Articles Archives

January 6, 2025

Manage Adobe Analytics Components in Excel Using APIs – Simplify Your Workflow

The Adobe Analytics Component Manager by Datacroft is an essential tool for many organizations using Adobe Analytics.

Several of our clients use and benefit from Datacroft's Component Manager. It's a fantastic tool, so a shout-out to Lukas Oldenburg for his work on it.

I strongly recommend you check it out if you're not familiar with it already.

If you are familiar with it, you know it works in Google Sheets, which not all our clients are excited about for various reasons. This has got me thinking that it should be possible to use the Adobe Analytics APIs in Excel.

Kasper Andersen, Partner

Like everyone else, we at Accrease are investigating how we can better use AI internally, with our clients, and in our deliveries.

So, last week, I decided to see how far I could push AI as a developer on something I had yet to gain exposure to. I thought it would be a good project to see if I could integrate the Adobe APIs into Microsoft Excel.

I am by no means a developer. Javascript I can read, but I used to Google it if I needed to write. Nowadays, I GPT it— if that's even a thing?

I use chatGPT daily, from being my personal developer to checking my grammar and summarizing notes. Even talk with it when driving to and from work. I know, I really need to find some friends.

My initial thought was to have chatGPT support me in building an Excel add-on, but after explaining my requirements, it suggested using Office Scripts?!

What are office scripts?

I've had no previous exposure to Office Scripts, so if you're the same, let me give you a quick intro to them.

Office Scripts is a tool in Microsoft Excel that helps you automate repetitive tasks. It makes it easy for everyone to use, even if you don't know how to code. Excel even comes with some out-of-the-box scripts to make your life easier. You can either record an action for it or code it using TypeScript, which is basically JavaScript.

They are only available in Office 365, but you can share them across users and organizations.

Now, you might be thinking. What's the difference between Office Scripts and traditional Excel Macros? Office Scripts are newer and easier to use than older Excel macros. Macros use a language called VBA, but Office Scripts uses TypeScript/JavaScript. Office Scripts also work with the online version of Excel, making them more convenient for using different devices.

Whether I've piqued your interest or not, if you're working in Excel daily, you should do yourself a favor and watch a few YouTube videos on the topic.

The result

It took about a week—a lot of back-and-forth with chatGPT and tests in Postman (for the Analytics APIs) and copy-pasting the office scripts. Whenever Excel generated an error, I would paste the error message back into chatGPT, asking it to address the error and creating a new script.

Once I had a functioning script, I would ask chatGPT to optimize it without breaking its functionality, which often resulted in a much smaller script.

I'm pretty impressed by the results myself and how AI can function as a developer — I have no real previous experience with the APIs and certainly not with Office Scripts, and today, we have a spreadsheet that works both on the desktop and online version of Excel.

For now, the functionality is the following:

Extract Workspaces: See usage and edit them straight in the spreadsheet.
Extract users: See their activities broken down by 4 previous quarters.
Extract report suites: Provides an overview of your Report Suites and Virtual Report Suites.
Edit report suites: Allows editing/enabling/disabling variables in one report suite at a time.

Let's get you started

Enough talk; let's get your set up to take it for a spin.

1. Download the sheet

Start by downloading the sheet by filling out the form below:

2. Configure API connection

Once you have the sheet, the first step is to go to the config tab. You must fill out A11 (Client ID) and C11 (Access Token).

To get this information, you need to create a project on Adobe.io. Here are the high-level steps:

Create a new project.
Click add API.
Select Adobe Analytics.
Select OAuth Server-to-Server.
Once the project is created you can find the credentials on the Project Overview page under Credentials.
Copy and paste them into the cells, and click the button, Get OrgId.

If successful, B11 will be populated with your Global Company ID, and you will be ready to use the other sheets.

