# Brand Kit Foundations

### About this export

| Field | Value |
| --- | --- |
| **content_type** | course |
| **platform** | contentstack-academy |
| **source_url** | https://www.contentstack.com/academy/courses/brand-kit-foundations |
| **language** | en |
| **product_area** | ai |
| **learning_path** | standalone |
| **course_id** | brand-kit-foundations |
| **slug** | brand-kit-foundations |
| **version** | 2026-03-01 |
| **last_updated** | 2026-05-04 |
| **status** | published |
| **keywords** | ["ai","brand kit","automate","brand voice","voice profiles"] |
| **summary_one_line** | Make Generative AI More Powerful Contentstack's Brand Kit allows you to maintain a consistent brand voice across all channels. Effortlessly generate content that aligns with your brand’s personality and resonates with yo… |
| **total_duration_minutes** | 13 |
| **lessons_count** | 3 |
| **video_lessons_count** | 3 |
| **text_lessons_count** | 0 |
| **linked_learning_path** | standalone |
| **linked_assessment_ref** | LMS_UNCONFIGURED_COURSE_ASSESSMENT |
| **markdown_file_url** | /academy/md/courses/brand-kit-foundations.md |
| **generated_at** | 2026-05-04T05:36:44.791Z |
| **intended_audience** | [] |
| **prerequisites** | [] |
| **related_courses** | [] |

> **Academy MD v3** — companion `.md` for Ask AI. Quizzes and graded assessments are **LMS-only**; this file never contains answer keys.

## Course Overview

| Metadata | Value |
| --- | --- |
| Catalog duration | 12m 45s |
| Released (if known) | 2026-03-01 |
| Product area | ai |

### Description

### Make Generative AI More Powerful

  
Contentstack's Brand Kit allows you to maintain a consistent brand voice across all channels. Effortlessly generate content that aligns with your brand’s personality and resonates with your audience. Brand Kit includes Knowledge Vault - a centralized way for you to store your company’s essential information. This will empower AI to produce factually accurate content that supports your business objectives.

### Learning objectives

1. Follow each lesson in order.
2. Practice in a training stack using placeholders **YOUR_STACK_API_KEY** and **YOUR_DELIVERY_TOKEN** in local `.env` files only.
3. Validate API responses against the official documentation.

### Topics covered

ai; brand kit; automate; brand voice; voice profiles

## Course structure

```text
brand-kit-foundations/
├── 01-introducing-brand-kit · video · 274s
├── 02-voice-profiles · video · 237s
├── 03-knowledge-vault · video · 254s
```

## Lessons

### Lesson 01 — Introducing Brand Kit

<!-- ai_metadata: {"lesson_id":"01","type":"video","duration_seconds":274,"video_url":"https://cdn.jwplayer.com/previews/WPNnZ8cY","thumbnail_url":"https://cdn.jwplayer.com/v2/media/WPNnZ8cY/poster.jpg?width=720","topics":["Introducing","Brand","Kit"]} -->

#### Video details

#### At a glance

- **Title:** Introducing Brandkit
- **Duration:** 4m 34s
- **Media link:** https://cdn.jwplayer.com/previews/WPNnZ8cY
- **Publish date (unix):** 1717202268

#### Streaming renditions

- application/vnd.apple.mpegurl
- audio/mp4 · AAC Audio · 113468 kbps
- video/mp4 · 180p · 180p · 169563 kbps
- video/mp4 · 270p · 270p · 208295 kbps
- video/mp4 · 360p · 360p · 235963 kbps
- video/mp4 · 406p · 406p · 258088 kbps
- video/mp4 · 540p · 540p · 329180 kbps
- video/mp4 · 720p · 720p · 451610 kbps
- video/mp4 · 1080p · 1080p · 859550 kbps

#### Timed text tracks (delivery)

- **thumbnails:** `https://cdn.jwplayer.com/strips/WPNnZ8cY-120.vtt`

#### Transcript

Content Stack now offers new capabilities designed to seamlessly integrate with AI-generated content using a customer's unique brand style and messaging. It's a game-changer for both content creators and for automated workflows utilizing AI that ensure a unified brand voice across various channels. BrandKit operates by tapping into a wealth of brand-specific data and documents, which we call Knowledge Vault. It then uses predefined voice profiles to steer AI-generated content towards the desired brand tone. This ensures that the AI consistently delivers content that truly embodies the essence of the brand. But before we get into any of that, let's back up and let's talk about some Content Stack products that already use AI and how BrandKit can further refine the results we get in these products. I'm assuming at this point you have experience with both our Automate and our AI Assistant products, but if not, feel free to jump into any one of those courses here on Academy. But I will take a moment to recap just some of the benefits of each of these products. First we have AI Assistant. AI Assistant appears in the Content Editor and allows you to interact with AI to perform things like translations or get new suggestions for content like titles, descriptions, or keywords, all without leaving the application. No more cutting and pasting, everything you need is available right within Content Stack. Next is Content Stack's Automate, which is a powerful visual workflow builder that uses an automation-based trigger and action system. With Automate, you can streamline business logic and complete tasks without human intervention. A basic automation might be to configure the platform to notify another app to take some action when an event occurs. For instance, you can create an automation to send Slack notifications whenever an entry in Content Stack is created, updated, published, or deleted. Moreover, Automate has an AI component associated with it. Let's say you set up an automation to extract SEO keywords, write summaries, generate daily reports, or even write blog posts. All that automated generated content is powered by artificial intelligence. But as many of us know, generative AI isn't perfect. Oftentimes, generated content can sound, well, generated. Not necessarily like you, in worst case scenario, the AI hallucinates, which means it generates content that may sound perfectly believable, but in fact is completely false. And while a human can fix these issues, it doesn't do much for the automation of generated content. Now let's get back to BrandKit. As I mentioned, there's two core capabilities here. First is customized profiles. These profiles give you the ability to speak in your brand's voice, whether that be your organization as a whole or individuals on your team. The voice profile is sent with your prompt, which in turn generates results in a familiar voice instead of the generic generated results typically returned by generative AI platforms. Second, we have Knowledge Vault, which is a database of specific organization information. This provides additional accuracy when generating specific information on your products and services that it otherwise wouldn't have. Knowledge Vault, in combination with voice profiles, makes BrandKit your personalized AI solution for generating automated content that's both on brand and accurate. So how does it work? Well, we'll take a look at that next.

