Ally Financial Inc. announced the early results of an experimental generative AI use case with its marketing team designed to help understand the real-world impact the technology could have on campaign development and employee productivity. Using the Ally.ai platform’s large language model (LLM) chat and prompt functionality, a select group of marketers were able to reduce the time needed to produce creative campaigns and content by up to 2-3 weeks and reported an average time savings of 34%, compared to typical processes without AI.
Additional results reported from the month-long experimental program include:
- The reduction in time needed for content creation and other marketing outputs of up to 2-3 weeks was primarily found in the early stages of the creative process in tasks such as research, first drafts and naming exercises
- The 34% time savings on tasks when using the Ally.ai interface could equate to an annual savings of almost 3,000 hours across the participant team
- More than 80 prompts developed by participants during the test had an 87% usefulness rate, showing the value of prompt engineering training for teammates and the effectiveness of outputs that could be used in assigned work
- Users also noted a prompt accuracy rate of 81%, signifying that they received the right type of content or response outputs from the Ally.ai platform to complete their assignments
- Participants also saw usefulness and accuracy in outputs when using Ally.ai for data analysis, quality controls and search engine optimization (SEO)
“The best marketing teams know the art of evolution, which includes evaluating and testing of transformative technologies,” said Andrea Brimmer, Chief Marketing and Public Relations Officer at Ally. “The early results demonstrate how AI could enable our talent to focus on creative tasks that needed the most human involvement, while speeding up the routine efforts of our day-to-day work. We look forward to the continued exploration of generative AI and learning more about how the technology can assist our team.”
Ally.ai, the company’s proprietary, cloud-based AI platform that serves as a secure, private bridge between enterprise-grade LLMs and Ally’s data and AI applications, launched in June 2023. The test with marketing comes on the heels of Ally’s pilot with its Customer Care and Experience group where Ally.ai was used to quickly and accurately summarize customer service calls, helping associates focus their energy on more meaningful customer interactions and reduce the time needed to service and close a customer inquiry.
Both use cases reflect a commitment to Ally’s core principles for the enterprise use of AI: initially focusing on use cases that support employee productivity and optimizing internal business processes; providing human intervention and controls for oversight and training; protecting against the disclosure of personal identifying information and prohibiting third party model providers from using Ally data to train their foundational models.
“The purpose of building Ally.ai in-house, and with foundational data, cloud and network infrastructure already in place, is to scale its use throughout the organization while also keeping a close eye on potential risks and opportunities for new use cases,” said Sathish Muthukrishnan, Chief Information, Data and Digital Officer at Ally. “The value that our marketing colleagues could get from Ally.ai demonstrates how important cross-functional collaboration is to the development of enterprise AI capabilities.”
Generative AI in Marketing: Supporting Productivity
In addition to the preset system prompts and security features already embedded in Ally.ai, the platform needed a place where teammates could engage with the LLM directly. Understanding that most LLMs and generative AI applications are good for summarization, content creation, ideation/brainstorming, and information analysis, Ally’s developers built an open prompt user interface (UI) for Ally.ai that would allow teammates to do two types of work: chat directly with the secured LLM and also develop specialized prompts that could be saved, used again, and shared with colleagues based on their accuracy and usefulness rates.
Teammates could change features of the platform depending on the work they did, including toggling between chat and prompt mode, adjusting prompt length as well as the model’s creative parameters and vocabulary. The participant team frequently used Ally.ai as a starting point for creative development, specifically for first drafts of advertising copy, video scripts or social media posts, or enhancing human-driven ideas that originated from brainstorms or naming exercises.