IMPORTANT: Be aware that the sheet is not automatically getting the access token; why you just copied and pasted it. This also means that it expires after 24 hours. If that happens, you'll get an error when using the spreadsheet. All you have to do is go to your Project Overview > Credentials and click Generate new Access Token, which you copy and paste into the sheet again.

3. Start using the sheet

Hopefully, the buttons are self-explanatory. The Pull- buttons will extract and fill the sheet. You cannot mess anything up if you don't click the Push- buttons, as this will push changes back into Analytics. But let's quickly go through the sheets:

Report Suite: Start here, as it functions as the source for the dropdown in the Report Suite Editor sheet. Other than that, it only has the purpose of giving you an overview of your report suites.

Report Suite Editor: Select your report suites from the dropdown along with the variable type you wish to load. Click the pull button, and they will load. If you make any edits, the row will be highlighted in yellow - if you edit it back to its original value, it will still be highlighted. So, to 'reset' the sheet, you click the pull button again. Edits will populate the last column with a 'True.' Clicking the Push button will push all rows marked with True back into Adobe Analytics. After the push, you should click the Pull button to refresh and see your updates have been applied.

Users: The purpose of this sheet is to provide an overview of the users and how often they log in. Besides giving a general overview over the past 4 quarters, it will also provide an overview of users who have not logged in the last 12 and 3 months. Please note: The user's list is pulled from within Adobe Analytics (Admin > Analytics Users & Assets). It will not correspond with the list of users from within the Product Profiles in the Admin Console.
Workspaces: This sheet allows you to pull all the workspaces in the account with all the details required to understand when it was created, which report suite it is used on, who created it, etc. You can delete and make edits to the workspaces.

If you edit details for a workspace, e.g., change its name, the row will be marked yellow, and column B (Action) will be pre-populated with EDIT.

If you want to delete a workspace, click the cell in column B (Action) and select DELETE. This will mark the row with red, indicating it will be deleted. If you remove the DELETE value or redo any edits to the original value, it will still be highlighted; you can pull data again to reset.

If you make any edits, the 'Action' column will have a value, and the rows will be highlighted in color. Based on this, the API will push updates for all workspaces with an EDIT or DELETE value set.

After the push, you should hit the pull button to see the updates to the Workspaces.

Final note

You are responsible for using this tool, and Accrease is not liable for any outcomes resulting from its use.
The sheet uses macros, which you should enable. Macros look for edits and highlight the row that is edited. I could not get this working with Office Scripts.
Remember, it requires a work/school account with Microsoft 365, as Office Scripts are unavailable on personal accounts.
Feel free to copy the Office Scripts if you can use them elsewhere.

I'd love to hear your thoughts!

Please comment below if you found this helpful or have suggestions for additional features. We'll continue to expand on the features based on the input we receive.

Published by Kasper Andersen in Articles

October 23, 2024

Use Adobe Target as a CDP: A step-by-step guide to unlocking personalization

If you work in digital marketing, you cannot avoid having been exposed to conversations about how a CDP fixes most, if not all, of your problems. For companies who have bought into the Adobe stack, you'll likely be looking at Adobe RTCDP.

A CDP isn't an impulse purchase (if such a thing exists in Marketing) but is often a long process. However, this shouldn't put you in a waiting position. If you have Adobe Target, you can start your CDP journey today and harvest some low-hanging fruits. If you need to consider a CDP, then using Target can help you build a business case for why purchasing one is valid.

Kasper Andersen, Partner

In this article, you will learn how Adobe Target can work like a Customer Data Platform (CDP), its good and bad parts, and how it can help you move towards using a complete CDP. This guide will show easy steps to help you get started.

Can Adobe Target Work as a CDP?

Adobe Target has many features similar to those of a real CDP. This means it can be a good choice for businesses that want to use customer data to make things more personal without paying for a full CDP. However, there are some differences between Adobe Target and a real CDP.