#### Subtitles (WebVTT)

```webvtt
WEBVTT

1
00:00:00.000 --> 00:00:15.680
Content Stack now offers new capabilities designed to seamlessly integrate with AI-generated

2
00:00:15.680 --> 00:00:21.360
content using a customer's unique brand style and messaging.

3
00:00:21.360 --> 00:00:28.460
It's a game-changer for both content creators and for automated workflows utilizing AI that

4
00:00:28.460 --> 00:00:34.360
ensure a unified brand voice across various channels.

5
00:00:34.360 --> 00:00:40.940
BrandKit operates by tapping into a wealth of brand-specific data and documents, which

6
00:00:40.940 --> 00:00:43.520
we call Knowledge Vault.

7
00:00:43.520 --> 00:00:50.480
It then uses predefined voice profiles to steer AI-generated content towards the desired

8
00:00:50.480 --> 00:00:52.020
brand tone.

9
00:00:52.020 --> 00:00:58.500
This ensures that the AI consistently delivers content that truly embodies the essence of

10
00:00:58.500 --> 00:00:59.940
the brand.

11
00:00:59.940 --> 00:01:05.460
But before we get into any of that, let's back up and let's talk about some Content

12
00:01:05.460 --> 00:01:12.000
Stack products that already use AI and how BrandKit can further refine the results we

13
00:01:12.000 --> 00:01:13.740
get in these products.

14
00:01:13.740 --> 00:01:19.420
I'm assuming at this point you have experience with both our Automate and our AI Assistant

15
00:01:19.420 --> 00:01:26.460
products, but if not, feel free to jump into any one of those courses here on Academy.

16
00:01:26.460 --> 00:01:33.020
But I will take a moment to recap just some of the benefits of each of these products.

17
00:01:33.020 --> 00:01:34.740
First we have AI Assistant.

18
00:01:34.740 --> 00:01:40.800
AI Assistant appears in the Content Editor and allows you to interact with AI to perform

19
00:01:40.800 --> 00:01:47.780
things like translations or get new suggestions for content like titles, descriptions, or

20
00:01:47.860 --> 00:01:51.860
keywords, all without leaving the application.

21
00:01:51.860 --> 00:01:58.220
No more cutting and pasting, everything you need is available right within Content Stack.

22
00:01:58.220 --> 00:02:04.220
Next is Content Stack's Automate, which is a powerful visual workflow builder that uses

23
00:02:04.220 --> 00:02:07.980
an automation-based trigger and action system.

24
00:02:07.980 --> 00:02:14.820
With Automate, you can streamline business logic and complete tasks without human intervention.

25
00:02:14.820 --> 00:02:19.980
A basic automation might be to configure the platform to notify another app to take

26
00:02:19.980 --> 00:02:23.300
some action when an event occurs.

27
00:02:23.300 --> 00:02:28.900
For instance, you can create an automation to send Slack notifications whenever an entry

28
00:02:28.900 --> 00:02:34.340
in Content Stack is created, updated, published, or deleted.

29
00:02:34.340 --> 00:02:39.140
Moreover, Automate has an AI component associated with it.

30
00:02:39.140 --> 00:02:45.660
Let's say you set up an automation to extract SEO keywords, write summaries, generate daily

31
00:02:45.660 --> 00:02:48.900
reports, or even write blog posts.

32
00:02:48.900 --> 00:02:55.540
All that automated generated content is powered by artificial intelligence.

33
00:02:55.540 --> 00:03:00.220
But as many of us know, generative AI isn't perfect.

34
00:03:00.220 --> 00:03:06.020
Oftentimes, generated content can sound, well, generated.

35
00:03:06.020 --> 00:03:13.140
Not necessarily like you, in worst case scenario, the AI hallucinates, which means it generates

36
00:03:13.140 --> 00:03:19.900
content that may sound perfectly believable, but in fact is completely false.

37
00:03:19.900 --> 00:03:28.460
And while a human can fix these issues, it doesn't do much for the automation of generated content.

38
00:03:28.460 --> 00:03:31.500
Now let's get back to BrandKit.

39
00:03:31.500 --> 00:03:34.860
As I mentioned, there's two core capabilities here.

40
00:03:34.860 --> 00:03:37.460
First is customized profiles.

41
00:03:37.460 --> 00:03:42.220
These profiles give you the ability to speak in your brand's voice, whether that be your

42
00:03:42.220 --> 00:03:46.380
organization as a whole or individuals on your team.

43
00:03:46.380 --> 00:03:52.180
The voice profile is sent with your prompt, which in turn generates results in a familiar

44
00:03:52.180 --> 00:03:59.220
voice instead of the generic generated results typically returned by generative AI platforms.

45
00:03:59.260 --> 00:04:05.820
Second, we have Knowledge Vault, which is a database of specific organization information.

46
00:04:05.820 --> 00:04:11.820
This provides additional accuracy when generating specific information on your products and

47
00:04:11.820 --> 00:04:15.660
services that it otherwise wouldn't have.

48
00:04:15.660 --> 00:04:21.900
Knowledge Vault, in combination with voice profiles, makes BrandKit your personalized

49
00:04:21.900 --> 00:04:28.620
AI solution for generating automated content that's both on brand and accurate.

50
00:04:28.780 --> 00:04:30.540
So how does it work?

51
00:04:30.540 --> 00:04:33.180
Well, we'll take a look at that next.

