Getting the most out of Dobbin means learning to treat it not as a generic tool, but as a thinking partner that is part of your team. The quality of your prompts directly impacts the quality of its strategic guidance.

Here are five tips to help you move from simple questions to powerful, valuable outcomes:

1. Give a Direct Instruction

Dobbin gives its best feedback when you give it a clear job to do. Instead of asking for a general opinion, which can lead to unhelpful praise, give it a direct instruction that leverages its knowledge of your brand.

Instead of asking for an Opinion

What do you think of this draft?

Give Dobbin a job to do

Review this draft against our brand voice guidelines and suggest three ways to make the tone more confident and direct.

2. Use Dobbin as a True Thought Partner

This is the most important mindset shift. Generic AI is for simple, transactional tasks. Dobbin is a system designed for strategic partnership. Because it's grounded in your company’s strategy, you can ask it to take on a role (like a marketing strategist) and it will reason from the perspective of your business. Don't just ask it to complete a task; ask it to have a perspective.

Don't just give it a task

Write a blog post about productivity.

Give it a role

Act as our lead content strategist. I need to write a blog post on productivity for an audience of busy founders. Outline a post that connects our product's core value of 'achieving velocity' to their daily struggles, and include a section that refutes a common productivity myth.

3. Specificity is Speed

Vague prompts require more processing time. Specificity makes Dobbin uniquely fast because it can immediately pinpoint the right information in your Knowledge Base. You don't need to know the exact document name; just describe the topic or type of document you're looking for, and Dobbin will find the most relevant source.

Instead of a broad question

What have we said about our pricing?

Make a pinpointed request

Summarize the key points from our 'Q3 Pricing Strategy' document. I need the main talking points for handling objections about our enterprise tier.

4. Context is King

The most common mistake is being too vague. While context helps any AI, for Dobbin it’s essential. Providing context allows it to access its entire memory of your company's strategy and voice, leading to a much richer, more aligned response.

No context

Make this sound better.

With context

I'm drafting an email to a customer who is at risk of churning because they've complained the product is too complicated. Here's my draft: [Paste draft here]. Rewrite it to be more empathetic, acknowledge their frustration, and clearly highlight the one feature that will solve their immediate problem.

5. Iterate in a Conversation

Don't try to write the perfect prompt. The best work happens in the back-and-forth. If an output isn't right, just reply in the same thread to refine it. Refining a prompt with Dobbin is like workshopping an idea with a partner who never forgets your core strategy, making each iteration more powerful because it's always grounded in your brand.

Instead of starting over

(Starting a new thread for every attempt at the same question.)

Refine the request

That's a good start, but the first version was too technical. Let's try again. This time, write it from the perspective of a user who has just saved an hour of work. Focus on the feeling of relief, not the feature itself.

What Not to Worry About

Getting great results doesn't require perfection. Here are a few things you can safely ignore when writing prompts for Dobbin:

Spelling and Grammar

Don't worry about typos or perfect punctuation in your prompts. Dobbin is designed to understand natural language, mistakes and all.

Perfect Prompt Structure

You don't need to be a professional prompt engineer. Just focus on providing clear context and instructions, and you'll get great results.

Having a Perfect Idea

It’s okay to be messy. You can upload a rambling voice memo or paste in rough notes and ask Dobbin to structure them into a coherent first draft.