A traditional CDP collects data from many sources, such as websites, emails, and in-store purchases, and combines it with ID stitching to create a complete picture of each customer. Adobe Target, however, doesn't have the same advanced ID stitching and is primarily intended for testing and improving experiences. Adobe Target can collect and use customer data, but it only has some tools a real CDP has to gather data from many sources.

Still, Adobe Target does have some CDP-like functions when used with Adobe Experience Cloud. Customer Attributes can bring in offline data like purchase history and CRM data. For example, Adobe Target can use information from a CRM to suggest products that customers might like. This can help create a consistent experience across all channels.

Adobe Target is good at combining customer data to create groups and personalize messages, just like a CDP. However, personalization will be on a group level rather than the 1:1 personalization a CDP can deliver.

How Adobe Target Collects Data

Adobe Target has some tools that help collect and organize customer data. When a visitor visits your site, Target creates a Profile based on the cookie ID and 'stores' information in the Profile that you tell it to.

This is done using mbox profile parameters.

This is great for returning visitors, as you can now personalize and promote messages based on the visitor's previous behavior.

Personalization goes wrong when the same visitor accesses the site from a new device. A new device means a new cookie and a new Target profile.

Target has a fix for that.

If your visitors are logging in, you can pass a DeclaredID to Adobe's Visitor ID Service or through a mbox parameter named mbox3rdPartyId.

When mbox3rdPartyId (or the Visitor ID Service) is set, it stitches the cookie IDs across devices to ensure that the Profile stays the same across devices.

Now, it begins getting interesting as Target begins behaving like a CDP.

One key feature available when setting the DeclaredID is using Customer Attributes. This lets Adobe Target collect data from CRM, Call centers, and other sources all in one place. For example, a business can use Adobe Target to target customers who are evaluated as having a high churn risk and use that information to show a relevant message to reduce the churn.

Adobe Target can also collect data in real time, which helps keep customer profiles up to date. If a customer leaves items in their shopping cart, Adobe Target can offer a discount immediately to help close the sale.

Benefits of Using Adobe Target as a CDP

- Personalized Experiences: Adobe Target allows grouping more significant segments and personalization of them based on their common attributes.

- Real-Time Responses: Adobe Target can react to customer actions instantly. For example, if a user is looking at a specific product, Adobe Target can immediately show them more related products.

- Works with Adobe Tools: Adobe Target works well with other tools like Adobe Analytics. This helps businesses get more information about their customers and use it to improve their marketing.

Challenges of Using Adobe Target as a CDP

Not as Powerful as a Full CDP: Adobe Target has tools different from those of a real CDP, which does not allow one to get a complete view of each customer.

Data Silos: Adobe Target is limited to how much offline data can be onboarded.

ID Stitching: Adobe Target has fundamental ID stitching, which cannot be compared to a real CDP.

Step-by-Step Guide: Setting Up Adobe Target as a CDP

Overall, there are three steps you need to perform to begin to do cross-device personalization with CRM data.

Pass DeclaredID to Adobe
Configure Customer Attributes (import CRM or other data sources)
Create an Audience off of your newly imported CRM data

Pass DeclaredID to Adobe

As mentioned, this can be done using Visitor ID Service and the mbox parameter mbox3rdPartyId. I'll only walk you through the simplest of the two, the mbox3rdPartyId approach.

Ideally, your visitor's DeclaredID is already exposed in your data layer when users log in. If not, your IT department might have to expose it for you.

I'm also assuming you're using Adobe Tags (aka. Adobe Launch) as your Tag Manager solution and already know how it works regarding libraries and publishing flows. Here goes:

1. Create a data element for the DeclaredID (if it doesn't already exist) and map it to your data layer element. For me, it looks like this:

2. You may already have an existing rule that triggers Target on all pages. You can edit this or create a new one.

3. In your actions of the rule, add or click on the existing Adobe Target - Add Params to All Requests.

4. Add a new parameter mapped to the data element in step 1. Make sure the parameter name is spelled correctly. Otherwise, it won't function as expected.

