Today we announced that our LAKEer agent was the most trustworthy agentic enterprise search solution as measured by the FACTS benchmarks.
It took a considerable amount of engineering effort by Sajjad, Nafisa, and others on the DaaX technical team to run the FACTS benchmarks. Now that we know how to run them, it'll be faster the next time, but the first run was very painful!
Why We Undertook This Benchmark Project
Because enterprise customers told us they need benchmarks!
We heard these needs from them:
- Need trustworthy agentic enterprise search,
- Need benchmarks to compare and contrast solutions,
- Their data is distributed, it's everywhere, and
- Need new business models, the old enterprise business models don't work for them for AI.
Proving "Most Trustworthy" With Benchmarks
Our neuro-symbolic AI technology was designed to deliver the most trustworthy answers. And that's where the customers' need #2 (benchmarks) comes in. When we said, "Most trustworthy", they said "Oh yeah? Prove it" (they said it more politely, but their message came through loud and clear).
Their point was that there are lot of bold claims by suppliers in AI nowadays (DaaX included!). It's very hard for customers to sift through these claims and determine which solution meets their needs. They want benchmarks to help them.
We agreed with them – suppliers should be confident enough in their solutions to publish benchmarks.
The Result: Industry-Leading 77.7
Fortunately, Sajjad, our CTO, was aware of the Google FACTS benchmarks and believed they could be used to measure LAKEer. And so off we went, benchmarking away LAKEer.
I can tell you when our team finished benchmarking, with the result we achieved a first-place result of 77.7, we all did a dance, had a party!
Benchmark results are temporal; someone may overtake us someday for first place. We are going to keep innovating and if we do get passed, we're just going to out innovate them and pass them right back!
Addressing the Other Customer Needs
Customers' need #3 was distributed data. Our Federated Search vision address this need.
Customers' need #4 was a new business model for AI solutions. We have a blog for it, read it here.
What's Next
We're not done yet benchmarking! Sign-up for our newsletter and be informed of additional benchmark results soon.