Akkio, the leading AI data platform for advertising agencies, has launched Agency Data LLM (AD LLM), the first Large Language Model (LLM) tuned to advertising analytics.
Akkio’s Agency Data LLM enables agency teams to effortlessly chat with campaign data to build audiences, uncover actionable insights, and create charts and dashboards – resulting in better client outcomes. AD LMM is the only fully customized AI experience fine-tuned for advertising data, context and tone.
“Agencies have tons of data, but tapping that data for answers is hard. We built this model to solve that problem,” said Akkio co-founder and CEO Jon Reilly. “Now anyone in an agency can get value from their data quickly and reliably, simply by asking it questions. It’s a step change in efficiency.”
When benchmarked against GPT-4, AD LLM gave a more accurate and informative answer to data insight questions 80% of the time. Furthermore, AD LLM responded 13.7 times faster than ChatGPT, showcasing its efficiency in high-demand scenarios.
The AD LLM system uses Akkio’s Generative Knowledge Graph, which maps the relationships in data and identifies and notes issues with the data where applicable. AD LLM is more performant than other LLM data tools due to its proprietary Retrieval-Augmented Generation (RAG) AI architecture. This innovative framework integrates advanced language models, vector databases, prompt engineering and sophisticated query parsing, allowing for fast, dependable responses, even across multiple tables.
Unlike general-purpose AI chatbots, Akkio’s platform was trained to learn common analyst and media team queries used to build and monitor campaigns. AD LLM connects directly to popular ad data sources, including Google Ads and Meta, and the most popular data warehouses such as Snowflake and BigQuery. It’s easy to then share answers into a dashboard that can be used internally for collaboration or externally with clients for reporting.
Large agencies can deploy Akkio directly in their infrastructure, so the data stays in their control. When deployed within an agency’s environment Akkio can take advantage of the variety of compute options available to optimize query speed. For instance, teams using the Snowflake Data Cloud love that Akkio can leverage Snowpark to query huge data quickly.
Example use cases for AD LLM include:
- Audience building. Prompt: “Create an audience segment of suburban high-income customers most likely to purchase.”
- Campaign optimization. Prompt: “Analyze the channels driving ROAS for the Q2 campaign.”
- Performance reporting. Prompt: “Build a visualization comparing CPC and conversion rates across ad groups.”
SOURCE: BusinessWire