AI Agents & Industry Workflows
We build AI that does a specific job in a specific industry — not another general-purpose tool for your team to find time to learn. Scoped to one process, integrated with the systems you already run, and measured on your own data before anyone talks about the next one.
Workflow or agent?
The two get used interchangeably and they are not the same thing. Choosing wrongly is the most common reason these projects overrun, so it is worth two minutes.
Workflow
- Triggered by
- A defined event — an email arrives, a record changes, a schedule fires
- Path
- Fixed steps, branches you can enumerate in advance
- Exceptions
- Routed to a human queue with the reason attached
- Predictability
- High — the same input produces the same output
- Best for
- Clear SOPs, high volume, repetitive work
- Cost to land
- Lower. Fastest visible payback
Agent
- Triggered by
- A goal you set; it decides which steps that goal needs
- Path
- Planned per case, calling tools as the situation requires
- Exceptions
- Attempts to resolve first, escalates what it cannot
- Predictability
- Moderate — needs guardrails, evaluation and logging
- Best for
- Messy inputs, judgement calls, exceptions as the norm
- Cost to land
- Higher. Tool integration plus an evaluation harness
What we usually recommend
Use a workflow to absorb the bulk of the volume first — it is cheaper, lands faster and produces the labelled data and the trust you will need later. Then bring in an agent for the long tail of exceptions the workflow keeps handing to a human. Doing it in the other order is how organisations spend a year building something impressive that nobody uses.
Industries we build for
Deep process knowledge in these sectors today, and a delivery method that transfers to any operation with a documented, repetitive bottleneck.
Logistics & Freight Forwarding
Shipping documents read, cross-checked and posted to your forwarding system, with your team involved only where something does not reconcile.
Professional Services
First-pass review and retrieval handled by workflows that cite their sources — with a qualified professional on every output that leaves the firm.
Retail & Brand
Launch content produced per channel and per market, checked against your brand guardrails, and held for approval before anything publishes.
Working in a sector that is not listed? Tell us the process — the assessment is the same in any industry.
How a build runs
Discovery
We sit with the people doing the work and map the process as it actually runs, including the exceptions nobody wrote down. We come back with a shortlist of candidates ranked by volume, error cost and how cleanly they can be automated.
Pilot
One workflow, built end to end and put into production against your live systems. We agree an accuracy target up front and measure against your data, not a benchmark set. If it misses, you see the number.
Deploy
Integration hardening, access control, audit logging and the human escalation paths. Your team is trained on the runbook and owns the operating procedure before we step back.
Operate & extend
Monitoring, drift review and tuning as your documents and edge cases change. Once one workflow is trusted, the next one costs far less than the first.
How you will know it worked
We are a young practice and we do not have a wall of client logos to point at. So instead of asking you to trust results you cannot check, here is exactly how the result gets measured on your data — and how you check it yourself.
The metric is agreed before anything is built
You and we write down what is being measured and what counts as good, in advance. If we cannot agree a measurable definition, that is a signal the process is not ready to automate — and we would rather find that out in week one.
The baseline comes from your records, not ours
The "before" figure is drawn from your own system logs for the period preceding go-live. No industry benchmark, no vendor-supplied comparison, no number we brought with us.
Measurement runs in production, on your data
Live operation over an agreed window, against the documents and cases you actually receive. Not a curated test set, and not a demo environment where the inputs were chosen by us.
You get the raw log, not a summary
Per-item results including every case the workflow got wrong and why it went wrong. You can recompute our headline number yourself, and we would rather you did.
The result is reported either way
Including when it falls short of the target we agreed. A supplier who only reports the wins is not measuring anything — they are selecting. You will see the misses in the same document as the hits.
Start with one workflow
Not a platform rollout and not a strategy deck. One process, in production, measured on your own data — so the decision to do the next one is made on evidence rather than enthusiasm.
What the pilot includes
- Discovery workshop: we map the process as it actually runs, not as the SOP describes it
- One workflow built end to end and integrated with your live systems
- An accuracy baseline measured on your own documents and data, published to you
- Handover documentation and an operating runbook your team owns
- 30 days of tuning after go-live
The stack underneath
These workflows are assembled from engines we build and operate ourselves. That is why a vertical build does not start from zero, and why we can be specific about how each part behaves.
AIGC Engine
Image, copy and voice generation as a component inside a workflow, not a tool someone has to remember to open.
Learn moreAI Digital Humans
Where a process needs a face — guided selling, multilingual video service, livestream hosting.
Learn moreAI Knowledge Base
Retrieval grounded in your own material, with citations and a private deployment option.
Learn moreIT Consultancy & Integration
The unglamorous half: connecting to the systems you actually run, including the ones with no API.
Learn moreAI agents and workflows: common questions
What is the difference between a workflow and an agent?
A workflow follows steps you defined, triggered by a defined event, and hands exceptions to a person. An agent is given a goal and decides its own steps, calling tools as needed, and tries to handle exceptions before escalating. Workflows are cheaper, faster to land and more predictable; agents earn their extra cost only where the inputs are genuinely messy. Most engagements start with a workflow.
Our data and documents are a mess. Do we need to fix that first?
No, and waiting until it is tidy is how these projects never start. We work with the material as it exists. Discovery will tell you honestly where the mess is genuinely blocking and where it simply looks bad, which is a useful distinction to have on record either way.
Our industry is not listed. Can you still help?
Yes. The listed industries are where we have the deepest process knowledge today, not the limit of what we build. The assessment is the same in any sector: map the process as it actually runs, find the highest-volume repetitive step, and prove it on one workflow before anyone commits to more.
How does pricing work beyond the pilot?
The pilot is a fixed scope at a fixed price, from US$6,800, so you can evaluate the result before committing further. Beyond that we quote per workflow or on a retained basis depending on how many processes you want running and who operates them. We will not quote a full programme before the assessment, because that number would be guesswork.
Can this run entirely on our own infrastructure?
Yes. Private deployment is available for the full pipeline, which is the usual requirement in regulated and document-sensitive work. Where a step uses a hosted model, we tell you which step, which provider, and exactly what crosses your boundary — before you sign, not after.
How is this different from subscribing to your AIGC platform?
The platform is self-serve: your team logs in and generates content. This is built and integrated: it triggers off your systems, applies your rules, and produces the output without anyone opening a tool. Plenty of clients run both — the platform for ad-hoc work, workflows for the repeating pipeline.
Bring us the process that hurts most
The assessment is a working session, not a pitch. If the honest answer is that your bottleneck is not worth automating yet, that is the answer you will get.