Why We Exist
The launch of OpenAI's ChatGPT in 2022 introduced a new way of interacting with software. As LLM technology has progressed over the last few years, the use cases this AI can be applied to have expanded rapidly. We believe that the future of work will be transformed by AI, and the way we build software today, automate workflows, start our days and live our lives will be fundamentally different in the coming years.
These benefits are not guarenteed to be evenly distributed however. Success requires adapting and adapting requires knowledge. In fact we have observed fully leveraging the potential of AI requires a deeper understanding of business domain problems compared to the traditional focus on implementation and expertise in a specific product or technology like Excel, Python or SalesForce. As this technology continues to evolve our mindset and approach to work must change as well.
Our goal is to ensure the most people can benefit from AI to live better lives and see their ideas come to life sooner by adapting to this new way of working.
Distribute AI Benefits More Evenly
As AI continues to advance, we believe (among others) that those who adapt effectively will gain the most. We aim to help more teams adapt effectively so advantages of advanced intelligence compound beyond a narrow set of incumbents. This is why we focus on products and workflows that make it easier for everyone to go from Idea to Product using AI. This entails both building products that help teams work with AI faster and understand AI system better, as well as helping teams learn how to develop their own workflows around AI.
We want everyone to be able to leverage AI to its fullest potential, not just a select few. By lowering the activation energy for rigorous evaluation, we enable teams to move faster with confidence instead of trading one for the other.
High Quality Personal Relationships
Even before AI, doing business with people\brand\ideals you have a connection with is where you would prefer to spend your time and money. In the Era of AI, utility and software options become comodities allowing us to focus on the human element of our work. We realize that and will go above and beyond to build high quality relationships.
Amazing User Experience
If AI enables everyone to build software faster and cheaper, then choosing products based on utility and function becomes less important. Similar to relationships, we believe UI\UX design will be critical to user adoption. Users will have many choices and ability to use products they enjoy. We strive to build those products.
Software Design Should Be Ambient Art
As we hope is visible on this website, we believe software should look and feel good to use. We should harken back to the days of Renaissance architecture and design, where the beauty of the product was as important as the utility. We strive to build software that is beautiful ambient art.
What We Do
We help teams go from idea to working AI capability faster by pairing evaluation‑first build loops with tight prompt / context engineering and calm, trustworthy UX. We battle‑test patterns inside real delivery (model benchmarking, instruction refinement, context shaping, cost & latency instrumentation) then turn the reusable parts into products like SparkEval and internal workflow tooling. Consulting engagements and tools reinforce each other—every client build hardens the playbooks the next team starts with.
LLM Performance Evaluation
Many project with AI start with picking a model, but many dont test out which one they should be using or how their prompts change the results. Our expertise and tools help us and our clients do that effectively.
Optimize Instruction Prompts
There is an increasing push towards letting AI systems determine or optimize their own system prompts, but hands on testing is still needed to evaluate the behavior of specific instructions
Context Engineering Workflows
Using our internal workflow process and our own products, we help customers optimize their workflows with AI tools to build higher quality software products faster.
Products
SparkEval (multi‑model chat + scoring) and other products in the works for AI development workflows enable us to deliver higher quality work to clients.
Consulting
Using our products and tested workflows, we are able to do work for and educate clients on ways to improve their own workflows and feel more independant and competant in the age of AI.
Where We Are Going
Identify and Build a suite of tools that improve the process of working with and understanding AI models and improving our ways of educating on and distributing those tools to customers.
Model Profiling
Our model evaluation app, SparkEval, as well as other tools we have planned will increase users understanding of models' behavior and performance, enabling better decision making.
Context Engineering
We have plans for products, documents and reports that will make context and prompt engineering easy for teams to manage and share with colaborators, increasing productivity.
Playbooks
With so many ai tools and prompts out their, we will develop curated advice on workflows and tools to achieve the highest quality results with minimal guessing.
Research
We have plans to use our product suite to develop datasets and reports that will help inform the community about AI models performance and capabilities.
Agents and MCP Servers
As our products and workflows mature we will work towards agentic tools to let users take an even farther step back if they choose too which will enable projects to complete even faster.
Community
Growing a community around MPL tools will be critical to improving our products and connecting people who can work with similar products towards similar goals.
improve our flywheel by extending and improving our offerings
Our Team
Robby Boney
Builder & workflow optimizer focused on collapsing distance between idea, validated behavior & production reliability. Background spans enterprise data, automation & AI integration — repurposed now into compression playbooks.
Current obsession: lowering the activation energy for rigorous evaluation so teams can move faster with confidence instead of trading one for the other.