```

```transcript
<!-- PLACEHOLDER: replace with real transcript before publish if cues were auto-derived from WebVTT -->
[00:00] Content Stack now offers new capabilities designed to seamlessly integrate with AI-generated
[00:15] content using a customer's unique brand style and messaging.
[00:21] It's a game-changer for both content creators and for automated workflows utilizing AI that
[00:28] ensure a unified brand voice across various channels.
[00:34] BrandKit operates by tapping into a wealth of brand-specific data and documents, which
[00:40] we call Knowledge Vault.
[00:43] It then uses predefined voice profiles to steer AI-generated content towards the desired
[00:50] brand tone.
[00:52] This ensures that the AI consistently delivers content that truly embodies the essence of
[00:58] the brand.
[00:59] But before we get into any of that, let's back up and let's talk about some Content
[01:05] Stack products that already use AI and how BrandKit can further refine the results we
[01:12] get in these products.
[01:13] I'm assuming at this point you have experience with both our Automate and our AI Assistant
[01:19] products, but if not, feel free to jump into any one of those courses here on Academy.
[01:26] But I will take a moment to recap just some of the benefits of each of these products.
[01:33] First we have AI Assistant.
[01:34] AI Assistant appears in the Content Editor and allows you to interact with AI to perform
[01:40] things like translations or get new suggestions for content like titles, descriptions, or
[01:47] keywords, all without leaving the application.
[01:51] No more cutting and pasting, everything you need is available right within Content Stack.
[01:58] Next is Content Stack's Automate, which is a powerful visual workflow builder that uses
[02:04] an automation-based trigger and action system.
[02:07] With Automate, you can streamline business logic and complete tasks without human intervention.
[02:14] A basic automation might be to configure the platform to notify another app to take
[02:19] some action when an event occurs.
[02:23] For instance, you can create an automation to send Slack notifications whenever an entry
[02:28] in Content Stack is created, updated, published, or deleted.
[02:34] Moreover, Automate has an AI component associated with it.
[02:39] Let's say you set up an automation to extract SEO keywords, write summaries, generate daily
[02:45] reports, or even write blog posts.
[02:48] All that automated generated content is powered by artificial intelligence.
[02:55] But as many of us know, generative AI isn't perfect.
[03:00] Oftentimes, generated content can sound, well, generated.
[03:06] Not necessarily like you, in worst case scenario, the AI hallucinates, which means it generates
[03:13] content that may sound perfectly believable, but in fact is completely false.
[03:19] And while a human can fix these issues, it doesn't do much for the automation of generated content.
[03:28] Now let's get back to BrandKit.
[03:31] As I mentioned, there's two core capabilities here.
[03:34] First is customized profiles.
[03:37] These profiles give you the ability to speak in your brand's voice, whether that be your
[03:42] organization as a whole or individuals on your team.
[03:46] The voice profile is sent with your prompt, which in turn generates results in a familiar
[03:52] voice instead of the generic generated results typically returned by generative AI platforms.
[03:59] Second, we have Knowledge Vault, which is a database of specific organization information.
[04:05] This provides additional accuracy when generating specific information on your products and
[04:11] services that it otherwise wouldn't have.
[04:15] Knowledge Vault, in combination with voice profiles, makes BrandKit your personalized
[04:21] AI solution for generating automated content that's both on brand and accurate.
[04:28] So how does it work?
[04:30] Well, we'll take a look at that next.
```

#### Key takeaways

- Connect **Introducing Brand Kit** back to your stack configuration before moving to the next module.
- Capture one concrete artifact (screenshot, Postman call, or code snippet) that proves the step works in your environment.
- Re-read the delivery versus management boundary for anything you changed in the entry model.

