Broadbent AI in production

Where it runs today

Our own brands, our clients’ websites and our apps all use the same Broadbent AI service, running on our own infrastructure — so the assistants we build for you are the ones we rely on every day.

  • Brand websites

    An assistant on every Broadbent site

    Visitors ask questions in plain English and get answers grounded in that site’s own content, with live pricing where it matters.

    • broadbent.group, broadbent.host, broadbent.cloud and broadbent.network
    • sftpapp.com, broadbent.digital and broadbent.ai
    • help.broadbent.group and community.broadbent.group
  • Client websites

    Assistants and catalogue automation

    The same service runs on websites we build and host for clients — answering customers and keeping large product catalogues organised.

    • Customer assistants that answer from each client’s products and policies
    • Automatic product categorisation for e-commerce catalogues
    • Generated breadcrumbs and category paths for SEO
  • Apps

    Built into sFTP Client 4.0

    Broadbent AI sets things up, writes commands and explains what a server is telling you, inside the desktop app for Windows, macOS and Linux.

    • AI Assistant for connections, monitors and schedules
    • Terminal command helper and “Is this server healthy?”
    • Passwords and keys are removed on the device before anything is sent
One service
Every site and app calls the same Broadbent AI API
Grounded answers
Each assistant gets its own site knowledge and pricing
Live replies
Answers stream in as they are written
Secrets stay local
Apps strip passwords and keys before sending

Applied AI

Practice

Assistants when you need an agent in the loop, scrapers when the knowledge has to stay live, and full systems when the work spans prototype through production.

  • Assistants

    AI assistants & agents

    Task-focused AI that understands your workflows, integrated into products and internal tools.

    • Natural language control
    • Product & ops copilots
    • Safe, reviewable actions
  • Knowledge

    Scrapers & knowledge systems

    Intelligent scrapers and retrieval that keep your AI in sync with the data your business actually uses.

    • Live data pipelines
    • Retrieval over your corpus
    • Kept current, not frozen
  • Solutions

    End-to-end AI systems

    From prototype to production — architecture, modelling, infrastructure, and operating AI at scale.

    • Architecture & stack
    • Implementation
    • Monitoring & iteration

How we work

From idea to a running AI system.

We meet you where you are — early exploration or scaling production. Each engagement has a path, so you know what happens next.

Request a quote
01 Discovery
Goals, data, constraints
02 Prototype
MVP on real data
03 Iterate
Quality against usage
04 Production
Harden, monitor, hand off

GPU servers

Lease Dell servers with NVIDIA Tesla GPUs.

Reliable compute for training, fine-tuning and inference, hosted in Colchester and sized for AI workloads — not a shared cloud GPU lottery.

  • Hardware

    Dedicated Dell + Tesla

    Dedicated machines fitted with NVIDIA Tesla GPUs, operated by the same UK team as the rest of the stack.

  • Workloads

    Train, tune, infer

    Sized for training, fine-tuning and large-scale inference, with predictable pricing and managed operations.

  • Connect

    Your stack, our facility

    Plug into your existing cloud or on-prem tools. Capacity and sizing through the Colchester team.

Operations

Make existing systems smarter — support, back office and product features — without throwing away the tools you already run.

  • Support

    Support & operations

    Copilots for support teams, documentation-aware chat, and triage automation.

  • Process

    Internal processes

    Smart workflows for approvals, compliance checks and repetitive back-office tasks.

  • Product

    Product features

    Embed AI directly into your SaaS, apps or customer-facing experiences.

Contact

Speak to the team

Call our Colchester sales team, or send a message anytime.

UK sales team · 8:00am–8:00pm