5. Publish your new changes.

Technically, you're done now.

By deploying the mbox3rdPartyId, you can now do cross-device personalization, as Target will now ID sti...or rather, cookie stitch visitors with a DeclaredID. Of course, the visitor must log in on both devices for Target to stitch.

Your segmentation is still limited to behavior collected by Target, so let's move on to ensure you can segment on offline/CRM data.

Configure Customer Attributes

Customer Attributes allows you to import offline data and tie it to the DeclaredID. Any offline data you can map to the DeclaredID can be uploaded into Customer Attributes, and by doing so, it becomes available to create Audiences in Adobe Target.

Data points examples:

- Membership level/type
- Membership points
- Next best offers
- Next best action
- Churn risk
- Customer Lifetime Value
- Lead score
- etc.

Before you start, ensure you know what data you want to upload and that the data can be mapped to the DeclaredID set on your site. Once you have that, you can get started with configuring the Customer Attributes:

1. Navigate to People

2. Navigate to Customer Attributes in the top menu.

3. Click +New

4. Creating the new Customer Attribute is pretty straightforward. The most important field is the Alias ID. When taking the mbox3rdPartyId approach, you can give it any name. If you've taken the Visitor ID Service, it is important that the alias matches the alias you've set in the implementation.

5. Upload your file with the data and confirm the schema. Ensure that the first column contains the DeclaredID.

6. Finally, activate the Customer Attribute for Target.

Depending on your Target license, you might be limited to activating only three columns at a time.

More importantly, you can automate the file upload. You could do this via your Marketing Automation tool (e.g., Adobe Campaign, Marketo, or something else).

The benefit here is that you—in most cases—already have your CRM data integrated into these solutions. So, instead of creating a new integration, you benefit from having your Marketing Automation spit out a daily, weekly, and monthly file that Customer Attributes then pick up and load to ensure your data is kept up to date.

Create an Audience with your new onboarded data

If you've been able to follow along this far, then it is time to be excited, as this is where the fun begins.

When creating a new Audience, you should select Visitor Profile. Here, your Customer Attribute will be auto-populated with the data columns that you activated.

It is simply going crazy building Audiences and using them in your activities.

Let's recap what you've achieved so far:

- You're setting a DeclaredID for your visitors who log in. This allows you to A/B test and personalize cross-device.

You've enabled and activated a Customer Attribute that imports offline data, allowing you to build segments on this data — a classic use case for why to invest in a CDP.

Recap of the benefits of using Adobe Target as a CDP

- Works with Adobe Tools: Adobe Target works well with Adobe Analytics and other products.

- Real-Time Personalization: It uses real-time data to give personalized content right away.

- Prepare for a CDP: Investing in a CDP involves getting teams together. You can slowly kickstart that process and even build a business case for why an investment in a CDP makes sense for your business.

- Eager Businesses: It's suitable for companies who wants to move NOW and cannot wait for when budget is available.

- Keep in mind that Adobe Target is not a CDP. However, it can be a great place to start for most companies considering a CDP but are struggling to build a proper business case or if you want to take your personalization program to the next level.

Conclusion: Is Adobe Target a Good CDP for You?

Adobe Target is a good option if you want to start creating personalized experiences for your customers without a complete CDP. It has tools for using customer data, creating audience groups, and providing real-time customized content. Adobe Target also works well with other Adobe tools, which makes it a good choice for many businesses.

If you are starting with personalization or already using Adobe products, Adobe Target can help you get many of the benefits of a CDP without the high cost. It's beneficial for smaller businesses that want to use customer data to improve the customer journey.

If you want to learn more about Adobe Target as a CDP or how other Adobe tools can help you use your customer data better, contact us! We'd love to hear from you. Leave a comment or share this post if you found it helpful!