### Lesson 02 — Voice Profiles

<!-- ai_metadata: {"lesson_id":"02","type":"video","duration_seconds":237,"video_url":"https://cdn.jwplayer.com/previews/XDtpLpU3","thumbnail_url":"https://cdn.jwplayer.com/v2/media/XDtpLpU3/poster.jpg?width=720","topics":["Voice","Profiles"]} -->

#### Video details

#### At a glance

- **Title:** Custom Profiles
- **Duration:** 3m 57s
- **Media link:** https://cdn.jwplayer.com/previews/XDtpLpU3
- **Publish date (unix):** 1717202268

#### Streaming renditions

- application/vnd.apple.mpegurl
- audio/mp4 · AAC Audio · 113465 kbps
- video/mp4 · 180p · 180p · 151652 kbps
- video/mp4 · 270p · 270p · 172629 kbps
- video/mp4 · 360p · 360p · 191876 kbps
- video/mp4 · 406p · 406p · 204841 kbps
- video/mp4 · 540p · 540p · 244901 kbps
- video/mp4 · 720p · 720p · 317290 kbps
- video/mp4 · 1080p · 1080p · 524489 kbps

#### Timed text tracks (delivery)

- **thumbnails:** `https://cdn.jwplayer.com/strips/XDtpLpU3-120.vtt`

#### Transcript

Setting up a new profile with Brand Kit is simple. First, you need to navigate to Brand Kit on the left-hand side. When you click the icon, it will bring you to a screen where you can create a new Brand Kit by clicking the New Brand Kit button. Here, you can name your Brand Kit, provide a description, and select what stack or stacks you want the Brand Kit to be associated with. This will then bring you into the Voice Profile screen where you can create a new Voice Profile. Start by giving the Voice Profile a name and a description. I'll call this Compass Travel Corporate. Next, you can adjust the sliders within the Communication Style Mixer. These settings are fairly self-explanatory. Do you want the formality of the tone to be more casual, business, or professional? You can adjust the tone of the voice to be more informative or persuasive, and the humor level. Do you want the voice to be more serious or a bit more lighthearted? Finally, you can set the language complexity level to either be more straightforward or technical. Of course, you can also leave the sliders set to None if you don't want to make an adjustment here, but we found that doing so gives you the best shot of getting the tone of the generated content to match that of someone on your team or your organization's overall communication style. Now, down below, there's a Custom Detail section. Insights can be used to provide additional information about your brand's tone to be sent with the prompt, again, giving you the ability to really enhance the results of the returned generated content. You can also provide some sample content. This will be used to guide the generated text to take on a certain style of writing. This sample doesn't provide any additional context about the generation, rather just a style to mimic. Once you're happy with the settings, you can click Save. Remember, you can create multiple voices. Maybe you want a general corporate voice like we created here, but you may also want to create profiles for others within your organization. Say, for example, Kim within the Compass Travel organization writes many blog posts for the website. We could create a voice profile matching her style, so any new blog posts written with Generative AI can sound like her. Okay, let's get back to our stacks by first clicking the Brand Kit icon in the top left, then the Stacks icon. To see one of these voice profiles in action, I'll create a new entry. To do that, I'll click the Entries icon on the left, then the new Entry button on the top right. I'll add a title to this entry, Great Summer Vacation Ideas, but maybe I'd like some help optimizing this title for SEO. I'll click the AI Assistant icon towards the top of the title text field. Notice Brand Kit appears at the top of the AI Assistant panel, where you can select the Brand Kit and the voice profile. Let's choose the one we just created, then in the Replace Field Content section, we can optimize this title for SEO, and when it replaces our initial title content, it'll use the voice of the Compass Travel corporate voice matching our other entries within our system.

