n marketing automation: 6 use cases that can benefit from data quality

Note from the Editor: This is part 2 of a 4-part series about how AI will be integrated into marketing automation platforms. Here is Part 1 of a four-part series on AI marketing automation.

The AI hype in 2023 has been centered on generative AI (copy, image, and video) use cases. Although some still doubt generative AI’s impact, the widespread adoption of content-focused capabilities shows that the focus is justified.

There’s a more fundamental movement underway: the integration of AI into each marketing technology application.

Martech leaders will see an increase in accuracy and productivity by incorporating AI into their core stack components, such as CRM and marketing automation platforms. I have prioritized data management within that scope. Most marketing operations leaders recognize this as the foundation.

Data management: the first (semi-) natural language processing

Data management was the first “natural language” change to fuel martech growth before the AI inflection. How? , the no-code transformation, gave us the ability to create new databases fields. This was a privilege that had previously been reserved for IT. It was possible to integrate internal and external fields into landing pages and website. This transformed digital engagement.

We still rely on humans to input data, even with automation. Training was still a barrier to adoption for (properly) entering data, despite the easier-to-use software. Early AI algorithms affected various data cleaning processes whendata were entered incorrectly or incompletely. We all knew that it was best to avoid inaccurate data entering the system as this would lead to erroneous downstream results.

I’ll use a common framework — garbage in, garbage out (GIGO) — to illustrate.

‘Garbage in’

1. Entering data

Martech leaders cringe if users complain that entering data is difficult. Empathy is due, especially if the interface has changed over time. If you are a Salesforce user, but still use Classic over the newer version of Salesforce, then this is your empathy reminder. Lightning is your reminder to be empathic!

including Salesforce has recently predicted that the generative AI “prompts” revolution would forever change the user interface. Each UI must now process natural language to reduce the friction for users to enter information.

As an example, ChatSpot‘s (HubSpot AI interface) user interface leverages the GPT Model. While I am vendor-agnostic I have used the tool to extract examples and it is available for testing in their public beta release.

Let’s begin with the basics – adding a new Contact.

The users won’t need to remember which part of HubSpot’s interface they click to “Add Contact”; instead, they can use a prompt that looks like this:

HubSpot’s alpha version has been running for three months. It also includes prompt templates, which trigger actions based upon common tasks. You can choose from a list of favorites like this.

2. Add data and research about individuals and companies

Many MAPs pull in basic information about customers from websites. AI simplifies this task. Now, a summary of key profiles that can be used to supplement contact personas and company firmographic information is just a click away. As an example:

3. Spreadsheets with Infused data

According to MarTech’s Salary and Career Study, approximately 70% of marketers work on spreadsheets more than 10 hours per week. They are the foundation of martech stacks.

In my March 2023 MarTech conference presentation, I discussed how these tools (and formulas, VLOOKUP abilities, etc.) are still our secret decoders for working across multiple data sources. In my presentation at the MarTech 2023 conference, I spoke about how these tools (and their formulas, VLOOKUP capabilities, etc.) are still our secret decoders to work across multiple data sources. In many large teams, this is supported by a full-time analyst. In smaller teams, a marketer who is Excel-savvy and has a good understanding of data often supports these efforts.

Programming VLOOKUP can be a bit too complex for some. Now, marketers are using AI-generated prompts to generate formulas. AI plug-in utilities allow AI-created prompts to be infused directly into spreadsheets.

The most popular and powerful additions will be these “no-code”, natural language capabilities. These capabilities will be integrated directly into the foundational tools for knowledge work (e.g. Google Workspace Labs, Microsoft Co-pilot). The AI assistant will extract domains and first/last name from email addresses. It can also pull up companies, company names, etc.

‘Garbage out’

Now let’s look at the other end of the spectrum.

4. Natural language interfaces for Analytics

We’ve been there. Someone asks you to download a report into Google Slides or PowerPoint rather than using the platform. The ability to get the report directly from the application using natural language prompts is a game changer.

“Can you provide me with a report on the basis of

” will be an prompt that lowers barriers for more people who want to access analytics directly.

In time, users will provide more quality entries if they are encouraged to enter data and to see the results properly reflected. Rather than fix the chart, users may decide to fix the problem at its source.

5. Visualization capabilities

The ability to create visualizations will be included in the capabilities. Plug-ins/interfaces will allow us to direct the platforms in order to create these visualizations.

As many others, I am eagerly awaiting access to OpenAI’s code interpreter. In the interim, I’ve followed others piloting the technology, including Ethan Mollick who gave a sneak preview of the capabilities in the One Useful Thing Newsletter — excerpted from his recent newsletter post.

6. Accessible big data

These data entry and output advantages are not limited to only the data that is the “source of truth” in CRM/MAP.

AI-based prompts will also be able to access other data attributes and augmenting data.

Despite the need for governance and training, blind trust cannot be avoided.

Martech leaders should be cautious not to rely solely on AI for data quality and management. A more robust governance is needed, given the infancy of the AI generative tools and the potential for them to negatively impact data quality without supervision.

Data management is a challenge that has a double impact. The prompts might not inherit the guidelines your organization has for associating accounts with contacts; you may have to develop more advanced prompts following those guidelines.

Anyone who imports data in a spreadsheet performs a sanity-check after applying formulas. Typos can cause issues in thousands of records. Faulty AI logic can cause thousands of records to be corrupted if users did not create the correct prompt.

What’s next? In Part 3 of the series, I will dive into the AI integration into the MAP campaigns processes.

The article 6 Marketing Automation Use Cases Where AI Can Help with Data Quality first appeared on MarTech.