Published by Kasper Andersen in Articles, Insights

Tags: Adobe Target, CDP

October 7, 2024

Enabling scroll depth tracking in Adobe Analytics with the new WebSDK Plugin

Tracking how much of a webpage users actually view is crucial to understanding engagement with your content. Whether you’re analyzing long-form articles, product pages, or campaign landing pages, scroll depth provides direct insight into how far users are going before leaving the page.

For years, the getPercentPageViewed plugin has helped marketers track scroll depth in Adobe Analytics. However, with the adoption of Adobe's new WebSDK, the old plugin is no longer compatible, leaving organizations that rely on modern architectures with a gap in their analytics setup.

We’ve developed a new scroll depth plugin specifically for Adobe’s WebSDK, offering a more robust, flexible, and scalable solution for tracking user engagement. In this post, we’ll explain the problem with the legacy plugin, demonstrate how to visualize scroll depth data, and guide you through the implementation of the new WebSDK-based solution

How to use scroll depth data in Adobe Analytics to identify drop-off points

One of the most valuable insights scroll depth tracking offers is identifying where users drop off on your pages. For landing pages and key CTA pages, this is critical, as it reveals whether visitors are engaging with your content long enough to reach the main calls to action.

Pinpointing drop-off points

Scroll depth data allows you to see exactly how far down the page users scroll before they exit or lose interest. If a significant portion of your visitors drop off before reaching the key messaging or CTA, this indicates a potential issue in your content flow, page design, or CTA placement. By identifying these drop-off points, you can take action to reduce friction and keep users engaged.

Consider screen size: percentages vs. absolute values

When analyzing scroll depth, it’s important to take screen size into account. Users on mobile devices may experience your page differently from desktop users, as content is displayed in a more compressed format. For this reason, it’s often helpful to measure scroll depth both in percentage terms (how far down the page they’ve scrolled) and absolute values(such as pixels or viewport units).

Percentage-based tracking: This helps normalize data across devices, making it easier to see the relative point where users drop off, regardless of screen size. For example, if 50% of mobile and desktop users are dropping off at around the same percentage mark, this provides a clear signal about engagement issues.
Absolute value tracking: This gives you more granular control over understanding specific user behaviors, especially when dealing with longer pages or dynamic content that may load differently across devices.

Mobile vs. non-mobile device analysis

Analyzing scroll depth data separately for mobile and non-mobile (desktop or tablet) devices is crucial for understanding the full picture. Given the different user behaviors and screen sizes, you may find that mobile users drop off sooner due to the nature of scrolling on smaller screens, while desktop users may scroll further before disengaging.

By segmenting your data, you can:

Mobile devices: Adjust content and CTA placement to ensure critical information is visible early on smaller screens, where users typically scroll faster.
Non-mobile devices: For desktops or larger screens, you may have more flexibility in content layout, allowing for CTAs to be placed further down the page, provided engagement levels remain high.

Example: improving CTA engagement on a landing Page

Imagine you’re running a campaign with a dedicated landing page designed to collect sign-ups. After reviewing the scroll depth data segmented by device type, you notice that the majority of mobile users drop off at 20%, while desktop users continue scrolling to 37%, but your sign-up form and CTA button are placed at 70%.

By moving the form higher, say around the 15% mark for mobile users and 35% for desktop users, you ensure that more visitors see the CTA before they drop off. This significantly increases the likelihood of conversions across both device types. Scroll depth tracking helps you make these data-driven adjustments, optimizing the performance of your landing pages for all users.

How to Implement the New Scroll Depth Plugin in Adobe WebSDK

Now that we've covered how scroll depth data helps identify drop-off points and optimize your landing pages, let’s look at how to implement the new plugin in Adobe WebSDK.

