Custom AI Software Development
Bespoke AI Software for Specific Business Problems
Quoting takes days when it should take hours. Years of ERP data sit unused. Staff retype information between systems. Critical workflow knowledge lives in one person's head. If any of that sounds familiar, bespoke AI software is usually the answer.
We do not just consult on AI. We build and run production AI products ourselves. Reactable AI is the Deep Purple team's own agent product, funded by Enterprise Ireland's Pre-Seed Start Fund. The original version has been retired; the new version, rebuilt as an agent, launches September 2026. The site is live at reactable.ai.
Deep Purple builds custom AI software for established businesses across Ireland, the UK and Europe. Not off-the-shelf platforms. Not chatbots. Purpose-built systems designed around your workflows, your data, and your specific operational challenges.
We work with established businesses and department-level teams in manufacturing, engineering, food production, construction, professional services, logistics, and medtech, with real processes that are currently manual, repetitive, or disconnected. For businesses looking for custom AI software development, this usually means building a system around an existing workflow rather than buying another generic tool.
Key facts
- Based in
- Longford, Ireland. Serving Ireland, the UK, the EU, and the US.
- Client size range
- 15 to 13,000+ employees. Function-level engagements at larger organisations.
- Typical project size
- €25,000 to €250,000. Most land between €40,000 and €100,000.
- Typical timeline
- 10 to 20 weeks
- Grant support
- Discovery typically 80% grant-funded. Build projects up to 50% for eligible businesses.
- IP ownership
- Yours. Handed over at completion.
- Hosting
- Ireland or the EU on our infrastructure, or deployed inside your own environment
- Delivery
- Senior-led team.
Problems Our Clients Typically Come to Us With
These are the problems that come up most often. There are plenty of others. If yours isn't here, it probably rhymes with one of them.
Quoting and estimating is eating days it shouldn't
Historical job data, pricing logic, and supplier rates rarely live in one place. We build AI pricing engines that learn from your past quotes and generate draft estimates in minutes. See how we built this for a building services company.
Years of ERP or operational data you can't get value from
You know the data is useful. Nobody has the time or tools to do anything with it. We build predictive models and integration layers that turn sitting data into decisions your team can act on. One example: a predictive quality model for food manufacturing.
Staff retyping information between systems
PDFs to spreadsheets. Spreadsheets to the ERP. ERP to the invoicing tool. We build document processing and workflow automation that reads, validates, and moves information between systems without the manual copying.
Critical workflow knowledge lives in one person's head
When they're out, things go wrong. When they retire, the business loses years of operational understanding. We build knowledge capture and decision-support systems that encode how the work actually gets done. The quoting engine mentioned above replaced exactly this kind of single-person bottleneck.
You can't hire enough skilled people to keep up
The labour market is not providing the staff you need. AI and automation do the work that headcount won't cover, so your senior people spend their time on what actually needs a person.
You tried off-the-shelf AI tools and they didn't touch the real problem
Copilot, ChatGPT, or Zapier helped with surface tasks but couldn't engage with your operational data. We build custom AI trained on your data and your process, not a generic model.
You paid for an AI report once and nothing got built
Most consultancies write reports. We write specifications and then build what they describe. See how our process works.
If any of this sounds like you, Discovery is the right next step. If your situation is a different shape, tell us anyway. We've probably built something close.
What Custom AI Software Looks Like
Every system we deliver is different because every business problem is different. Recent projects include:
AI-powered quoting systems that learn from historical pricing data and generate estimates in minutes instead of hours
Document processing and compliance automation that extracts, cross-references, and flags data across specifications and regulatory standards
Predictive analytics and machine learning models that forecast outcomes, detect anomalies, and optimise processes using your historical operational data
Conversational AI on your proprietary data — systems that let your team query databases, documents, and knowledge bases in plain English, built on your data, not public models
Agentic AI and autonomous workflows — intelligent systems that execute multi-step business processes independently, with human oversight where it matters
Computer vision applications for field measurement, quality inspection, and progress tracking
ERP integration layers that add AI decision-making to your existing business systems (Sage, Odoo, or custom ERPs)
Custom dashboards and reporting tools that surface insights from operational data your team is already generating
We do not always build from scratch. Often, we build AI integration layers that sit on top of your existing systems and make them smarter, without replacing the tools your team already uses.
