Introducing Aesop:
Optimizing Your Knowledge for AI

Aesop is known for his fables – one of the most famous, the ant and the grasshopper, teaches us the value of preparation vs. living life in the moment. While the grasshopper is chasing shiny things and looking for quick fixes, the ant is plodding away getting ready for winter.

New technological breakthroughs often cause companies to act like grasshoppers and disregard the work of the ants. The promise (and hype) of AI has caused many leaders to push the technology on their organizations. But these changes have come so fast that few, if any, large organizations are actually prepared to use it.

It all comes down to data, and no matter how advanced the technology is, the output is only as good as the data you feed in. If Aesop were around today, he would likely share a story about "garbage in equals garbage out", and if he were to use Chat-GPT to write it he might warn against the risks of "garbage in equals gospel out".

Our Approach

We designed Aesop to be a methodology that helps you prepare the knowledge within your organization to be optimized so that AI can assist you in a trusted, reliable manner. Perhaps ironically, Aesop does this by using AI to challenge your current knowledge base and find the areas that will cause various models to fail. In this process we iterate through your knowledge with your experts and produce clean documentation and data that accurately represents your organization as new models are being trained.

For most organizations their knowledge is captured in a complex web of policies, procedures, processes and trainings. In many industries this level of knowledge management is supported by large teams and can span the globe. This, of course, leads to inconsistency and redundancy in documentation; and in many cases our "effective SOPs" have long passed their expiration date. It should be no surprise that an AI model trained on such procedures might provide inaccurate answers, or even worse hallucinate an answer that "sounds really good".

Aesop addresses the root of the problem by focusing on improving the quality and reliability of your data and procedural records before you apply AI. Think of it as the preparation the ant takes to clean up his house and inventory his supplies before stocking up for winter.

How Does Aesop Work?

Aesop is not a tool; it is a methodology supported by technology that helps you create and/or update knowledge documents that are optimized for AI. We do this by following a set of principles:

Start With the End in Mind: Your knowledge documents represent your organization your culture, your priorities, your values - they should not be generated from off the shelf templates. Aesop starts by understanding what is important to you - this can be as easy as a "golden template SOP" or a reimagining of your knowledge management.
Human-in-the-Loop: While Aesop leverages AI to enhance your documentation, we never lose sight of the importance of human oversight. Accountability stays where it belongs-with the Process Owner. Aesop works alongside your team, keeping the human element front and center. This means you still get that crucial context and detailed info that only comes from real conversations with experts.
Bridging the Experience Gap: While documentation is never a substitute for experience, good documentation is an accelerator. Aesop attempts to bridge this gap by expanding the source data from existing documentation and data to include in the views with relevant experts and process owners. The output of those interviews is fed back into the generation of the updated procedures.
Precision and Accuracy Supported by Context: Aesop prioritizes precision and accuracy over brevity. Our goal is to provide clear, reliable documentation that can easily be referenced within context. This often means more verbose documentation to provide the right amount of context so that a reference can stand on its own.
Zero Hallucinations: One of our core goals with Aesop is "zero hallucinations." When the AI doesn't know something, it doesn't guess-it asks for more context or flags the need for additional data. This keeps your processes grounded in reality, preventing costly mistakes.

Aesop Technology Stack

As a methodology supported by technology it follows that we have a technology stack. Given the rapid pace of advancement in LLMs we are constantly swapping out models for ones that help us achieve better results, so our tech stack is modular, and we have several deployment models. We treat each engagement as a separate environment and remove all installations once the output is produced.

Source Documentation Repository: Where you drop the files / data sets that are to be used as inputs. This can remain behind your firewall if needed.
LLMs and Vision Models: This is a menu of models that we select from once we understand the source documents and desired outputs. They are primarily open-sourced models that we deploy locally, but we can also use any enterprise APIs that you might have available.
Aesop Prompting: Our intelligent prompt generation module that we use to interrogate source documentation.
Aesop Interview: Our unique approach to guiding process owners during an interview to fill the gaps in procedural documentation.
Aesop User Interface: A novel way for your experts to review the proposed documentation and accept, reject, or improve responses.
Output Documentation Repository: Where we store the newly generated knowledge document and any additional supporting requirements.

How to Use Aesop and What Does It Cost?

As a methodology Aesop is best used by service providers who understand their client's business. The outputs are knowledge documents that are free from reliance on the Aesop tooling and the execution environment requires maintenance and upkeep; it is logical that a service provider bear this cost so they can achieve economies of scale. To that end expect the pricing to be what you typically pay your service provider for such work.

If your organization has an internal knowledge management team Aesop can be deployed locally, and Zeroth Technology will provide training, maintenance, and licensable versions of the Aesop specific software.

The Aesop Advantage

Aesop is designed to elevate the quality of your existing procedures, documents, and datasets, ensuring that your foundation is strong enough to support advanced technologies like LLMs and intelligent knowledge retrieval systems. Even if use of those systems is not on your horizon, the Aesop advantage makes sense because it delivers:

Referenceable Documentation: Aesop transforms your knowledge documents into reliable, easy-to-use documentation with consistent formatting and concise wording. This makes them not only user-friendly but also ready for integration with future technologies.
Eliminating Contradictions: Consistency across multiple sources is one of the biggest challenges in procedural documentation. Aesop identifies and resolves contradictions, ensuring your records are clear, aligned, and free of conflicting information.
Reducing Repetition: Redundant information clutters your documentation and increases the risk of errors. Aesop streamlines your records by eliminating unnecessary repetition, making your procedures more efficient and easier to follow.
Ensuring Accuracy + Precision: Provides clear and reliable documentation with the right amount of context that can stand alone and stand up to scrutiny.

By addressing these critical aspects of knowledge management, Aesop tackles the "not exciting but incredibly important" foundation that positions your organization for future success. This proactive approach, much like the ant in Aesop's fable, prepares you for the challenges and opportunities ahead.

When you implement AI and other advanced technologies, you'll be building on a strong, reliable foundation of well-organized, consistent, and accurate information. The Aesop method prepares you with the foresight of the ant, enabling your organization to not just weather challenges but to thrive and enjoy its days in the sun.