Step 1: Add the following code snippet at the end of the "send event" action in your page view rule:

//clear previous page data

_satellite.cookie.remove('initialPercent');

_satellite.cookie.remove('highestPercent');

// Function to set a cookie with an optional expiration in minutes

function setCookie(name, value, minutes) {

    var expires = "";

    if (minutes) {

        var date = new Date();

        date.setTime(date.getTime() + (minutes * 60 * 1000));

        expires = "; expires=" + date.toUTCString();

    }

    document.cookie = name + "=" + (value || "") + expires + "; path=/";  

}

// Function to calculate the percentage of the page viewed

function getScrollPercent() {

    const scrollTop = window.pageYOffset || document.documentElement.scrollTop;

    const scrollHeight = document.documentElement.scrollHeight - document.documentElement.clientHeight;

    return Math.min(Math.round((scrollTop / scrollHeight) * 100), 100);

}

// Function to calculate the initial viewport percentage

function getInitialViewportPercent() {

    const viewportHeight = window.innerHeight || document.documentElement.clientHeight;

    const scrollHeight = document.documentElement.scrollHeight;

    return Math.min(Math.round((viewportHeight / scrollHeight) * 100), 100);

}

// Function to track page view data

function trackPageViewData() {

    // Set the initial percent viewed on page load

    if (!_satellite.cookie.get('initialPercent')) {

        // Calculate the initial viewport percentage

        const initialPercent = getInitialViewportPercent();

        _satellite.cookie.set('initialPercent',initialPercent);

        _satellite.cookie.set('highestPercent',initialPercent); // initially set with initial percent value, once user scrolls the page it gets updated.

    }

    // Get the current scroll percent

    const currentPercent = getScrollPercent();

    // Update the highest percent viewed

    const storedHighestPercent = parseFloat(_satellite.cookie.get('highestPercent') || 0);

    if (currentPercent > storedHighestPercent) {

       _satellite.cookie.set('highestPercent',currentPercent);

    }

}

// Initialize page view tracking

trackPageViewData();

// Initialize page view tracking

window.addEventListener('scroll', trackPageViewData); // Track on scroll events

window.addEventListener('click', trackPageViewData); // Track on click events

Step 2: Create a data element that reads the initialPercent and highestPercent values from cookies and builds a string:

var initialPercent = _satellite.cookie.get('initialPercent');
var highestPercent = _satellite.cookie.get('highestPercent');
if(initialPercent!= null && initialPercent!= 'undefined' && highestPercent != null && highestPercent != "undefined"){
  return "initialPercent="+initialPercent + " | "+ "highestPercent="+ highestPercent;
}
else{
  return "";
}

Step 3: Map the data element to the desired field in the XDM object in the same page view tracking rule.

Step 4: Build your changes in your development and make sure the eVar is captured with the desired value.

Step 5: Create rule-based classifications to separate initial percent and highest percent values into separate dimensions and then create sub classifications to bucket them if needed.

Considerations for scroll depth tracking

Scroll depth tracking offers valuable insights into how users engage with your content, but there are a few important considerations to be aware of when implementing this feature in Adobe WebSDK.

Scroll data isn’t captured for users who bounce

Currently, our scroll depth plugin does not fire if a user bounces, meaning scroll depth data won’t be captured if a visitor leaves the page without interacting further. This could lead to missed insights, especially on high bounce-rate pages.

Testing before implementing to ensure a smooth rollout

Before implementing the scroll depth tracking plugin in a live environment, it's crucial to conduct thorough testing to ensure it works as expected and doesn't introduce any issues that could affect performance or data accuracy.

Maximizing insights with no additional cost

Scroll depth tracking provides invaluable insights into how users engage with your content and where they drop off, especially for landing pages and key CTA sections. However, it’s important to be mindful of factors like server calls, bounce data, and fast-scrolling behavior when implementing this functionality.

If you are looking to capture real-time scroll data or need help optimizing your scroll depth tracking, we’re happy to help. Contact us for guidance on the best approach for your business and to ensure you’re getting the most out of Adobe Analytics while keeping costs and performance in check.