Need predictive models or statistical analysis before building a full system? Our Data Analytics service delivers the analytical foundation — feature importance, predictive models, and prototype tools — that can then be productionised into custom software.
All systems are built on secure, GDPR-compliant infrastructure hosted in Ireland or the EU, with full audit trails and human-in-the-loop controls where decisions matter. Your data stays yours. We never use client data to train public AI models.
AI agents, in plain English
People use the words 'AI agent' to mean a lot of different things. A useful working definition: an AI agent is a system that takes in data, makes a decision with some confidence, produces an output, and hands back to a person when the answer is uncertain or the stakes are high. The agent is not a chatbot. It is the engine behind a workflow.
By that definition, the AI quoting engine we built for a building services company is an agent. It reads historical job data, produces a draft quote with a confidence score, and routes anything uncertain to a human reviewer. That is a working production AI agent, not a science project.
We also build and run our own AI agent product. Reactable AI is the Deep Purple team's own agent product, funded by Enterprise Ireland's Pre-Seed Start Fund. The original version has been retired; the new version, rebuilt as an agent, launches September 2026. When we say we build AI agents for our clients, we mean it in the practical sense: systems that do real work, with real data, in real operational settings. For the full picture of how we design, build and run production agents, see our AI agents page.
What a Good First Project Looks Like
The best first AI project for most businesses shares a few characteristics:
- One specific workflow with clear volume (not "everything at once")
- Existing data, even if it is messy
- A measurable outcome (time saved, errors reduced, capacity gained)
- A human review step where it matters
This keeps scope manageable, costs predictable, and gives you a working system to evaluate before committing to larger projects.
The Business Case: Build vs Hire
For most of our clients, the decision is not "should we use AI?" — it is "should we hire more people or build a system?" Here is how a typical AI project compares to the alternative.
| Hire 2 Staff | Build an AI System | |
|---|---|---|
| Year 1 cost | €80,000 – €120,000 | €25,000 – €60,000 (once) |
| Year 2+ cost | Same again (recurring) | Hosting + minor updates only |
| Time to impact | 3–6 months (recruit + train) | 10–20 weeks |
| Scales with demand | Hire more people | Handles more volume automatically |
| Grant eligible | No | Yes — up to 50% funded |
| Consistency | Varies by person | Same quality every time |
This is not about replacing people. It is about freeing your team to do the work that actually requires human judgement, while the system handles the repetitive, pattern-based tasks that slow them down.
AI Integration With Your Existing Systems
Most of the systems we build do not replace what you have already paid for. They sit on top of it and make it smarter.
The business systems we integrate with include:
- Sage, Odoo, and SAP, for pulling historical transaction data and writing back updated records
- Microsoft Dynamics and Salesforce, for reading customer records and automating opportunities and quotes
- Xero and other accounting platforms, for pulling invoice and cost data into pricing and margin models
- Custom ERPs and proprietary systems, integrating via API, database view, or scheduled extract depending on what the system supports
- Project management tools, document management systems, and operational platforms, as needed for the workflow
We work in three patterns depending on the integration:
API layer. When your existing system has a stable API, we sit the AI model on top of it. Clean, reversible, and no changes to your underlying data.
Middleware. When there is no usable API, we build a middleware service that reads from the source system, processes with AI, and writes results to wherever they are needed. Good for legacy systems that are not going away.
Event-driven sync. When the AI needs to respond to changes in near-real-time, we build event triggers that fire the AI process when the trigger event happens. Useful for new quotes, purchase orders, inspections, or any data that updates continuously.
The AI quoting engine we built for a building services company is a clear example. Five years of historical job data from their ERP, processed with an AI pricing model, and the draft quote written back into their existing workflow. No rip-and-replace.
