Hi, OpenAI team! Chat Norris welcomes you to
The Agile Monkeys experience.
Let us share our latest findings:
Clients are seeking AI agents, and we advise them to begin with the most essential one: search.
Why? Working on Search means focusing on data first, on AI readiness. Every AI operation at the business level relies on querying data. If the data returned is not relevant, incomplete, outdated, or inaccurate, everything built on top falls apart.
Search is also the impactful and measurable low-hanging fruit.
Becoming AI ready:
- 1) Data, data, data...
- - Extract, analyze, combine, enrich, prepare, connect.
- Latency, contradictions and outdated data are our enemies.
- 2) Biological threat lab (A.K.A. Extreme Observability and testing):
- - Everything must be tracked, notated, registered.
- 3) Security:
- - Of course. :)
- 4) Evaluations, evaluations, evaluations:
- "Improvement without evaluation is like karaoke without speakers - technically happening, but no one can tell if it's any good."
"When your system is as organized as Marie Kondo's closet, and your search engine always nails the answer, it's agent time—and the sky's the limit."
Let us brag a bit...
We beat Salesforce’s Einstein Search.
We created an AI search engine for an e-commerce company ("Bimba y Lola"$250MM in ARR) with the goal of allowing semantic searches and more complex business rules. Our competition, Salesforce Einstein, was left behind.
- We created a Feature Extraction Retrieval pattern, combining several models and enrichment workflows.
- Search service diagram
Ingestion service diagram - We built a very cool evaluation system that included automated regression testing, automated evaluations, debugging utilities, LLM explainability and more.
Other cool things we are doing:
- We are contributing to the environment and health, helping law firms fight document dumps (when polluting companies bury law firms in documents as a way to obstruct the process).
- We are developing a growing ecosystem of tools built around a core knowledge base for Harvard University.
- Climateverse, a project that leverages AI to query metadata and take action on the impacts of climate change for risk mitigation.
- We’re going beyond chatbots and designing other AI experiences.
- We are going way beyond RAG:
- - We model search engines as research agents.
- We created a retrieval pattern called Feature Augmented Retrieval. - We are working with USAFacts using AI to help understand the Government’s data.
Our commitment to you as a partner:
Our goal is to become your preferred partner, and we are committed to giving our utmost effort and then some.
We will communicate effectively with you by providing clear and concise insights from the clients we work with, making it easy for you to understand and digest the information. We will take up minimal time from you while delivering substantial value.
🤝
How can you help us?
- Referrals would make it. And having access to the latest updates would be awesome too.
Coffee?
📍
You can enjoy coffee with us in Brooklyn, San Francisco, Boston, Madrid and the Canary Islands.