Santosh would be happy to help! Reach out, get started and ask your questions to: [email protected]

Published by Santosh Kumar in Articles, Insights

Tags: Adobe, getPercentPageViewed, Scroll depth tracking, WEBSDK

August 29, 2024

The role and impact of the AI Assistant

Recently, there have been many discussions and articles on Artificial Intelligence (AI), which is an important topic, now changing how many industries operate. We had the chance to talk to Dana Icikzone, Senior Solution Consultant at Adobe, about this and discuss the release of their new AI Assistant within Adobe Experience Platform.

This blog post is built on our interview with Dana, by reading it you´ll learn more details about the AI Assistant within the Adobe Experience Platform: How different roles can benefit, the general business impact, and perhaps most importantly; the trust and privacy of it. You will also gain valuable perspectives, guidance, and recommendations on how to work with and fully utilize it, from an expert’s point of view.

This is just an introduction. If you are interested in diving deeper into this topic, we are constantly creating blog posts, Mini Masterclasses and content deep diving into the most recent AI prompts, trends and impacts. Follow our LinkedIn page and get noticed when we release something new. But first...

What is the AI Assistant?

To briefly introduce you to the AI Assistant - it is a conversational interface powered by generative AI models. It allows users to ask questions and receive answers based on a combination of base models, custom models, decision-making algorithms and business goals. Embedded within the experience platform, it operates across all applications, including the Real-Time Customer Data Platform (RT-CDP), Adobe Journey Optimizer (AJO), and Customer Journey Analytics (CJA).

The AI Assistant is designed to work seamlessly across various applications within the Adobe Experience Platform. This integration ensures that users can have consistent conversations and obtain relevant answers regardless of the specific application they are using. For instance, an RT-CDP user can still get insights based on CJA data, making workflows more efficient.

Some key technical features

The AI Assistant offers several key features that enhance its utility:

Conversational interface: Allows users to interact naturally and obtain quick answers.

Custom models: Tailored to specific customer needs, ensuring data privacy and relevance.

Role-based access control: Ensures that users can only access data they are authorized to view.

Operational insights: Provides actionable insights based on enterprise data.

Knowledge expansion: Helps users expand their understanding of the platform and their roles.

Verifiable layers: Ensures transparency by providing sources and explanations for all data.

How can the AI assistant benefit my role?

Many different departments and roles benefit from the AI Assistant, including, for example, IT teams, data analysts, and the marketing department. Dana highlighted how it can serve as a companion to developing expertise, managing routine tasks, and providing quick answers to workflow-related questions.

"It should be a companion helpig anyone to become kind of an extended expert," she said.

The IT department

As responsible for ensuring data is collected and being assessable, the IT department can utilize the AI Assistant for data exploration, management, insights, and discovery. For example, in automating routine tasks, thereby freeing up time for more strategic activities. The AI Assistant can answer operational questions, such as how often a segment is used or where a schema field is applied, making data management more efficient.

The data analysts

Data analysts are the professionals ensuring data is thoroughly analyzed and interpreted. They can use the AI Assistant to dive deeper into data sets, perform complex queries, and gain insights quickly. The AI Assistant helps in understanding data structures, troubleshooting specific scenarios, and optimizing workflows. It can also assist in finding and analyzing audiences, making data analysis more streamlined and effective.

The marketing department

Marketing teams ensure data is activated and utilized in strategic decisions, and they also benefit significantly from the AI Assistant. For example, by obtaining quick answers to workflow-related questions and troubleshooting issues. It can aid in campaign creation, audience discovery, and optimizing customer journeys. What the AI Assistant does is help filter out information available within the platform. This can support the decisions you make in terms of creating campaigns, making marketing operations both more efficient and effective.

Prompts and usage

Using the AI Assistant effectively involves specific prompting skills. Users can ask knowledge questions, operational insights questions, and troubleshooting queries. For instance, a digital analyst might ask, "How do I build a segment?" or "What is an identity map?" These prompts help users quickly access necessary information without sifting through extensive documentation.

