What our AI implementation services cover
We start from a defined business outcome, not a list of tools, then pick the technology that fits the job. These are the six kinds of work we deploy.
- An AI agent for one task. One clear job, clear access, review points and a named owner.
- Business workflows. Approvals, reporting, customer follow-up, document processing and hand-offs between systems.
- Microsoft Copilot rollout. Practical setup across Word, Excel, Outlook, Teams, SharePoint and Power Automate.
- Power Automate and integrations. Triggers, connectors, approvals and notifications that move information between approved services.
- ChatGPT for teams. Shared ways of working, reusable instructions and access to the knowledge the team needs.
- Developer workflows. GitHub Copilot, Codex and Claude Code set up for repeatable coding tasks, internal tools, documentation and testing.
When an assistant needs to stay available and use approved tools through a gateway you control, we also set up OpenClaw agents.
We build into what you already use. Rip-and-replace is rarely needed.
How AI implementation for SMEs works with us
There are five stages, and you see progress at each one.
1. Audit, or start from your plan
Not sure which job to fix first? Our AI audit costs a fixed £750. It reviews your workflows and gives you ranked opportunities, the risks and a written plan. The plan is yours to keep, whoever builds it. If you already know what you want built, we skip this and scope from your plan.
2. Design and scope
We agree the first useful outcome, the estimated delivery days and any access or licences the build needs. On a Microsoft 365 or Power Platform build, we agree permissions, exceptions and human review points before rollout. Before work starts, the written quote sets out what is included, the timetable and how payment works.
3. Build and test
We configure and connect the system in your environment. Then we test it with the people who will run it. Testing covers normal work, the awkward edge cases and the points where the system should stop and hand back to a person.
4. Train and hand over
Your team gets documentation and hands-on training to use, check, pause and maintain what we built. We use examples that match your normal work, not a generic list of prompts. For a wider rollout, we start with a pilot group, review the questions and adoption, then agree the next stage. For training beyond the build, see our AI training courses.
5. Support
The support window after launch is agreed in your scope. If you would rather not maintain the system yourself, our managed AI service monitors, fixes and improves the agreed workflows.
Typical first builds, by team
Each of these can be tested with real work before you commit to more. The right column stays with a person.
| Team | A sensible first build | What a person still decides |
|---|---|---|
| Sales | Lead follow-ups and CRM updates after an enquiry | Anything that commits you to a price or a date |
| Finance | Invoice processing that flags anomalies | Payments, and every flagged invoice |
| Marketing | Content drafts and a posting schedule across platforms | Final copy before it is published |
| Operations | Reporting, inventory alerts and internal updates | Figures that feed a decision |
| Whole business | Microsoft Copilot rollout with a pilot group | Who can see which files |
| Developers | GitHub Copilot, Codex or Claude Code workflows | Code before it is merged |
Already have an AI setup that is not delivering? We can start there instead.
Examples of our work
Explore our case studies for examples of training, AI implementation and software builds. When choosing an implementation partner, ask to see work relevant to your systems and the type of handover you need. Work in progress is not evidence of a completed result, and a training session is different from a production integration.
AI implementation in the UK and Ireland: how pricing works
- First call: free, 15 minutes, no obligation.
- AI audit: £750, fixed price.
- Build and automation: quoted. The written scope sets out estimated delivery days and covers configuration, integration, testing, documentation and handover.
- Team and Copilot training: quoted on team size, format and duration.
- Managed AI: from £250 a month, plus setup from £1,750.
Software subscriptions, platform fees, hosting and licences are extra. You pay those directly or as agreed in the quote. Prices are in sterling, and your quote confirms the currency, VAT and payment terms, including for businesses in the Republic of Ireland.
We do not promise grants. Invest NI lists working with a private sector provider as not eligible for its Innovation Vouchers, so a voucher cannot pay for our work. Full detail is on our prices page.
What makes an AI implementation stick
These are the rules we build to.
One job, one owner. Every build starts with a single job and the person in your business who owns it. Broad access waits until there is a reason for it.
Permissions come before prompts. Microsoft 365 Copilot follows a user’s existing access permissions. Overshared files can therefore remain accessible through Copilot, making permissions review part of a sensible rollout. We look at sharing and permissions before a Copilot rollout. Where a tenant change is needed, we work with your IT provider rather than change controls without the right owner.
Agreed review points. Payments, sensitive records, mass messages and decisions about customers or staff get explicit limits and a human check. We agree those points before rollout.
An easy way to pause. Your team knows how to pause the system and remove a connection.
Data protection checked early. The ICO lists the novel application of existing technologies, including AI, as a factor that can mean a data protection impact assessment is required. If a workflow handles personal data, we flag it during scoping so your data protection lead can decide whether one is needed. Our AI policy template sets ground rules for staff.
Your AI implementation partner: who does the work
On our team, these are the two roles closest to a build:
- Davy Anthony, AI Implementation Consultant, brings enterprise workflow and process improvement experience. He implements and automates AI systems that fit your existing tools.
- Jess Mason, AI Expert, builds and teaches AI assistants and leads architecture and integration work.
Phil Patterson, founder and director, has helped teams adopt AI since 2022. Priyanka Kasture, AI Specialist, helps teams find useful AI opportunities and fit solutions into their existing workflows.
Blue Canvas is run from Derry by Blue Canvas AI Ltd, company number NI737502. On-site sessions are agreed before work starts.
We are independent. Naming a tool on this page does not mean its maker endorses or partners with us. Meet the team.
Bring us the plan, or the problem
Tell us what you want built, which systems are involved and who owns the work. In 15 minutes we will tell you whether it is a sensible first build and how we would scope it.
Plan the rollout and the way back
A successful test is a release decision, not permission to connect every system at once. Before launch, record the tested version, access granted, known limits and the person who approves it. Start with a small group or volume and compare the result with the existing process.
Agree what would pause the workflow: an unexpected recipient, a missing source, a wrong amount, a permission change or a failure that cannot be explained. Keep a practical manual route available. The recovery plan should say who can disable connections, which records may need correction and how work resumes without duplicating messages or transactions.
After rollout, review quality, exception volume and actual staff use alongside time saved. A pilot that creates more checking than it removes may need a simpler design. These acceptance and recovery arrangements are agreed as part of the scope, rather than assumed from a working demo.
FAQ
How much do AI implementation services cost?
Builds are quoted after we agree a written scope, which sets out the estimated delivery days. If you want help choosing the first job, the AI audit is a fixed £750. Managed AI starts at £250 a month, plus setup from £1,750. Software, platform, hosting and licence costs are separate.
Do we need an AI audit before you build anything?
No. If you already know which workflow you want built, we scope it from your plan. The audit helps when you are not sure where to start, and buying it does not commit you to a build.
Will it work with the tools we already use?
Usually, yes. We build into your current setup rather than replace it. If a build needs new access or a security or tenant change, we agree it with whoever runs those systems for you, such as your IT provider.
How long does an implementation take?
That depends on the job and the systems we need to connect. The written scope gives you the timeline before any work starts.
What happens after you hand over?
Your team gets documentation and training to use, check, pause and maintain the system. Support after launch is agreed in the scope. If you want us to keep running it, managed AI covers monitoring, fixes and agreed improvements each month.
Do you work outside Northern Ireland?
Yes. Most projects run remotely, and we work with clients in Great Britain and the Republic of Ireland as well as Northern Ireland. Prices are in sterling, and your quote confirms currency, VAT and payment terms.
.svg){kind=small}
.svg){kind=small}