#### Subtitles (WebVTT)

```webvtt
WEBVTT

1
00:00:00.000 --> 00:00:08.000
Setting up a new profile with Brand Kit is simple. First, you need to navigate to Brand

2
00:00:08.000 --> 00:00:13.760
Kit on the left-hand side. When you click the icon, it will bring you to a screen where

3
00:00:13.760 --> 00:00:19.980
you can create a new Brand Kit by clicking the New Brand Kit button. Here, you can name

4
00:00:19.980 --> 00:00:25.760
your Brand Kit, provide a description, and select what stack or stacks you want the Brand

5
00:00:25.760 --> 00:00:31.360
Kit to be associated with. This will then bring you into the Voice Profile screen where

6
00:00:31.360 --> 00:00:40.200
you can create a new Voice Profile. Start by giving the Voice Profile a name and a description.

7
00:00:40.200 --> 00:00:46.860
I'll call this Compass Travel Corporate. Next, you can adjust the sliders within the

8
00:00:46.860 --> 00:00:52.860
Communication Style Mixer. These settings are fairly self-explanatory. Do you want the

9
00:00:52.860 --> 00:00:58.300
formality of the tone to be more casual, business, or professional? You can adjust

10
00:00:58.300 --> 00:01:05.680
the tone of the voice to be more informative or persuasive, and the humor level. Do you

11
00:01:05.680 --> 00:01:11.380
want the voice to be more serious or a bit more lighthearted? Finally, you can set the

12
00:01:11.380 --> 00:01:17.660
language complexity level to either be more straightforward or technical. Of course, you

13
00:01:17.660 --> 00:01:24.120
can also leave the sliders set to None if you don't want to make an adjustment here,

14
00:01:24.120 --> 00:01:29.980
but we found that doing so gives you the best shot of getting the tone of the generated

15
00:01:29.980 --> 00:01:36.560
content to match that of someone on your team or your organization's overall communication

16
00:01:36.560 --> 00:01:44.880
style. Now, down below, there's a Custom Detail section. Insights can be used to provide additional

17
00:01:44.880 --> 00:01:50.920
information about your brand's tone to be sent with the prompt, again, giving you the

18
00:01:50.920 --> 00:01:58.440
ability to really enhance the results of the returned generated content. You can also provide

19
00:01:58.440 --> 00:02:04.840
some sample content. This will be used to guide the generated text to take on a certain

20
00:02:04.840 --> 00:02:12.760
style of writing. This sample doesn't provide any additional context about the generation,

21
00:02:12.800 --> 00:02:20.360
rather just a style to mimic. Once you're happy with the settings, you can click Save. Remember,

22
00:02:20.360 --> 00:02:27.400
you can create multiple voices. Maybe you want a general corporate voice like we created here,

23
00:02:27.400 --> 00:02:33.880
but you may also want to create profiles for others within your organization. Say,

24
00:02:33.880 --> 00:02:39.360
for example, Kim within the Compass Travel organization writes many blog posts for the

25
00:02:39.360 --> 00:02:46.200
website. We could create a voice profile matching her style, so any new blog posts

26
00:02:46.200 --> 00:02:53.320
written with Generative AI can sound like her. Okay, let's get back to our stacks by first

27
00:02:53.320 --> 00:02:59.720
clicking the Brand Kit icon in the top left, then the Stacks icon. To see one of these voice

28
00:02:59.720 --> 00:03:06.720
profiles in action, I'll create a new entry. To do that, I'll click the Entries icon on the left,

29
00:03:07.080 --> 00:03:12.680
then the new Entry button on the top right. I'll add a title to this entry,

30
00:03:12.680 --> 00:03:21.240
Great Summer Vacation Ideas, but maybe I'd like some help optimizing this title for SEO. I'll

31
00:03:21.240 --> 00:03:28.880
click the AI Assistant icon towards the top of the title text field. Notice Brand Kit appears

32
00:03:28.880 --> 00:03:35.960
at the top of the AI Assistant panel, where you can select the Brand Kit and the voice profile.

33
00:03:36.160 --> 00:03:42.400
Let's choose the one we just created, then in the Replace Field Content section,

34
00:03:42.400 --> 00:03:49.120
we can optimize this title for SEO, and when it replaces our initial title content, it'll use

35
00:03:49.120 --> 00:03:55.600
the voice of the Compass Travel corporate voice matching our other entries within our system.

```

```transcript
<!-- PLACEHOLDER: replace with real transcript before publish if cues were auto-derived from WebVTT -->
[00:00] Setting up a new profile with Brand Kit is simple. First, you need to navigate to Brand
[00:08] Kit on the left-hand side. When you click the icon, it will bring you to a screen where
[00:13] you can create a new Brand Kit by clicking the New Brand Kit button. Here, you can name
[00:19] your Brand Kit, provide a description, and select what stack or stacks you want the Brand
[00:25] Kit to be associated with. This will then bring you into the Voice Profile screen where
[00:31] you can create a new Voice Profile. Start by giving the Voice Profile a name and a description.
[00:40] I'll call this Compass Travel Corporate. Next, you can adjust the sliders within the
[00:46] Communication Style Mixer. These settings are fairly self-explanatory. Do you want the
[00:52] formality of the tone to be more casual, business, or professional? You can adjust
[00:58] the tone of the voice to be more informative or persuasive, and the humor level. Do you
[01:05] want the voice to be more serious or a bit more lighthearted? Finally, you can set the
[01:11] language complexity level to either be more straightforward or technical. Of course, you
[01:17] can also leave the sliders set to None if you don't want to make an adjustment here,
[01:24] but we found that doing so gives you the best shot of getting the tone of the generated
[01:29] content to match that of someone on your team or your organization's overall communication
[01:36] style. Now, down below, there's a Custom Detail section. Insights can be used to provide additional
[01:44] information about your brand's tone to be sent with the prompt, again, giving you the
[01:50] ability to really enhance the results of the returned generated content. You can also provide
[01:58] some sample content. This will be used to guide the generated text to take on a certain
[02:04] style of writing. This sample doesn't provide any additional context about the generation,
[02:12] rather just a style to mimic. Once you're happy with the settings, you can click Save. Remember,
[02:20] you can create multiple voices. Maybe you want a general corporate voice like we created here,
[02:27] but you may also want to create profiles for others within your organization. Say,
[02:33] for example, Kim within the Compass Travel organization writes many blog posts for the
[02:39] website. We could create a voice profile matching her style, so any new blog posts
[02:46] written with Generative AI can sound like her. Okay, let's get back to our stacks by first
[02:53] clicking the Brand Kit icon in the top left, then the Stacks icon. To see one of these voice
[02:59] profiles in action, I'll create a new entry. To do that, I'll click the Entries icon on the left,
[03:07] then the new Entry button on the top right. I'll add a title to this entry,
[03:12] Great Summer Vacation Ideas, but maybe I'd like some help optimizing this title for SEO. I'll
[03:21] click the AI Assistant icon towards the top of the title text field. Notice Brand Kit appears
[03:28] at the top of the AI Assistant panel, where you can select the Brand Kit and the voice profile.
[03:36] Let's choose the one we just created, then in the Replace Field Content section,
[03:42] we can optimize this title for SEO, and when it replaces our initial title content, it'll use
[03:49] the voice of the Compass Travel corporate voice matching our other entries within our system.
```