How We Build Custom AI Systems
Every project follows a structured process from initial conversation to production deployment. The goal is to reduce risk at every stage — you never commit to a large build without evidence it will work.
Discovery and Scoping
Every project starts with a Discovery phase. We visit your business, understand your processes, and scope the build before you commit to implementation. Discovery costs €5,000 as a standalone engagement. If you qualify for Enterprise Ireland's Digital Discovery programme, that drops to €1,000 out of pocket. Discovery produces a detailed specification with architecture, integrations, timeline, and budget — not a vague proposal.
Build and Iteration
We build iteratively, not in a black box. You see working software early and often. Milestones are structured so you can evaluate progress at each stage before committing to the next. If something needs to change mid-project, we adapt — the specification is a starting point, not a contract written in stone.
Integration
Most of the systems we build connect to existing tools — Sage, Odoo, SAP, custom ERPs, project management platforms, document management systems. We build AI layers that sit on top of what you already use. Where a hybrid approach is faster or lower risk — combining custom components with existing tools or automation platforms — we will recommend that instead of building everything from scratch.
Testing and Deployment
Every system goes through structured testing before it reaches your team. We deploy to a staging environment first, run your real scenarios through it, and fix issues before go-live. User training is included — we do not hand over a system and walk away.
Pricing and Timeline
Timeline. Typical projects run from 10 to 20 weeks from Discovery handoff to production deployment. Simpler projects with one integration and one core AI function can be faster. More complex multi-system builds take longer. Discovery itself is a separate 2 to 4 week engagement that produces the specification the build is priced against.
Pricing. Typical projects run from €25,000 for a small build up to €250,000 for a large multi-system build. Most land in the €40,000 to €100,000 range. Larger programmes are typically delivered in phases and can exceed this range over time through sustained engagement. Function-level engagements at larger organisations can scale further depending on complexity and integration scope. We give you a firm number at the end of Discovery, based on the detailed specification, not a rough estimate from a sales call.
Grants. Discovery engagements are typically 80% grant-funded through Enterprise Ireland's Digital Discovery programme. Your out-of-pocket cost for Discovery is €1,000 (ex-VAT). Build projects are typically up to 50% grant-funded for eligible businesses. We help clients structure applications and align scope with grant criteria. The process, the team, and the deliverables are the same whether you use a grant or not.
All prices ex-VAT.
Who is delivering. Senior engineers only. The same people who design the architecture also write the code. Direct access to the team building your system.
Your Data, Your Control
Every system we build runs on private cloud infrastructure hosted in Ireland or the EU. Your business data is never used to train public AI models, never shared with third parties, and never leaves the jurisdiction you agreed to. This is contractually protected.
We build with data sovereignty as a design principle, not an afterthought. Access controls, audit trails, encryption at rest and in transit, and GDPR compliance are built into the architecture from day one — not bolted on at the end.
If the engagement ends, your data is exported, handed over, and securely deleted. You own everything we build for you.
You own the IP
Every line of code we write for you belongs to you. No vendor lock-in. No licensing fees. No dependency on Deep Purple to keep your system running. We build on your infrastructure, hand over everything, and document it so your team — or any future developer — can maintain and extend it.
What Happens After Launch
Launching the system is not the end of the project. We provide a structured handover that includes full technical documentation, user training for your team, and a support period to address anything that emerges in the first weeks of live operation.
After handover, you have three options:
Run it independently
The system is yours, the code is yours, and the documentation is written so any competent developer can maintain it.
Retained advisory
Ongoing access to our team for monitoring, performance reviews, architecture guidance, and iterative improvements. Typically €1,500 to €3,000 per month.
Come back for the next project
Most clients who start with one AI system end up building a second within 12 months, usually targeting a different workflow.
We build systems that are designed to outlast the engagement. No proprietary frameworks. No vendor dependency. No recurring licence fees to Deep Purple.
When Custom AI Software Is Not the Right Answer
Custom AI software is not always the right fit. Sometimes the honest answer is not to build. Here are the most common situations where we tell a prospect they should not spend money on a custom build.
