow pilot
- Run the new approach alongside the old one.
- Have a person review everything before it reaches a customer.
- Note where people work around the pilot, since that tells you what to fix.
Decision worksheet
- Keep: the review step holds up and the owner can explain the output.
- Revise: the task is right, but the prompt, tool, or handoff needs work.
- Stop: the risk showed up, or nobody will own it next month.
Book discovery
A one-page map of your tools, your AI agents, and the person who makes the final call on each workflow.
Tools you already use
- Email, calendar, CRM or job board, accounting, files.
- What each one is the official record for.
- Which ones people still open on a busy day.
Agents and assistants
- Which agents you run (a Grok Bot per workflow, for example) and which chat assistants people use alongside them.
- Where an agent may read, and where it’s allowed to write.
- Shared memory: what the next session or bot is allowed to know.
Who owns what
- Who approves a draft before it goes out.
- Who can switch the workflow off.
- What stays a person’s call even when the draft is good.
See LaunchpadOS plans
What to bring to a Free AI evaluation so the call covers your actual week instead of a feature tour.
Bring these
- The three workflows that take up the most of your time.
- The tools you use now, and which one is the official record.
- Who’s responsible for AI today, even if the answer is nobody yet.
- Constraints like customer data rules, busy seasons, headcount, or locations.
- The questions you want answered, including “What shouldn’t we automate?”
You won’t need
- A target percentage. Nobody can honestly promise one in advance.
- A list of every license you own. The ones people use are enough.
- A decision to buy anything.
Book a Free AI evaluation