#### Key takeaways

- Connect **Voice Profiles** back to your stack configuration before moving to the next module.
- Capture one concrete artifact (screenshot, Postman call, or code snippet) that proves the step works in your environment.
- Re-read the delivery versus management boundary for anything you changed in the entry model.

### Lesson 03 — Knowledge Vault

<!-- ai_metadata: {"lesson_id":"03","type":"video","duration_seconds":254,"video_url":"https://cdn.jwplayer.com/previews/M1oETwD7","thumbnail_url":"https://cdn.jwplayer.com/v2/media/M1oETwD7/poster.jpg?width=720","topics":["Knowledge","Vault"]} -->

#### Video details

#### At a glance

- **Title:** Knowledge Vault-
- **Duration:** 4m 14s
- **Media link:** https://cdn.jwplayer.com/previews/M1oETwD7
- **Publish date (unix):** 1717202348

#### Streaming renditions

- application/vnd.apple.mpegurl
- audio/mp4 · AAC Audio · 113582 kbps
- video/mp4 · 180p · 180p · 154642 kbps
- video/mp4 · 270p · 270p · 184264 kbps
- video/mp4 · 360p · 360p · 208745 kbps
- video/mp4 · 406p · 406p · 225191 kbps
- video/mp4 · 540p · 540p · 279025 kbps
- video/mp4 · 720p · 720p · 375567 kbps
- video/mp4 · 1080p · 1080p · 663873 kbps

#### Timed text tracks (delivery)

- **thumbnails:** `https://cdn.jwplayer.com/strips/M1oETwD7-120.vtt`

#### Transcript

Knowledge Vault will act as the central hub for all your brand's essential materials. It's where you can manage, store, and organize your documents, data, and content related to the brand. This repository acts as a reference point for Generative AI, ensuring that the content it creates is based on accurate brand information. The Vault can house everything from brand guidelines and product specifications to help center resources, historical content, and any other documents that capture the personality of your brand's voice and style. So how does it work? A Knowledge Vault functions as a specialized vector database, tailor-made for storing and retrieving information that you have curated in a format that's ideal for AI language models. Unlike traditional databases that keep text in its raw form, a vector database transforms text into numerical formats known as embeddings or vectors. These embeddings encapsulate the semantic meanings and relationships between words and concepts, enabling AI to grasp the context and subtleties of the information contained within the Knowledge Vault. When a user or program like Automate requests a content generation task, such as creating a product description, the AI first consults the Knowledge Vault. The Knowledge Vault then identifies the most relevant information in the database by comparing the query with the stored embeddings for semantic similarity. This relevant information, along with the additional details like a voice profile, is compiled to form the prompt sent to the AI model. This accurate background information grounds the AI in factual data specific to the company, significantly enhancing the relevancy and accuracy of the generated content. Knowing how it works is really helpful as you think about what type of content to include in the Vault. The most effective content for a Knowledge Vault centers on factual, company-specific information. General information is likely already included in the large language models training and doesn't need to be added. Things like company background, which would include company history, a mission statement, core values, brand guidelines, and style guidelines. You may also want to include information on your products and services, like descriptions, technical specifications, and use cases. Maybe you want to add training materials, sales presentations, or white papers. All of these are highly relevant to your organization. Of course, there's pieces of content that you probably want to avoid including in your Knowledge Vault. Outdated documentation, old blog posts, confidential information, and ambiguous content. Ultimately, you're looking for quality content that will support your brand's overall voice. There's two ways you can populate a Knowledge Vault, either through APIs or you can use our recipes within Automate. To show you where to find BrandKit within Automate, navigate to Automate on the left-hand side. Once you click the icon, you'll enter the Automate screen. And here, you can create a new automation. We'll go ahead and provide it a name. I'll call this Compass Knowledge Vault. Then, if you move to the Action step, you'll notice there's a new connector for BrandKit. When you select it, notice all the different actions for Knowledge Vault and Voice Profile. Setting up an automation is simple for BrandKit. If you need to know more about automations, check out our Automate Foundation course here on Academy.