Another practical example Dana shared is the AI assistant's ability to help find specific audiences within a platform. She explains the problem: "Imagine that you would have to go through every single audience, and there might be thousands of different audiences within a platform". Traditionally, managing your audiences would require a lot of time, creating new audiences, often duplicating existing ones. The AI assistant helps avoid these inefficiencies by providing quick access to the necessary information. For instance, it can help a data analyst quickly identify the most relevant audience for a marketing campaign, saving you a lot of time.

Business impact

The AI Assistant impacts businesses by enabling quick access to enterprise data, facilitating knowledge expansion, and automating tasks. This leads to increased productivity, faster campaign creation, and overall improved efficiency. Dana emphasized how "...being more productive, more efficient and faster as a resource, will enhance your operations." The integration of AI in business processes also helps bridge the gap between different roles, making teams more versatile and efficient. For instance, it can help data analysts understand marketing strategies, and vice versa, leading to more comprehensive and effective campaigns.

Increasing productivity: By automating routine tasks and providing quick access to information, the AI Assistant allows employees to focus on more strategic activities. This leads to increased productivity and faster decision-making processes.

Efficiency improvement: The AI Assistant helps in reducing the time and effort required to perform various tasks. Faster campaign creation and efficient data management contribute to overall easier, efficient and enhanced operations.

Enhancing knowledge and expertise: The AI Assistant aids in expanding the knowledge and expertise of employees by providing quick answers to complex questions. This helps in improving product proficiency and role expansion, making employees more versatile and valuable to the organization.

Trust and privacy

Trust and privacy matters are crucial in the implementation of AI as assistants. Dana emphasized that the AI Assistant on the Adobe Experience Platform is built with these considerations heavily in mind. It uses custom models specific to each customer, ensuring that data is never accessed outside the customer's environment. Role-based access controls further ensure that users can only access data they are permitted to see.

Privacy, security, and governance

The AI Assistant was built with privacy, security, and governance at the forefront. Users must be granted permission to interact with the AI Assistant, and role-based access control policies are strictly honored. This ensures that only authorized personnel can access specific data sets and information.

Customer data protection

The AI Assistant is also designed honoring customer data stewardship. Data is not used or shared across customers, and filters can be leveraged to scrub Personally Identifiable Information (PII). All data provided by the AI Assistant comes with verifiable layers, such as source and explanation, ensuring transparency and trust. Importantly, no third-party data is used to provide answers, which further safeguards customer information.

Dana highlighted, "The AI Assistant uses a combination of models, and one is the custom models that are customer-specific, and those models would never be used, or the data would never be accessed outside of that customer."

Verifiable layers

One of the spotlight features of the AI Assistant is the provision of verifiable layers. Users can always verify where the answer comes from, which is crucial for maintaining trust and accuracy. Dana noted, "There's always a source you can verify, where the answer comes from, which is really important within an AI system."

Future of AI

The future of AI serving as assistants is promising, with potential advancements in automating tasks, generating new segments, and even suggesting optimal strategies based on set goals. Dana believes that as technology evolves, it will continue to drive innovation and efficiency in business operations. "I think generative AI is probably the biggest game changer for Adobe in the past decade and has an incredible potential for customer experience solutions” she explains.

AI Assistants are transforming how businesses operate by providing quick access to data, enhancing productivity, and ensuring trust and privacy. As Dana highlighted, the technology's impact on various roles and business processes will only grow, making it an indispensable tool in the modern business landscape.

Want to know more? Read more about the AI Assistant at Adobe and join the conversation: Adobe AI Assistant

We are constantly creating more blog posts, Mini Masterclasses, and content diving deep into AI. If you don't want to miss out follow our LinkedIn page here -> Accrease (Partner of the year)

Published by Vide Blomdahl in Articles, Insights

Tags: Accrease, Adobe, AdobeExperiencePlattform, AEP, AI, AI Assistant