AI Agents & Workflows
AI that does actual work, not a chatbot in the corner
AI is everywhere, and many businesses across the world are trying to jam it into all software wherever possible. Some of them just shove the phrase AI to their software even if it doesn’t have any AI built in at all. It’s become a marketing gimmick. That’s not what we do. We carefully review the hours your team are actually losing, work out which of them can be automated (by AI or not), and build a solution and workflow that solves your problem. AI may not be the correct tool for the job, but we’ll let you know that in advance.
There is an enormous amount of noise around AI right now, and we’re committed to helping businesses utilise it where possible. We want to develop solutions to problems that businesses have, and if AI helps solve that problem, then that’s brilliant.
Prime use-cases for AI that we see a lot are narrow and specific. Find a specific, high-volume piece of work. Reading emails, extracting the information manually, inserting that information manually into a piece of software your business already uses or relies upon. Something like that is perfect for AI and provides real value for money. It’s quite amazing how much time is freed up on something so simple.
A project we did with Airwaves is a perfect example for this. They’re a facilities management company that grew fast, and the thing holding them back was inbound jobs; almost every job arrived by email, was manually read, inserted into their job management system by hand, and often information was missing and went unnoticed until the job was being picked up by an engineer. You can view the case study below about what we implemented to help them. Engineer information requests dropped by 98%!
What we cover
Finding where AI actually pays
Before anything gets built, working out which parts of your operation are worth automating and which aren't. Usually a handful of processes account for most of the wasted hours. Sometimes the answer is that a simple integration would do it and AI is overkill, which is a cheaper and more reliable outcome for you.
Document & email processing
The highest-value automation for most businesses, because it's where the manual typing lives. Reading inbound emails, invoices, orders, forms, and PDFs, pulling out the information that matters, and putting it into your systems correctly, including the ones with no API worth speaking of.
Tool-using agents
Agents that don't just answer questions, but actually do things; create records, send emails, update your CRM, chase missing information, and hand off to a person when they hit something they shouldn't decide alone. Connected to the systems you already run rather than a separate place to check.
Retrieval over your own knowledge
Answers grounded in your documents, your policies, and your data, rather than whatever the model absorbed from the internet. This is what stops an AI confidently inventing an answer (known as hallucinating), and it's the difference between something your team trusts and something they stop using.
Guardrails & human-in-the-loop
Deciding upfront what the system is allowed to do on its own, what needs a person to approve it, and what it must never touch. Anything irreversible or expensive gets a human in front of it. This gets designed in from the start, not bolted on after something goes wrong.
Evaluation & monitoring
Measuring whether it's actually working, on your real data, with the failures visible rather than buried. Models and vendors change underneath you, so something that worked in March can drift by September. Without monitoring you find out from a customer, which is the expensive way.
How it works
Find the hours
We sit with you and work out where the time genuinely goes; the repetitive, manual, error-prone work that's eating away at your time each week. This is also where we'll tell you which parts AI shouldn't touch.
Build with guardrails
We build the agents, workflows, and integrations, with the limits agreed in advance and a person in the loop wherever the decision warrants one. It goes live with a phased deployment first so you can see it working before it's trusted with everything all at once!
Measure and expand
We check the numbers against what you were doing before, make tweaks, and widen the scope where it's working well.
What's included
- A hunt for where automation actually pays
- Agents wired into the systems you already run
- Guardrails and human sign-off on risky actions
- Answers grounded in your own documents and data
- Measured against real before-and-after numbers
- Start on one process, expand once it's proven
Ways to work with us
No fixed packages, because every business is different - these are the shapes an engagement usually takes. We'll scope the detail, timeline, and cost with you before anything starts.
Discovery
Find where AI actually pays.
- We map where the hours go
- Honest on what AI shouldn't touch
- A costed plan before any build
Build & pilot
Live on one narrow process.
- Guardrails agreed upfront
- Human-in-the-loop where it matters
- Proven on real work first
Ongoing & expand
Measure, tune, widen.
