ServicesCustom AI systems
An AI development company in London, building AI software you own
AI that does a whole job at volume, with the accuracy agreed in writing first.
- Fixed price
- £8,000 to £25,000
- Time
- Two to four weeks to a working system
SolvesPaperwork, screening and replies that grow with every customer
The short answer
Augustova is an AI development company in London. Our AI development services build custom AI software for UK businesses: systems that read documents, score applications, answer calls and act inside your own tools, priced at £8,000 to £25,000 as a fixed price and delivered in two to four weeks. You own the code and the provider accounts from day one.
We design, build and deploy AI systems around one specific business process: reading documents, screening and scoring applications, answering and making calls, and acting inside the CRM, inbox and calendar you already run. They run in production on your own accounts, with the cost of every model call measured and capped.
Fixed price, agreed in writing before anything is built. Every engagement is quoted as a fixed price with a written scope, including what is not included.
The full guide: AI automation agency in LondonWhat you get
Included
- A working system in production, not a prototype on a laptop
- The repository, the provider accounts and the credentials, in your name
- Cost instrumentation per request, per feature and per model, with enforced ceilings
- An accuracy threshold agreed in writing and measured on a sample of your real inputs
- A runbook, written documentation and a recorded walkthrough
- A training session for the people who will use it daily
- A written support arrangement that starts at go live rather than being negotiated afterwards
Not included
- Training a model from scratch when an existing one does the job for a fraction of the cost
- Data cleaning at a scale that should be its own project, which we scope and price separately
- A guarantee of accuracy we have not measured on your own data first
How it runs
- 01
Paid discovery: written scope, acceptance threshold and fixed price
- 02
The core loop working on your real inputs in week one
- 03
Integration with your systems, with a demo every week
- 04
Testing on real inputs, with spending ceilings triggered in production
- 05
Handover, or monitoring and support if you want it
CoversDocument reading and extraction · Screening and scoring with written reasons · Voice and chat agents · Knowledge assistants over your own documents · Agents that act inside your systems · Cost instrumentation and spending ceilings
Signs it is the right call
- Somebody spends hours a week reading, sorting or retyping information that arrives in a predictable shape
- A decision gets made the same way every time by a person who has to be available for it to happen
- Work arrives in the evening and at the weekend and waits until somebody is back
- You have been quoted for an AI project and cannot tell whether the number is reasonable
Where projects like this go wrong
The demo that never becomes a product
A model that works on ten hand picked examples usually fails on the eleventh. We test against a sample of your real inputs, including the messy ones, before agreeing that a system is finished.
Cost that scales faster than usage
AI spend fails quietly, because nothing breaks until the invoice arrives. Cost is attributed per request and per feature from the first day, and spending ceilings are triggered deliberately in production before anyone relies on them.
Accuracy nobody agreed on
Good enough has to be a number written down before the build starts. We agree the acceptance threshold and how it will be measured, so finished is a fact rather than an opinion.
What AI software development covers
An AI development company builds software in which a model does part of the work a person used to do, and builds everything around the model that makes it safe to rely on: the integration with your systems, the rules about what it may decide alone, the record of what it did, and the meter on what it costs. The model is usually the smallest part of the job.
These are the systems we build, each one a kind we have built and run in our own products before quoting it to anyone else.
- 01
Document intelligence: invoices, CVs, referrals, contracts and forms read as they arrive and turned into structured records, with the exceptions sent to a person.
- 02
Screening and scoring: every application or enquiry read against your criteria, scored, and the score explained in writing so a person can check it.
- 03
Voice and chat agents: calls answered and made, applicants and customers screened or booked into a real calendar, with a text fallback designed before the voice.
- 04
Knowledge assistants: answers drawn only from your own documents and systems, with the source shown and a hand-off to a person when the answer is not there.
- 05
Agents that act: an agent that updates the CRM, files the document, books the slot or chases the payment inside the tools you already use, proposing anything risky for a person to approve.
- 06
AI inside a product: features in a SaaS product or app, attributed to a customer and a feature from the first request so the price per seat still makes sense at scale.
How AI software development runs, from scope to production
A paid discovery step comes first, usually the audit above or a shorter scoping workshop when the problem is already well defined. It produces the written scope, the acceptance threshold as a number, a sample of your real inputs including the messy ones, and the fixed price. Nothing is built before that document exists, because the scope conversation is what decides whether the project succeeds.