- Off-the-shelf already solves 80% or more of the problem. If a standard tool covers most of what you need, the marginal benefit of a custom build rarely justifies the cost.
- The transaction volume is too low to pay back the build. Custom AI is a one-off cost paying down a recurring inefficiency. If the inefficiency is too small, the maths does not work.
- The data does not exist, or is not captured reliably. AI systems learn from data. If the data is not being recorded, or is recorded inconsistently, the right first step is fixing the data capture, not building the AI on top.
- The process changes every few months. Custom AI encodes a process. If the underlying process is genuinely unstable, the build will be out of date by the time it ships. Automate the stable parts, leave the unstable parts to people.
- The real problem is organisational, not technical. If the blocker is accountability, communication, or decision rights, a software system will not fix it. In those cases we say so and walk away.
We tell you this during Discovery. We would rather lose the build than build the wrong thing.
Recent Projects
Four recent examples of what we have built. Different industries, different problems, same approach: working software in production, measurable outcomes, senior delivery team.
AI-powered quoting engine
A building services contractor was quoting by hand, with complex quotes taking 1-2 hours of senior time and 24-48 hour turnaround. We built an AI system that reads five years of historical job data and drafts complex quotes in 15-20 minutes. Same-day turnaround, with human review before anything is sent.
AI Discovery for a 150-person engineering firm
A 150-person international engineering firm wanted to identify where AI could genuinely help across their operations. The Discovery engagement produced a prioritised roadmap with estimated effort, cost, and return for each opportunity. Enterprise Ireland funded.
Predictive quality model
A food manufacturer wanted to understand which of 150+ production variables actually drive product quality. We delivered a predictive ML model, feature importance rankings, and a working prototype tool. Delivered as a four-week Data Analytics engagement that can be productionised into custom software.
Computer vision for field measurement
A specialist construction contractor operating across Ireland, the UK and Europe was losing revenue to measurement disputes and paper dockets. We built a computer vision app that calculates area from photographs with ±3% accuracy and updates a real-time dashboard across all sites. Enterprise Ireland funded.
Not sure if custom software is the right approach?
An AI consulting assessment will evaluate your options — custom software, workflow automation, off-the-shelf tools, or a hybrid approach — before you commit to a build.
Learn more about AI ConsultingFrequently Asked Questions
Start With a Conversation
Not sure if custom AI software is the right approach? Start with a Discovery assessment. €1,000 out of pocket with grant funding, or €5,000 self-funded. No obligation to proceed.
Book a 20-minute intro callWant to understand the process first? See how we work →
About Barry Gough
CTO, Deep Purple AI Consulting
Barry Gough is the CTO of Deep Purple AI Consulting. With an MSc in Computer Science from University College Dublin, where machine learning was a core focus of his studies, and over 20 years building production software systems, Barry brings formal ML training and deep hands-on engineering experience to every AI and data analytics engagement.
Barry completed his masters at UCD in 2011, studying ML algorithms, statistical modelling and data-driven systems just as big data techniques were maturing and deep learning was about to transform the industry. Barry joined Purpledecks in 2016 and has led the technical delivery of enterprise projects incorporating machine learning, computer vision, data classification, predictive features and recommendation engines for commercial clients across Ireland and the UK.
In 2023, Barry architected and built Reactable AI from the ground up as an internal Deep Purple product, a system that generated and optimised marketing campaigns across channels against measured results. Reactable AI ran in production as an autonomous agent system, requiring him to design systems where AI made real decisions with real consequences.
At Deep Purple, Barry leads all technical delivery: AI system architecture, machine learning model development, data pipeline engineering, and manages a team of experienced ML engineers and applied statisticians. His combination of formal ML education, years of incorporating AI into commercial projects and hands-on experience architecting autonomous AI systems means clients work with a technical lead who can make genuine engineering decisions about AI.
Deep Purple AI Consulting (deeppurple.ai) is an AI consultancy and custom software development company based in Longford, serving businesses across Ireland, the UK, the EU, and the US. We help established businesses identify where AI can make a real difference, then build the systems to make it happen. Senior-led delivery. Grant-funded where possible. No hype.