- Monitored against real numbers
- Tuned as models change
- Expanded where it earns its keep
Proof
These are real figures from real clients, not modelled averages or vendor benchmarks. Airwaves' inbound job processing is now fully automated; administrative staff no longer spend their day on data entry, jobs reach engineers faster, and engineers turn up with the information they need instead of ringing the client from the car park.
Read the Airwaves case study →Our work
See all work →Frequently asked questions
What's the difference between an AI agent and a chatbot?+
A chatbot answers questions. An agent should theoretically do real work. The agent we built for Airwaves reads inbound job emails, extracts the details, creates the job record in their system, and emails the client directly to chase anything missing, without anyone asking it to. Nobody sits in front of it typing. That's the distinction that really matters; one is a smarter search box, the other removes a job from someone's day.
Is this just automation with extra steps?+
Sometimes it can be! We'll say when it is and build the plain automation instead. It's cheaper and it fails less. AI earns its place specifically where the input is messy and unpredictable, like free-text emails from a hundred different clients who all format things differently. Traditional automation needs the input to be consistent. Where yours already is, you don't need a model in the middle of it.
Where does AI genuinely make sense, and where doesn't it?+
It makes sense on high-volume, repetitive work where the input varies but the outcome is well-defined; reading documents, triaging inboxes, extracting data, drafting routine replies. It makes far less sense for judgement calls with real consequences, anything needing guaranteed accuracy on every single item, or work you only do occasionally. The occasional stuff isn't really worth automating regardless of the technology.
What happens to our data? Does it leave the business?+
That depends on the design, and it's a decision we'll need to agree upon up front rather than discover afterwards. There are setups where your data never leaves infrastructure you control, and setups using third-party model providers with contractual guarantees about training and retention. We'll lay out the options with the actual trade-offs, including cost and capability, before anything is built. If you're in a regulated sector, we'll work to your constraints, not from what's most convenient to us.
What happens when it gets something wrong?+
It will occasionally. AI is non deterministic, meaning it doesn't always produce the same results. This is precisely why the guardrails get designed first. The system knows what it's allowed to do alone and what needs a person to sign off, and anything irreversible or expensive sits behind a human. When it isn't confident, it escalates instead of guessing. The right question isn't whether it will ever be wrong, but what happens when it is. This will all be designed by ourselves and we'll ensure you're aware of these risks.
Do we need our own AI model?+
Almost certainly not. Training your own model is expensive, time consuming and requires a lot of data. What actually makes the difference is giving a good existing model the right context from your own systems and documents, and wiring it properly into your processes. If anyone is proposing training something bespoke for a standard business automation problem, ask them what it buys you.
How do you know whether it's actually working?+
By measuring it against what you were doing before, and monitoring and analysis the results. That means agreeing upfront what we're counting, whether that's hours, error rates, or turnaround time, and then checking. The Airwaves 98% figure exists because we knew how often engineers were requesting information on site beforehand. Anything running without that measurement in place is asking you to take its usefulness on faith. We use facts and figures to inform our development.
Will it work with the systems we already use?+
Usually, yes, and that's the point; an AI that lives in a separate window is another thing to check rather than a saving. If your systems have APIs, connecting should be straightforward. If they don't, there are normally still work arounds that can be implemented, though we'll be honest about which ones are more fragile and might not be a good fit.
What does this cost, and how soon does it pay back?+
There's a build cost and an ongoing running cost, and you'll see both before you need to commit to anything. What drives the build is how many systems it touches and how messy the input is; the running cost scales with volume and is usually modest against the hours saved. Payback on a well-chosen first automation is typically fast, because we tend to deliberately start with the process wasting the most time. If the numbers don't work, that's a reason not to build it and we'll tell you.
Can you start small rather than automating everything?+
That's how we'd prefer to do it. One narrow process, live and measured, so you can see whether it works on your real data before committing further. It's lower risk for you and it settles the argument with the numbers rather than a demo. Once one is genuinely working and you're happy with it, the case for the next one often makes itself.