The build runs in weekly demos against your real data. Week one usually ends with the core loop working end to end on a narrow slice, so the shape is visible early. Integration with the systems you already run comes next, because that is where most of the value and most of the risk sits. Cost instrumentation goes in from the first request, attributed to a customer, a feature and a model, and the spending ceilings are triggered deliberately in production before you rely on them. Go-live is followed by handover: the repository, the provider accounts and the credentials in your name, a runbook, documentation and a recorded walkthrough.
What AI development costs in the UK, and what we charge
Published UK prices for AI development cluster in three places. A first implementation at a consultancy runs from about £12,000 to £40,000, a departmental automation build £5,000 to £20,000, and a multi-system programme £20,000 to £50,000 and up, with several larger firms setting a minimum project of £25,000. Day-rate engagements cost more again once the days are counted.
Ours are fixed: £8,000 to £25,000 for a custom AI system, and £4,000 to £15,000 for a single agent or an AI feature added to software you already run. The difference is mostly scope discipline and experience: we have built each kind of system before, so fewer weeks go on discovering how.
| What you are buying | Typical UK market | Augustova, fixed |
|---|---|---|
| A single agent or AI feature, in production | £5,000 to £20,000 | £4,000 to £15,000 |
| A first custom AI system | £12,000 to £40,000 | £8,000 to £25,000 |
| Several agents sharing data and hand-offs | £20,000 to £50,000+ | £8,000 to £25,000 |
| Minimum project | Often £25,000 at larger firms | None |
Market figures are the published UK ranges surveyed in September 2026 in our consultancy and automation cost guides; they are context, not a quote. All prices exclude VAT.
What a custom AI system costs to run once it is live
The build price is fixed. The running cost scales with usage, and it is the number most quotes leave out. From operating our own AI products, three things hold. The model is rarely the expensive part; a text message or a minute of telephony costs more than the model behind it. The small model does most of the work once the workflow is designed properly, with the frontier model reserved for the few steps that need it, which makes a well-built system several times cheaper than the per-token headline suggests. And a spending ceiling that has never fired is an assumption, not a control.
Every system we build reports cost per feature and per customer from day one, so the running cost is a number you watch rather than an invoice you discover. Our running-cost guides give worked examples at current prices.
Custom, off the shelf, or a workflow tool
Custom is the right answer when the work has to happen unattended, at volume, inside your own systems: reading every application that arrives, answering the phone at 9pm, chasing every invoice, reconciling three systems that were never meant to talk. It is the wrong answer when a paid ChatGPT or Claude workspace would do, or when an off-the-shelf product already matches the workflow closely enough. Between the two sit the workflow tools, n8n, Make and Zapier, which are right for a single process with modest volume and wrong once reliability, cost at scale or data handling matter.
We say which of the three applies in the discovery step and we quote only for the one that fits. Half the value of a paid discovery is being told not to build.
| Workflow tool | Off the shelf | Custom AI system | |
|---|---|---|---|
| Best for | One process, modest volume | A standard workflow a product already covers | Unattended work at volume inside your systems |
| Typical UK cost | Subscription plus setup | Per-seat licence | £8,000 to £25,000 fixed, plus running cost |
| Who owns it | The platform | The vendor | You: code, accounts, credentials |
| Where it breaks | Reliability and data handling at scale | The 20 per cent of your process it does not cover | Scope that was never written down |
AI software we have built and run
Onrolo, our recruitment platform, reads and scores every CV against the role with the reasoning written down, contacts applicants by message and voice within minutes at any hour, and books interviews into the recruiter’s own calendar. The published outcome is more than 90 hours of manual CV screening removed per office every month. GoRizzume, our careers platform, scores a CV against a specific advert and rewrites the weak sections, and runs spoken interview practice with a live AI voice on the web and in its iOS and Android apps.
For a London driving school we built an AI voice assistant that answers in the school’s name, quotes the real prices, checks the caller’s area and books into the real diary, and hands reschedules, complaints and refunds to a person. It must pass 174 scripted test calls before any change goes live. Behind all of them sits the cost governance described in our case study: every model call attributed to a feature and a customer, with ceilings that have been watched firing.
