Agency’s Custom AI Model Fuses Advertising with Behavioral Science
Signal Theory, a leading independent advertising firm known for leveraging behavioral science, announces the launch of AL, an innovative AI “employee” assistant that enhances the agency’s internal processes and select client projects. The name “AL” is a tribute to one of Signal Theory’s founders, Al Higdon. Developed entirely in-house, AL showcases Signal Theory’s technological capabilities, commitment to safe and ethical use of AI and dedication to remaining human-centered.
“Right now, there are questions all clients should be asking their agency partners, like, ‘What are you doing to protect my brand’s private data from public AI systems?’” says Elijah Kleinsmith, Signal Theory’s technology and innovation director. “We built AL because clients expect us to keep their data safe, even as we incorporate new technology into our work.”
Over the past year, Kleinsmith says AL has successfully supported various Signal Theory disciplines and contributed to select client projects. He analyzes the internal use of AL’s chat and automation functions weekly, noting that the tool’s use has doubled every month since its launch. Nearly 80% of Signal Theory employees are using AL, and nearly a third say AL has significantly impacted their work, improving efficiency, generating actionable insights and driving better outcomes.
A key feature Kleinsmith has seen growth in is the use of AL’s research capabilities. The tool has helped Signal Theory staff generate more than 80 cited research reports. Another is the incorporation of behavioral science principles. AL has been designed with a human-centered approach, integrating into existing workflows and channels to better assist and respond to the needs of Signal Theory’s various account teams and clients. AL also provides actionable behavioral science insights with each response.
“It’s easy to see the potential efficiencies and effectiveness of AI, but to truly be useful, you have to ensure your data is safe.”
The firm says some clients were initially concerned about AI-related issues like data security, liability and copyright, and the team took those worries seriously in the development of AL.
“It’s easy to see the potential efficiencies and effectiveness of AI, but to truly be useful, you have to ensure your data is safe,” says Tony Robinson, CFO of Signal Theory. “AL has been designed to keep inputs confidential and secure.”
Signal Theory has built trust one client at a time through open communication about the agency’s robust AI policy, thoughtful integration practices, ongoing staff training and regular updates.
“The creation of AL as our own proprietary tool has been met with very positive feedback.”
“When you’re crafting communication strategies for new product or service launches, or managing customer and sales data, confidentiality and security are critical,” says Diane Young, vice president of account management. “We’ve been very transparent with our clients about how and where AI fits into the workstream at Signal Theory. The creation of AL as our own proprietary tool has been met with very positive feedback.”
Kleinsmith says AL is always learning: AL has read more than 10 million words and written more than a million. “We will continue to innovate and focus on improving AL’s performance and usefulness,” says Kleinsmith. “The model is now twice as fast as when we introduced it. We are currently testing new capabilities, such as generating synthetic survey data.”
Four Ways Signal Theory’s Custom AI Supports Employees, Clients to Keep AI Secure
- Behavioral Science Integration: AL is built on a foundation of human behavior research applied to advertising. It shares psychologically informed recommendations and strategies with each output, ensuring the focus remains on the human experience.
- Workflow Enhancement: By automating routine tasks and offering intelligent suggestions, AL streamlines processes and frees up the Signal Theory team to focus on higher-level strategic work.
- Client Support: AL’s capabilities extend to assisting with client projects, offering data-driven insights and creative solutions that are grounded in behavioral science principles.
- Secure Infrastructure: Because AL is a closed system, none of the data processed is stored or used in public AI models.
“We’re eager to introduce AL to the world,” says John January, Co-CEO of Signal Theory. “We made a conscious decision to incorporate AI into our workflow to streamline our internal operations and deliver real value for our clients. However, since it is a bespoke system, we also had the opportunity to teach AL all about behavioral science and how to apply that to our work. We think this sets AL apart.”
