Return to blog

Company

02
Dec
2024
-
4
min read

Arcee AI, From Small Language Model Pioneer, to Pioneering SLM-Powered Agentic AI Workflows

First, we pioneered small language models (SLMs). Now, we're elevating them to their full potential, leveraging them in our end-to-end, easy-to-use agentic AI workflow platform called Arcee Orchestra. Here's a look at how we got started with SLMs, and how we're now taking them to the next level.

Mark McQuade
,
Brian Benedict
,
Jacob Solawetz
,

The future of enterprise AI is now: with the launch of Arcee Orchestra today, we bring you a no-code platform to build custom AI workflows powered by our state-of-the-art SLMs.

When we founded Arcee AI in 2023, we had one goal in mind: to make world-class small language models (SLMs) available to companies across all industries. This vision was based on our firm belief that only SLMs could offer the domain-specificity, efficiency, scalability, and security that enterprises need for generative AI that can truly boost the productivity, efficiency, and performance of their teams.

We were certain that where large language models (LLMs) fall short, SLMs would thrive–and we’re happy to say that, as we look back at our first full year in business, we’ve been proven right. The market agrees with us: SLMs, not LLMs, are the future of enterprise AI.

Arcee AI co-founders (from left to right) Jacob Solawetz (CTO), Mark McQuade (CEO), and Brian Benedict (CRO).

Thanks in part to our team having pioneered the SLM category, and productized a range of game-changing SLMs, it’s not an exaggeration to say that 2024 became the year of the SLM in the generative AI space. Headlines in both the business media and tech outlets have described the distinct market shift away from LLMs and towards SLMs.

As the initial hype for LLMs turned into frustration with their high costs and ongoing performance issues, the market started to praise SLMs as a solution that brings immediate value at a reasonable price.

Arcee AI… helps companies train and roll out an increasingly popular — and much tinier — approach to AI: small language models. Rather than try to do everything ChatGPT can, Arcee’s software helps accomplish a more limited set of day-to-day corporate tasks — like building a service that only fields tax-related questions, for example — without requiring as much data.Rachel Metz, “In AI, Smaller, Cheaper Models Are Getting Big Attention” in Bloomberg, August 8, 2024

The power–and popularity–of our models resides not just in the fact that they’re smaller than their LLM counterparts. Their high performance and incredible efficiency is the result of our research team’s innovative work building a world-class post-training pipeline that includes state-of-the-art techniques like model merging, Spectrum, and logit distillation.

What’s Next with SLMs: Leveraging Them to Power Agentic AI Workflows

We're thrilled with the success of SLMs and, as passionately as we believe in our vision of SLM-led enterprise AI, market adoption and SLM performance advanced much faster than we could have predicted.

Meanwhile, in the field, we’ve begun to see more and more use cases that Arcee AI better served through an ensemble of SLM models operating on our customer’s internal data and taking action through their internal APIs–which is more than just flashy chat interfaces that enterprise product managers release to check a box for management.

Increasingly, we’re hearing customers and prospects say that they want their AI to be more than a model. Getting answers from language models–even the highly accurate and extremely nuanced answers that Arcee AI SLMs provide–is no longer enough. Going from input to output is no longer enough. Now, companies want their AI to take them from an input to an output that is an actual business outcome or action.

Fortunately for us, this market trend towards AI that can take action, or what’s widely called "AI agents" or "agentic workflows," is in fact the most natural evolution of SLMs. We already build industry-leading accurate and specialized SLMs, and we can now take advantage of the nimbleness of these expert models–empowering our models to work to their full potential by working together, in a platform that also integrates with other business tools.

Today, we’re thrilled to announce the launch of Arcee Orchestra, an enterprise AI solution powered by SLMs that represents the culmination of Arcee AI's progression from SLM training toolset provider to SLM builder to SLM-based agentic system.

Now, instead of just submitting your query to an SLM, you can submit it to Arcee Orchestra–which breaks it down into tasks, routes each task to the most appropriate SLM, then consolidates a response. You get detailed, trustworthy answers fast, but also so much more. Our no-code UI provides templates with which you can create customized workflows to handle multi-step tasks, eliminate busywork, and get better outputs.

Read the full introduction to Arcee Orchestra by Lucas Atkins, the Co-Lead of our research division, Arcee Labs.

And you can learn more about the models that power Arcee Orchestra in this article by our CTO Jacob Solawetz.

Fifteen Months: From Three Founders to a Team of 35 Across Five Continents

As we look back on 2024, the year of the SLM, and head into 2025 with the launch of Arcee Orchestra, we'd like to take a moment to reflect on just how far we have come. Arcee AI formally announced our seed funding in January of this year, which was quickly followed by our Series A led by Emergence Capital in July. We started with just the three of us, and now we’re 35 people spanning five continents. We’re very proud of our team, and we’re also extremely grateful for our many customers–who bet on our vision of an SLM-led future.

How to Get Started with Arcee Orchestra for SLM-powered agentic workflows

You can be among the first to get started with Arcee Orchestra, our new solution that will transform how teams across all industries do business. Available today as a private beta, reach out to get started, and a team member will be in touch with you. In the meantime, you’re welcome to try the models that power Arcee Orchestra right now by signing up on our Model Engine!

Give Arcee a Try

Lorem ipsum dolor sit amet consectetur. Vitae enim libero lectus urna blandit sapien. In egestas ac dolor dictum.
Book a Demo

Sign up for the Arcee AI newsletter

Subscribe to get the latest news and insights on SLM-powered AI agents

Thank you!

We will get back
to you soon.
Oops! Something went wrong while submitting the form.