AI Refinery: Smarter, faster marketing

The journey of reinvention continues for our marketing function, which is no small feat—especially for a large, global, complex enterprise like Accenture, with nearly 2,000 marketing and communications employees in many locations around the world. And with hundreds of programs launching in any given year, a marketer can spend dozens of hours completing research and analysis needed to make strategic decisions that drive business outcomes.

Enabling gen AI: A strong digital core

Our existing strong, secure digital core that includes a secure technology platform in the cloud, coupled with a strong data foundation, was a key driver of our reinvention and enabled our adoption of gen AI.

We had also implemented our proprietary cloud-based AI and automation-enabled platform—SynOps—to create end-to-end visibility for marketing programs and better governance. The platform provides a single, unified view of all programs that has reduced duplicative content production—saving valuable employee hours and reducing manual steps for managing a campaign by 55%, decreasing operational costs and increasing efficiency. To make it work, we integrated data across disparate systems, standardized it with centralized reporting and KPIs, and created a governance plan. This enabled real-time performance insights across channels and audiences to enhance decision making and campaign optimization—leading to optimal impact in-market.

Next, it was time to accelerate our use of gen AI-powered tools. We initially aimed to enhance our existing workflows using individual AI apps for each marketing task, but quickly realized this approach would not create the enterprise-wide value we sought. We took a step back and reevaluated our strategy: our mission was to reinvent the way we work.

Enter: AI Refinery™.

Refining our intentions

The AI Refinery is built on NVIDIA AI Foundry, NVIDIA AI Enterprise and NVIDIA Omniverse Enterprise and enables businesses to rapidly customize and deploy high-precision AI models and solutions to meet specific needs.     

Using the AI Refinery as an all-encompassing framework, we were able to provide our marketers with all the necessary tools at their fingertips—in one place. AI Refinery converts raw AI technologies and tools into scaled, enterprise AI systems with a faster path to AI benefits and outcomes.

Powered by SynOps and other proprietary data and third-party sources, our agentic solution prioritizes the marketer’s intentions over detailed process documentation. We moved from a deterministic workflow, where the outcome is predictable, to an intention-driven one where the tasks are flexible to meet an end goal. 

This change is expected to further reduce manual steps in a marketing campaign by 25-35%, from 135 to 85. For example, where once the marketers would have needed to contact multiple colleagues with questions and await their responses, they can now expect nearly instant answers from the agents, keeping their momentum. This alleviates the most time-consuming aspects of marketers’ roles and allows them to focus on more creative, strategic and rewarding work.

Accelerating the marketing process

AI Refinery houses 14 AI-powered agents, each of which is modeled after a key skill such as research, analytics, editorial scheduling, critical thinking or forecasting. Our agents enable our marketers to perform activities with increasing efficiency, honing decision making and speeding up campaign work that relies on data gathered from disconnected systems.

For instance, when a marketer starts a campaign, they need to know about competitors’ positioning or marketing strategy on a given topic or offering or capability. Where previously they would have spent upwards of two days manually researching internal and external data sources, the marketer can now ask the Research Agent in AI Refinery to enable them to do the work. Within seconds, marketers receive a concise summary from sources from Accenture proprietary research, our and third-party data such as Moody’s Analytics NewsEdge and the web. A campaign strategy brief used to take up to three weeks to complete; now marketers can produce an initial draft in minutes. This is expected to increase speed-to-market by 25-55% and provide 6% cost savings.

A new form of collaboration

Marketers communicate with the agents through an intuitive interface. For example, a marketer may ask the agents, “When should I launch a thought leadership campaign focused on gen AI in healthcare?” The relevant agents then form a huddle in which they can ‘discuss’ solutions autonomously—in this case, the Research agent would surface information on other planned campaigns targeting the same audience and the Calendar agent would identify their launch dates to inform the marketer’s decision on when to launch. Marketers can watch these conversations unfold in real time and prompt as needed, bringing transparency and giving them control of the outcomes.

Initially, the agents functioned almost like recent university graduates: they had good foundational knowledge but didn’t yet benefit from relevant experience. However, with every interaction and piece of feedback, they learn and improve, delivering increasingly refined outcomes. The result is a cycle of learning between people and AI, whereby people’s use of AI agents teaches them to be better, which makes them more useful to those people.

Marketing revolutionized, with limitless potential

To achieve true enterprise value, we needed to revolutionize our marketing function. Gen AI represents the next wave of workplace transformation, empowering human workers to become the best versions of themselves at work—and with less of the frustration that comes with many manual, tedious tasks. With AI Refinery, the promise of reinvention via agentic architecture is becoming a reality.