Environment Setup & Prompt Builder Best Practices

Completing these steps before the hands-on session ensures everything runs smoothly.

This opens the Prompt Library — the main page where all prompts for your organization are listed.

​What the Prompt Library Shows

The Prompt Library page displays:

If Something Isn't Working

If the Admin Settings option or the Studio navigation item isn't visible, admin permissions may not be enabled on the account. Reach out to a Databook admin or the bootcamp facilitator to get access resolved before the hands-on session.

The Prompt Builder is where conversation starters are created — pre-built prompts that appear as clickable cards for end users. Rather than typing out a question from scratch, a user clicks a card and gets an AI-generated response instantly.

It's a way to package high-quality questions so that an entire team can benefit from them without needing to know exactly what to ask.

Dynamic parameters like {account} are inserted directly from the Prompt Builder UI. They make a prompt reusable — write it once, and it works for every company in the territory.

When an end user clicks a prompt card that contains a dynamic parameter, they'll be asked to select an account. The AI then runs the prompt using that specific company's data. There's no need to create separate prompts for every account.

Categories group prompts by use case or sales stage. Some examples:

Any category structure that fits the team's workflow can be created here. The goal is to make it easy for users to find the right prompt at the right moment.

The right side of the Prompt Builder page has a testing panel. To test:

Testing is the best way to iterate on a prompt and refine the output before making it available to the team.

A prompt doesn't need to be finished in one sitting. Save it as a draft and come back to it. Once the output is consistent and useful, activate it — the prompt will then appear in the Prompt Library for the entire organization.

The more specific the prompt, the more useful the response. Vague questions produce vague answers.

A prompt performs best when it reads like instructions to a research analyst — what to do, what sources to prioritize, and how to structure the output.

Example:

Analyze {account}'s most recent earnings call transcript. Identify the top 3 strategic priorities mentioned by the CEO or CFO. For each priority, explain why it matters for a sales conversation and suggest one question I could ask a VP-level contact about it.

If the desired output is a list, a table, or bullet points with headers — say so. The AI follows formatting instructions.

Example:

For {account}, create a competitive landscape summary in the following format:

Limit to the top 5 competitors.

Telling the AI what NOT to include is just as important as telling it what to include.

Example:

Summarize the latest quarterly earnings for {account}. Focus only on revenue, growth rate, and any forward guidance changes. Do not include stock price commentary or analyst ratings. Keep the summary under 200 words.

The 10,000-character limit is there for when precision demands it — not to be filled for the sake of length. A focused 200-character prompt will outperform a scattered 5,000-character one.

The key is density of intent: every sentence in a prompt should either clarify what is wanted, how it should be formatted, or what to avoid. If a sentence doesn't do one of those three things, it's probably not adding value.

Dropping {account} into a generic sentence produces a generic response. Building the entire prompt around the parameter produces a response that's deeply contextualized to that company.

Weak usage:

Tell me about {account}.

Strong usage:

Based on {account}'s 10-K filing, earnings call transcripts, and recent news from the past 6 months, identify 2-3 signals that suggest they may be evaluating new solutions in their go-to-market function. For each signal, explain the business pressure behind it and recommend a conversation angle for outreach.

The prompts created in the Prompt Builder are for the team. The output should be something a rep, AE, or CSM can use immediately — without heavy editing. If the response consistently requires rework, the prompt needs refinement.

These are ready to copy into the Prompt Builder. Test them with a few different accounts and adjust as needed.

Summarize the most recent earnings call for {account}. Include: (1) the top 3 strategic priorities mentioned by leadership, (2) any guidance changes or revised targets, (3) key themes from the analyst Q&A. Keep the summary under 300 words and use bullet points.

Identify, validate, and rank 5 to 10 peer companies for {account} using official filings from the last two fiscal years. For each peer, include: company name, industry, approximate revenue, and one sentence on how they compete with or relate to {account}.

Create a research brief for {account} that can be reviewed before a first meeting. Include: (1) company overview — what they do, size, and industry position, (2) recent news from the last 6 months, (3) any leadership changes in their go-to-market or technology functions, (4) 2-3 potential pain points based on public information. Format as a one-page brief with headers.

Based on all available data for {account}, identify the 2 strongest buying signals from the following categories: financial performance pressure, revenue growth deceleration, leadership changes, hiring surges, strategic or operational shifts, or earnings Q&A and external pressure. For each signal, provide 2-3 sentences of evidence and explain why it creates urgency.