Use Case 1:
Content Creation

Across industries, marketers are already using AI tools to help their teams create and optimize many types of content, from written text to video and audio content, to imagery and graphics.

We have seen the benefits of AI applications in four key content areas:

1. Text: Leveraging text-based tools like ChatGPT, Bard, and Copy.ai can help create or optimize first drafts, repurpose content for different mediums, or align copy more closely to your brand voice. Consider providing prompts such as “make the following text more concise,” “create a series of social posts from the below article,” or “analyze the writing style from the text below and write a 200-word piece on [topic].”
2. Video: We have seen teams begin to use tools like Synthesia to convert text scripts to avatar videos that can be included in product demo recordings and sales outreach emails. Synthesia and other, similar tools work with multiple languages, too, which has been valuable for companies looking to scale these efforts across a global footprint.
3. Audio: Marketers are using Altered to change the vocal range and timbre of narrators of audio ads, podcasts, and videos, removing the need for multiple voice actors. This is useful for companies with multiple brands, products, and audiences who may respond differently to voice cues.
4. Imagery and graphics: Tools like Midjourney, DALL·E 2, and a new AI module in Canva are being used to create campaign imagery, starting designs, and visual assets for social media and web.

In each of these use cases, it’s important to keep the following considerations in mind as you work to create and enhance content using AI:

Importance of good prompt engineering: Precise prompt engineering is key for generating quality outputs from AI tools. The “garbage in, garbage out” mantra may never be more true than it is in this case. The more details and specificity you can provide in your prompts, the closer the output will be to a finished, usable product that addresses your content needs.

Need for oversight: Beware of model “hallucinations.” Large language models (“LLMs”) and other AI technologies are known to produce inaccurate, fake or otherwise unhelpful content – which has obvious implications for content quality. If published, this can negatively impact the perception of your brand, as well as your SEO and social engagement. We recommend human oversight and intervention – to avoid this common pitfall.

Positive impact of training and iteration: The more iterations of content inputs to generative AI tools, the better their content outputs become over time. As an example, marketing teams are intentionally uploading brand guides and content pieces in their brand voice to train AI tools on their brand.

Evolving legal environment: Companies are also grappling with the intellectual property implications of AI-generated content, which itself is trained on other pre-existing content. (Harvard Business Review published a helpful preliminary perspective on this and other issues in April 2023)

‍

Use Case 2:
Idea Generation

In addition to core content creation applications, marketing teams are using AI tools to generate ideas for campaigns, creative and engagements with customers and other stakeholders. In our portfolio and beyond, we are seeing marketers use AI to support idea generation in several key ways:

Campaign ideation: Prompts that contain industry, product, audience, market trend, season, location and channel information can help your marketing team generate campaign taglines and messaging pillars, and develop new ways to describe value propositions.

Creative brainstorming: Some teams apply generative AI tools to boost brainstorming sessions, generating more creative ideas for several content types: blog titles and body content, social posts, product descriptions, calls-to-action, emails, imagery, and several others. Many marketers intentionally generate several ideas from a single prompt, reducing the number of iterations needed and enabling creative leaders to choose the best fit ideas.

Strategy: AI tools can also be applied to help your team think more holistically about ways to engage with and convert customers. Consider prompts such as: “Give a list of 5 strategies to boost customer engagement for [your company products/services], targeting [your audience] and trying to get them to [take an action]. For each strategy, which channels are most effective for engagement?”

As with content creation, we recommend that marketing leaders and their teams tread carefully in their use of these applications, taking the following into consideration in idea generation:

Timing: Leaders have different perspectives on whether AI should be used to generate creative ideas, or whether it should be used only to refine creative ideas brought by team members.

Confidentiality: Many tools, including the free version of ChatGPT, use inputs as training data sets. We recommend that marketers use caution when considering what types of information they provide to the tools through their prompts and uploads. (Read more on that here.)

‍

Use Case 3:
Research

Marketers leverage AI tools to help research markets, customers, competitors, news, and other subjects relevant for their programs. We have seen this come in handy in three key areas.

Market research: Teams frequently use AI to research markets, using prompts like: “How competitive is the market for [product or service] in [geography] for an audience consisting of [segment]? Who are the primary competitors?”

Customer research: Many marketers are using AI to conduct customer research that can underpin campaign and messaging strategy. Output from prompts such as “What do customers in [segment] care about when it comes to purchasing [product or service]?” can be used to inform a product’s value proposition or campaign messaging.

Best practices research: Marketers also use AI to understand how they can better strategize, execute and measure their campaigns, using prompts such as: “What are the best channels to drive [retention, demand generation, awareness]?” and “What is the best way to measure a [retention, awareness, demand generation] campaign?”

Here, too, we recommend proceeding with caution, taking the time to consider the risks and limitations of AI-supported research. For example, some tools remain unconnected to the web or may falter in certain functions or use cases. Others might provide outdated information based on the most recent data sets they has been trained on, and there have been several cases of LLM tools providing inaccurate answers to relatively simple math problems. Marketing teams should be aware of these common concerns and work to review and validate LLM outputs independently.

‍

Use Case 4: Automation

Marketers are also automating several functions both within their teams and in partnership with other departments.

A few useful examples of automation we have seen from Summit companies include:

Spend and targeting optimization: Leaders are beginning to use tools like Pixis and Ortto to automate ad spend and optimize targeting across more than one platform.

Coding: Marketing teams use ChatGPT to generate, validate and provide coding recommendations for marketing assets. This has been helpful for team members who use regular expressions, conduct analysis using Python, build HTML, and create Apache HTTP server scripts.

Cross-functional automation: Marketers are connecting AI tools to other operational platforms in their stacks. One interesting example we’ve seen involved Monday.com, where a change in status would prompt GPT-3 to auto-generate outreach emails for editing by content writers.

We believe there are several other opportunities for automation that these tools enable across notetaking, customer service operations, and other key functions as well. For marketers who want to explore this use case, we recommend automating for operational efficiency first rather than “automation for automation’s sake.” Work with your teams to understand their biggest operational pain points, map AI capabilities to solve those pain points through automation, and experiment!