#### Subtitles (WebVTT)

```webvtt
WEBVTT

1
00:00:00.000 --> 00:00:05.520
Knowledge Vault will act as the central hub for all your brand's essential

2
00:00:05.520 --> 00:00:11.280
materials. It's where you can manage, store, and organize your documents, data,

3
00:00:11.280 --> 00:00:18.160
and content related to the brand. This repository acts as a reference point for

4
00:00:18.160 --> 00:00:23.840
Generative AI, ensuring that the content it creates is based on accurate brand

5
00:00:23.840 --> 00:00:29.120
information. The Vault can house everything from brand guidelines and

6
00:00:29.120 --> 00:00:35.040
product specifications to help center resources, historical content, and any

7
00:00:35.040 --> 00:00:39.840
other documents that capture the personality of your brand's voice and

8
00:00:39.840 --> 00:00:46.160
style. So how does it work? A Knowledge Vault functions as a specialized vector

9
00:00:46.160 --> 00:00:51.360
database, tailor-made for storing and retrieving information that you have

10
00:00:51.360 --> 00:00:58.160
curated in a format that's ideal for AI language models. Unlike traditional

11
00:00:58.160 --> 00:01:04.480
databases that keep text in its raw form, a vector database transforms text

12
00:01:04.480 --> 00:01:10.400
into numerical formats known as embeddings or vectors. These embeddings

13
00:01:10.400 --> 00:01:16.160
encapsulate the semantic meanings and relationships between words and concepts,

14
00:01:16.160 --> 00:01:22.000
enabling AI to grasp the context and subtleties of the information contained

15
00:01:22.000 --> 00:01:28.160
within the Knowledge Vault. When a user or program like Automate requests a

16
00:01:28.160 --> 00:01:33.920
content generation task, such as creating a product description, the AI first

17
00:01:33.920 --> 00:01:38.560
consults the Knowledge Vault. The Knowledge Vault then identifies the most

18
00:01:38.560 --> 00:01:43.920
relevant information in the database by comparing the query with the stored

19
00:01:43.920 --> 00:01:49.520
embeddings for semantic similarity. This relevant information, along with the

20
00:01:49.520 --> 00:01:55.840
additional details like a voice profile, is compiled to form the prompt sent to

21
00:01:55.840 --> 00:02:01.360
the AI model. This accurate background information grounds the AI in factual

22
00:02:01.360 --> 00:02:06.320
data specific to the company, significantly enhancing the relevancy

23
00:02:06.320 --> 00:02:12.320
and accuracy of the generated content. Knowing how it works is really helpful

24
00:02:12.320 --> 00:02:17.200
as you think about what type of content to include in the Vault. The most

25
00:02:17.200 --> 00:02:22.080
effective content for a Knowledge Vault centers on factual, company-specific

26
00:02:22.080 --> 00:02:27.440
information. General information is likely already included in the large

27
00:02:27.440 --> 00:02:33.040
language models training and doesn't need to be added. Things like company

28
00:02:33.040 --> 00:02:38.080
background, which would include company history, a mission statement, core values,

29
00:02:38.080 --> 00:02:42.960
brand guidelines, and style guidelines. You may also want to include information

30
00:02:42.960 --> 00:02:48.640
on your products and services, like descriptions, technical specifications, and

31
00:02:48.640 --> 00:02:54.080
use cases. Maybe you want to add training materials, sales presentations, or

32
00:02:54.080 --> 00:03:00.720
white papers. All of these are highly relevant to your organization. Of course,

33
00:03:00.720 --> 00:03:05.040
there's pieces of content that you probably want to avoid including in your

34
00:03:05.040 --> 00:03:11.360
Knowledge Vault. Outdated documentation, old blog posts, confidential information,

35
00:03:11.360 --> 00:03:17.040
and ambiguous content. Ultimately, you're looking for quality content that will

36
00:03:17.040 --> 00:03:22.720
support your brand's overall voice. There's two ways you can populate a

37
00:03:22.720 --> 00:03:28.960
Knowledge Vault, either through APIs or you can use our recipes within Automate.

38
00:03:28.960 --> 00:03:34.080
To show you where to find BrandKit within Automate, navigate to Automate on

39
00:03:34.080 --> 00:03:39.760
the left-hand side. Once you click the icon, you'll enter the Automate screen.

40
00:03:39.760 --> 00:03:45.920
And here, you can create a new automation. We'll go ahead and provide it a name.

41
00:03:45.920 --> 00:03:51.600
I'll call this Compass Knowledge Vault. Then, if you move to the Action step,

42
00:03:51.600 --> 00:03:58.400
you'll notice there's a new connector for BrandKit. When you select it, notice all

43
00:03:58.400 --> 00:04:03.920
the different actions for Knowledge Vault and Voice Profile. Setting up an

44
00:04:03.920 --> 00:04:09.360
automation is simple for BrandKit. If you need to know more about automations,

45
00:04:09.360 --> 00:04:13.680
check out our Automate Foundation course here on Academy.

