Your Path to Agentic AI Starts Here.

Architect is a small Leeds based software consultancy specialising in AI & Machine Learning

Get on a call
Powered by
The World of AI
Lets Clarify Some Terms.

We operate across all things AI, from traditional machine learning through to the cutting edge and agentic AI, ensuring that our solution is embedded in your software.
‍

Generative AI (Gen AI)

Generative AI takes text, image, audio input and follows instructions to generate an output based on the input.

Works Well For:‍

  • Text and Document Q&A
  • Image Q&A
  • Chatbots
  • Enhancing user-software natural language interactions

Common Myths: ‍

  • It costs too much
  • Its not secure or GDPR compliant
  • Its hard to control

Machine Learning

Machine Learning (including Deep Learning) uses data from inside and outside of your business to train models that can make predictions.

Works Well For:‍

  • Reccommendation
  • Optimisation
  • Predictive Analytics
  • Probabilistic Modelling
  • Image Classification & Object Detection
  • Time Series forecasting

Common Myths:‍

  • We need lots of high quality data first
  • Its hard to explain
  • Its hard to maintain

Agentic AI

Giving Gen AI models the ability to take actions & make decisions.
‍
Works Well For:‍

  • Conversational Agents (Co-Agents): These are advanced chatbots that are deeply embedded with the UI and can take actions on behalf of a user.
  • Back Office Automation: Even robust software comes with edge cases that often require manual intervention. Task Agents can connect to internal APIs to replace back office processes such as Quality Assurance and Data Inputting.

Common Myths:

  • Agents cant be monitored easily
  • User facing agents will cost too much
  • They are too complex and not worth the hassle
Work
Featured Projects.

Our work combines cutting-edge AI technology with a human-centered design philosophy.

We design AI solutions that blend innovation with user-centered functionality, creating impactful and enjoyable experiences while delivering measurable results.

Car Finance 247

FinTech Broker based in the United Kingdom.
‍

  • Introduced Cost Effective Dual Channel Call Transcription for a 300 strong sales floor
  • Built a complex customer value model for optimising Google Value Based Bidding
  • Replaced expensive SaaS softwares with internal WebApps
  • Built a customer facing agentic chatbot, with a RAG vector database

Bayonette

Telecommunications Company based in Oslo, Norway.
‍

  • Develop a lightweight, real-time toxicity filter for resource-constrained devices using Rust.
  • Trained and fine-tuned a model using Jigsaw datasets and custom examples.

QuickAI

Startup based in San Francisco, USA
‍

  • Developed machine learning models in Rust for operating inside the browser (WASM)
  • Optimised the data processing for browser use
Offerings

How we can Help.


We operate from first principles in all areas of Artificial Intelligence.  

About Us
AI  Education

Get your organisation in the best shape for taking advantage of the latest AI tools, with workshops & integration support on topics like:
‍
AI Development Environments (Cursor and Github Copilot)
AI Assisted Pull Request Reviews
Effective employee use of LLMs

Agentic Ai

Whether its developing Co-Agent Chatbots, internal task based agents for automating back office processes (QA), we have experience developing agentic AI bespoke to your use case.

Generative AI

We have experience with Open and Closed source LLMs, Retrieval Augmented Generation (RAG), Document Q&A, LLM Ops & monitoring, Prompt Optimisation and Validation

Machine Learning

Speak to our Machine Learning Specialists, who can build bespoke models based on your data, for your use cases.

- Exploratory Data Analysis: We start by exploring your data, giving us and you a detailed analysis of what is possible.
- Model Development: Develop models & track their performance against key metrics & performance baselines
- Machine Learning Engineering: Deploy the models either on-device (edge), on your infrastructure, or in the cloud.

Delivery

Our Process.

In general we prefer to follow this process in order to have sufficient time to identify how we can aid you best. It might be that we can't help you right now, and thats ok!

1. Introductory Call
Get to meet the Team, understand how we operate and start building trust.
2.  Discovery Sprint: AI Project Identification
For the first few weeks of our engagement, we identify and prioritise opportunities for AI projects that align with your priorities.
3. Discovery Sprint: Stakeholder Interviews
Depending on the application, we will then conduct user interviews
4. Discovery Sprint: Proof Of Concept
Yes, even machine learning projects need a proof of concept. This is the minimal amount of work we can do in order to demonstrate value or gain feedback on the implementation.
5. Minimum Viable Product
If you are happy with the proof of concept, we can discuss pricing and timescales for turning a PoC into a full product.
6. Development & Maintainance
We wont just dissapear, if you need ongoing development and maintainance support then we will always be available to continue work on previous projects as required.

Testimonial

"Great work. Thanks! We will definitely be working together in the future"

Rohit Patel
Director, Generative AI (Meta) and Founder of QuickAI

Testimonial

"Great work!"

Avre Paalsrud
CEO Bayonette
Connect

Book a Call with Us.

Privacy Policy

Crafted by the Architect Design Team