Which models we build on, and why it matters less than you think
We choose the model per feature, on your real inputs, against the acceptance threshold agreed in the scope: Anthropic, OpenAI and Google models, and specialist platforms for voice. The choice matters less than the design around it. A workflow that sends the routine steps to a small, cheap model and reserves the largest model for the few decisions that need it is usually faster, cheaper and just as accurate, and in our own products moving one high-volume feature to the small model made it several times cheaper with no drop in success.
Where your data has to stay in the UK or the EU, the models are called through cloud regions that keep it there, with storage and logs alongside, and the position is written down so a client questionnaire is one line. Models are called through business accounts whose terms exclude training on your data.
Data protection and ownership
A written data processing agreement is signed before any personal data moves, and Augustova is registered with the Information Commissioner under reference ZC152144. Each system states which data it processes, through which provider, in which region and on what lawful basis. Anywhere money, a person’s rights or regulated data are involved, the system proposes and a person approves, and every action is logged with its reasoning.
The code, the repositories and the provider accounts are in your name from day one. If you stop working with us, the system keeps running and another developer can pick it up from the handover pack.
AI development company, AI consultancy or AI automation agency?
The labels overlap and buyers use them loosely. An AI consultancy decides what is worth building and whether to build anything; an AI automation agency builds agents and workflows around processes a business already runs; an AI development company builds AI software, which may be an internal system or a product you sell. Augustova does all three, with the same senior team, at fixed prices, so the choice is about where you start rather than who you hire.
Start with the audit if you do not yet know which process is worth it. Start with a build if you can describe the job in a paragraph and the systems it has to touch. If you are comparing firms, our comparison of twelve London AI consultancies and studios names the ones that suit a large programme better than we do.
Common questions
How much does a custom AI system cost in the UK?
Augustova prices custom AI builds between £8,000 and £25,000 as a fixed price, agreed in writing before any work starts, with an explicit list of what is not included. A single agent or AI feature added to software you already run is £4,000 to £15,000. Larger platform work is quoted per phase so the budget cannot drift mid project.
What do your AI development services include?
AI software development from a written scope to a system in production: discovery, architecture, the build in weekly demos against your real data, integration with your existing systems, cost instrumentation with spending ceilings, testing, deployment and a handover pack. Where the product needs one, the web and mobile apps around the AI are built too.
What does an AI development company actually build?
Software that does a specific job a person currently does: reading invoices, CVs or referrals into structured records, scoring applications with written reasons, answering and making calls, answering questions from your own documents, and acting inside your CRM, inbox and calendar. We have built and run each of these in our own products.
How long does it take to build a custom AI system?
Two to four weeks from agreed scope to a working system in production. Most delays on projects like these are decisions waiting on somebody rather than engineering time, which is why the scope and the acceptance criteria are agreed in writing before anything is built.
Which AI models do you build on?
Whichever does the job best on your real inputs at the lowest running cost, chosen per feature from Anthropic, OpenAI and Google models, with specialist platforms for voice. The small model does most of the work once the workflow is designed properly, which keeps the running cost down.
What will it cost to run each month?
Stated in the quote, per feature, at your volume, and capped in production. The model is rarely the expensive part; messages, call minutes and platform fees usually cost more. We pay the same per-message and per-minute prices in our own products.
Can you add AI to software we already have?
Yes. Adding an AI feature to an existing product or system is usually an implementation project at £4,000 to £15,000: integrated with what you run, tested on your real inputs, and instrumented for cost per customer and per feature.
Who owns the code and the accounts?
You do, from day one. The code, the repositories and the provider accounts are yours, and at go live you receive documentation, credentials, a runbook and a recorded walkthrough so another developer could pick the work up.
What if AI is not the right answer to our problem?
We say so. Plenty of problems are better solved with ordinary software, a process change, or nothing at all. The discovery step is paid and its output is yours to keep, so an honest no still leaves you with a written scope you can take anywhere.
- 5.0on Google, every review five stars
- Certified
- Certified
- Certified
- for Startups
Not sure yet? Talk it through.
Tell us what happens today and roughly when you need it changed. A senior member of the team replies personally, within 24 hours, or book a free 30-minute consultation and talk it through.