Gemini, ChatGPT, Copilot and Claude — rolled out properly.
Most companies already pay for AI they do not use well, or use AI they do not pay for on private accounts. We set up the platform that fits your workplace — Google Gemini, OpenAI ChatGPT, Microsoft 365 Copilot or Anthropic Claude — with the licences, data protection, policies, training and integrations that turn it from a novelty into a tool. And because we are a cloud company first, the same platforms serve us as building blocks on Google Cloud, Azure and AWS for the solutions we build.
Google Gemini
Gemini inside Google Workspace — Gmail, Docs, Sheets, Meet — for companies on Google; NotebookLM and Gemini Enterprise for company knowledge and agents; Vertex AI and the Gemini API on Google Cloud for custom solutions.
- Strong with long documents, images, audio and video
- Natural fit for Google Workspace companies
- EU regions on Google Cloud for custom work
OpenAI ChatGPT
ChatGPT Business and Enterprise for staff, with a company workspace, administration and no training on your data; custom GPTs and projects on company documents; the OpenAI API and Azure OpenAI for custom solutions, with EU data-residency options.
- The most widely known assistant — least training effort
- Agent and browser-based features as they mature
- Azure OpenAI for companies that want it inside their Azure tenant
Microsoft 365 Copilot
Copilot inside Word, Excel, Outlook, PowerPoint and Teams, grounded in your SharePoint and OneDrive with the permissions you already have; Copilot Studio for agents; Azure AI Foundry for custom solutions.
- Natural fit for Microsoft 365 companies
- Permissions and data boundary inherited from the tenant
- Agents on your Microsoft data without new infrastructure
Anthropic Claude
Claude for teams and enterprises, with a company workspace and projects; the Claude API on AWS Bedrock and Google Vertex AI in EU regions for custom solutions.
- Strong at long, careful writing, analysis and coding
- Available inside AWS and Google Cloud accounts you already have
Open-weight models
Llama, Mistral, Gemma and similar models run on your own cloud environment — for data that may not leave, for high volumes, or for full control over versions.
- Data never leaves your environment
- Lower cost per task at volume
- Mistral as a European provider
Which one fits
Companies on Microsoft 365 usually start with Copilot; companies on Google Workspace with Gemini; mixed or browser-first teams with ChatGPT Business or Claude. Behind custom solutions, the model is chosen per task — and can be changed, because we keep prompts, evaluation sets and knowledge sources portable.
Plans and terms
Business or enterprise plans only, with administration, audit logs and terms that exclude training on your data. Processor agreements signed; EU processing or data residency chosen where the provider offers it.
Identity and access
Single sign-on, multi-factor authentication, licences assigned by role, access removed with offboarding — the same identity you already manage.
Data classification
A short, workable rule for what may go into which tool: public, internal, confidential, personal data. Connected sources respect existing permissions.
Retention and logging
Conversation retention set to your policy; logs available for review; no shadow accounts.
Transparency and the AI Act
Users know when they interact with an AI system; documentation of what is used for what. Most workplace uses are minimal-risk; we identify the category.
Shadow AI replaced
Private ChatGPT and Gemini accounts used for company work are replaced by the governed platform — the single largest data-protection improvement most companies can make in a month.
Assess week 1
Who does what, which platform is already licensed, which data may be used where, which five tasks would benefit most.
Configure weeks 2–3
Workspace or tenant settings, single sign-on, data controls, retention, administrator roles, connected sources with permissions.
Connect and template weeks 3–5
Company documents and knowledge sources connected; templates, custom GPTs or agents for the top tasks built with the people who do them.
Train and govern weeks 4–6
Sessions by role, a usage policy people can follow, champions in each team, and a monthly usage and value review.
- Platform choice with reasons written down.
- Configuration of tenant, identity, data controls and retention.
- Connected knowledge sources with permissions respected.
- Templates and agents for the five tasks with the most hours.
- