New Client Brand - Setup Playbook

Start by confirming that the topics you set up during onboarding actually map to the parts of the business your client cares about.

There's more than one right way to structure this. A consumer brand selling multiple product lines usually wants topics organized by product line, since that's how they think about performance internally. A B2B SaaS brand might care more about funnel stage, in which case you'd set topics like "product line, top of funnel" and "product line, mid funnel" with prompts layered underneath.

The test is simple: do these topics match how the business is organized and how you plan to report results back to them? If not, fix it now before you have a bunch of data sitting under the wrong structure.

Learn more about Topic Strategy & Best Practices

Next, look at the prompt list and get a feel for what you're tracking.

Be sure keep watch for volume. If your highest volume prompt is sitting under 10, you're tracking a very narrow slice of user queries. You want a mix of prompt volumes, some niche and specific prompts that reflect what they sell, plus broader prompts that catch people earlier in their awareness journey. For a premium tea brand, someone shopping for "direct trade organic green tea brands" already knows quite a bit. Someone searching for "a relaxing drink before bed" is the person the client probably wants to reach as well.

Learn more about Prompt Strategy & Best Practices

If you need more prompts, suggestions are a great starting point. Pick your region, pick a topic or select all of them, and let the platform generate ideas.
​

Learn more about Prompt Volume & Suggestions
​

Three ways to dig deeper here, and they're all useful for different reasons.

Keyword research. Search a broad term related to the client's category and see what surfaces. This is usually where the higher volume, top of funnel prompts come from. When you find one you like, you can add it to an existing topic or create a brand new topic right there if the prompt doesn't fit anywhere.

Domain research. Scan a domain and pull prompts associated with it. Handy for finding prompts tied to a competitor's space or to your client's own content.

Competitor Gap Analysis. Select your tracked brand, select a competitor, and explore. The platform shows you where the competitor has coverage that your client doesn't. These gaps can be very interesting prompts to add, especially when they come with good volume, because they show you where your client is losing ground in areas they care about.

Once you've added a batch, they'll run on the next cycle and you'll start getting a read on visibility.

Learn more about Prompt Research

This is the technical side. The platform crawls the client’s domain and checks that everything AI bots need in order to find and crawl the content efficiently is in place.

The foundation analysis looks at three files:

The first two are usually there. llms.txt is the one that's most often missing. Right now, this is a nice to have rather than a need to have, and if it's absent or out of date the platform can generate it for you.

If a client asks what llms.txt even is, here's the short version: it's a text file added to the root of the domain that gives a curated map and summary of the site's content, written specifically for agents and bots rather than people. In super simple terms: it helps the robots find, read, and understand your content.

The Site Audit reviews all pages within the client's domain to find suggestions to improve crawlability.

Review suggestions and export them to review with the client.

Walk the client through the integration options. Cognizo has out-of-the-box integrations for Cloudflare Workers and WordPress, plus a custom integration route.
​

If the client isn't on Cloudflare or WordPress, Cognizo integrates with plenty of other providers, it just means the setup takes a bit more exploration on your end. Reach out to your CSM or [email protected] for any customer integration questions.

Learn more about Traffic Analytics

Be sure to review the information we captured when scanning the client's domain. The visual identity including brand colors, brand voice & tone, and audiences can all be tailored to match the brand's existing online voice. Set this once and it will apply to all content generated via Cognizo.

​

Once the setup work is done, head into Content Studio. Based on the data collected for the brand so far, you'll get a set of recommendations for what to publish to move visibility.

The recommendation types vary and can include refresh recommendations on existing pages, blog content like articles, listicles, and comparisons, social content like reddit and Youtube, or backlink opportunities with earned media. You can regenerate the list up to three times a day, and regenerating often surfaces different formats than the first pass, so it's worth doing if nothing in the initial set feels right for the client.

Click into any single recommendation and you'll see what to include in the piece, the reasoning behind the recommendation (usually something like "this competitor is winning on all three of these prompts and the strongest citation signal is article pages at 66% of citable format"), and the sources AI is citing today.

If a recommendation is good but not urgent, save it and it moves to your saved section for later. When you're ready, hit create and Content Studio generates a brief you can either turn into a full article or share directly with the client.

The ongoing rhythm looks like this: come in regularly (at least once per week), look at the new recommendations, and make sure the client is hitting the maximum amount of content they can create each month.