Skip to main content

Setting Up Your Brand in Cognizo

Written by Stevi
Updated over a month ago

Brand setup is where Cognizo learns how your business shows up across AI-powered search. The goal is not perfection on day one. It is to get a strong baseline you can refine as data starts coming in.

Create a New Brand

Start by clicking Create New Brand.

You will be asked to name your brand. Use the version of your name that appears most consistently online. Avoid legal suffixes like “Inc.” or “Corporation” unless they are used everywhere. Shorter, commonly used brand names are easier for AI systems to recognize and cite.

Once submitted, Cognizo will begin building a profile using your website, public sources, and user-generated content.

This step usually takes about a minute.

Review Your Brand Overview

After the profile loads, you will see a brand overview that includes your industry categorization, high-level focus areas, and early assumptions about how your business is described online.

Review this carefully. This is an early discovery moment. What we generate here can surface real content gaps and opportunities. If the overview leans too heavily into one offering, that is often a signal that your broader story may be underrepresented in existing content online.

Select and Add Competitors

Next, you will see a carousel list of competitors Cognizo identified automatically. Review them and check the ones that are relevant. If you do not recognize a name, skip it. You can always come back.

Add any that are missing by typing them into the field below the carousel.

We recommend starting with at least five competitors to ensure meaningful early data.

More is better over time and many teams end up with 15-20 or more!

You can move forward even if this list is not perfect. Competitors can be reviewed and added at anytime after brand setup is complete.

Step 5: Define Topics

Topics are organizational buckets for your prompts. They exist to keep your data readable as your setup grows and to make it easier to understand visibility patterns at a higher level. A good topic represents a meaningful slice of how customers discover, evaluate, or engage with your business.

There is no single correct way to define topics. The most useful setups tend to mirror how the business actually operates rather than how it is marketed.

For some teams, that means creating topics around core product lines so each offering can be tracked independently. Others organize topics around major services or use cases, especially when different parts of the business solve very different problems.

Topics can also reflect how customers move through your funnel. You might group prompts that capture early research questions in one topic and more purchase-oriented questions in another. This approach can surface where visibility drops off as intent becomes more commercial. In seasonal businesses, topics sometimes align to business cycles, making it easier to compare performance during peak and off-peak periods.

Some teams use topics to separate audiences or channels. For example, prompts related to enterprise buyers may live in a different topic than prompts aimed at small businesses or partners. Over time, this makes it easier to see how visibility shifts across customer cohorts or stages in the lifecycle.

In short

  • Think in terms of how your business is structured, not how you market it.

  • Topics are flexible. You can add, remove, or rename them later.

  • Prompts will still run correctly even if a topic is not ideal, but mismatched topics can skew roll-up reporting.

If the setup feels close enough, continue.

Review Cognizo’s Suggested Prompts



Cognizo will generate 5–20 prompts per topic using industry research, user-generated content, and real-world questions from forums and social platforms.


These prompts represent how people actually ask about your space. When reviewing prompts, remember to look for the intent of the prompt, not the specific wording. AI breaks everything down to intent so it’s ok if the prompt doesn’t read the exact way a person might ask something.

You know your business better than we ever will, so be sure to remove any prompts that are clearly irrelevant to your business.

Expect some prompts to feel technical or unfamiliar, especially in specialized industries. You can remove these or track them for a short period of time to analyze if they are valuable to you.

Add Custom Prompts

If you collect prospect or customer questions as part of your business model, this is a great time to review them and add real user-questions you’ve collected to your prompt tracking. You probably have some gold-standard prompt material in your contact forms, support tickets, and sales calls.

Also consider adding prompts that cover differentiators your competitors do not offer, end-to-end services, geographic reach or targeting locations of interest, legacy or credibility signals

You can always edit or deactivate prompts later.

Complete Setup and Let It Run

Once you finish, Cognizo will finalize your brand setup. This typically takes about an hour. You will receive an email when everything is ready.

From here prompts will run daily across your enabled AI platforms and data starts accumulating immediately. The first 7–10 days are about establishing a baseline but feel free to explore as the data is rolling in. Take notes on questions or patterns you notice.

Plan your first internal review once initial data is available.

What Happens Next

Most teams plan an internal review of early results after one or two weeks of data have collected. In this review look at where you are showing up, where competitors are cited and your brand is not, and which prompts reveal the biggest gaps or opportunities

From there, you can tighten prompts, expand coverage, and start connecting visibility data to content and optimization work. Check out the linked guides below to help you analyze your data and continue to refine your brand buildout in Cognizo.

If you have questions reach out to our support team via live chat or email us at [email protected]

Did this answer your question?