```

```transcript
<!-- PLACEHOLDER: replace with real transcript before publish if cues were auto-derived from WebVTT -->
[00:00] Knowledge Vault will act as the central hub for all your brand's essential
[00:05] materials. It's where you can manage, store, and organize your documents, data,
[00:11] and content related to the brand. This repository acts as a reference point for
[00:18] Generative AI, ensuring that the content it creates is based on accurate brand
[00:23] information. The Vault can house everything from brand guidelines and
[00:29] product specifications to help center resources, historical content, and any
[00:35] other documents that capture the personality of your brand's voice and
[00:39] style. So how does it work? A Knowledge Vault functions as a specialized vector
[00:46] database, tailor-made for storing and retrieving information that you have
[00:51] curated in a format that's ideal for AI language models. Unlike traditional
[00:58] databases that keep text in its raw form, a vector database transforms text
[01:04] into numerical formats known as embeddings or vectors. These embeddings
[01:10] encapsulate the semantic meanings and relationships between words and concepts,
[01:16] enabling AI to grasp the context and subtleties of the information contained
[01:22] within the Knowledge Vault. When a user or program like Automate requests a
[01:28] content generation task, such as creating a product description, the AI first
[01:33] consults the Knowledge Vault. The Knowledge Vault then identifies the most
[01:38] relevant information in the database by comparing the query with the stored
[01:43] embeddings for semantic similarity. This relevant information, along with the
[01:49] additional details like a voice profile, is compiled to form the prompt sent to
[01:55] the AI model. This accurate background information grounds the AI in factual
[02:01] data specific to the company, significantly enhancing the relevancy
[02:06] and accuracy of the generated content. Knowing how it works is really helpful
[02:12] as you think about what type of content to include in the Vault. The most
[02:17] effective content for a Knowledge Vault centers on factual, company-specific
[02:22] information. General information is likely already included in the large
[02:27] language models training and doesn't need to be added. Things like company
[02:33] background, which would include company history, a mission statement, core values,
[02:38] brand guidelines, and style guidelines. You may also want to include information
[02:42] on your products and services, like descriptions, technical specifications, and
[02:48] use cases. Maybe you want to add training materials, sales presentations, or
[02:54] white papers. All of these are highly relevant to your organization. Of course,
[03:00] there's pieces of content that you probably want to avoid including in your
[03:05] Knowledge Vault. Outdated documentation, old blog posts, confidential information,
[03:11] and ambiguous content. Ultimately, you're looking for quality content that will
[03:17] support your brand's overall voice. There's two ways you can populate a
[03:22] Knowledge Vault, either through APIs or you can use our recipes within Automate.
[03:28] To show you where to find BrandKit within Automate, navigate to Automate on
[03:34] the left-hand side. Once you click the icon, you'll enter the Automate screen.
[03:39] And here, you can create a new automation. We'll go ahead and provide it a name.
[03:45] I'll call this Compass Knowledge Vault. Then, if you move to the Action step,
[03:51] you'll notice there's a new connector for BrandKit. When you select it, notice all
[03:58] the different actions for Knowledge Vault and Voice Profile. Setting up an
[04:03] automation is simple for BrandKit. If you need to know more about automations,
[04:09] check out our Automate Foundation course here on Academy.
```

#### Key takeaways

- Connect **Knowledge Vault** back to your stack configuration before moving to the next module.
- Capture one concrete artifact (screenshot, Postman call, or code snippet) that proves the step works in your environment.
- Re-read the delivery versus management boundary for anything you changed in the entry model.

## Resources & references

| Page | Companion Markdown |
| --- | --- |
| /courses/brand-kit-foundations/introducing-brand-kit | /academy/md/courses/brand-kit-foundations/introducing-brand-kit.md |
| /courses/brand-kit-foundations/voice-profiles | /academy/md/courses/brand-kit-foundations/voice-profiles.md |
| /courses/brand-kit-foundations/knowledge-vault | /academy/md/courses/brand-kit-foundations/knowledge-vault.md |

## Supplement for indexing

### Content summary

Make Generative AI More Powerful Contentstack's Brand Kit allows you to maintain a consistent brand voice across all channels. Effortlessly generate content that aligns with your brand’s personality and resonates with yo… Make Generative AI More Powerful Contentstack's Brand Kit allows you to maintain a consistent brand voice across all channels. Effortlessly generate content that aligns with your brand’s personality and resonates with your audience. Brand Kit includes Knowledge Vault - a centralized way for you to store your company’s essential information. This will empower AI to produce factually accurate content that supports your business objectives.

### Retrieval tags

- ai
- brand kit
- automate
- brand voice
- voice profiles
- brand-kit-foundations
- Introducing
- Brand
- Kit
- Voice
- Profiles
- Knowledge
- Vault
- brand-kit-foundations course

### Indexing notes

Chunk at each "### Lesson NN — Title" heading; copy lesson_id and topics from the preceding HTML comment into chunk metadata for RAG filters.
Course slug: brand-kit-foundations. Union of lesson topic tokens: Introducing, Brand, Kit, Voice, Profiles, Knowledge, Vault.
Do not embed or retrieve LMS-only quiz items or mastery exam answer keys from this export.

### Asset references

| Label | URL |
| --- | --- |
| Video thumbnail: Introducing Brand Kit | `https://cdn.jwplayer.com/v2/media/WPNnZ8cY/poster.jpg?width=720` |
| Video thumbnail: Voice Profiles | `https://cdn.jwplayer.com/v2/media/XDtpLpU3/poster.jpg?width=720` |
| Video thumbnail: Knowledge Vault | `https://cdn.jwplayer.com/v2/media/M1oETwD7/poster.jpg?width=720` |

### External links

| Label | URL |
| --- | --- |
| Contentstack Academy home | `https://www.contentstack.com/academy/` |
| Training instance setup | `https://www.contentstack.com/academy/training-instance` |
| Academy playground (GitHub) | `https://github.com/contentstack/contentstack-academy-playground` |
| Contentstack documentation | `https://www.contentstack.com/docs/` |
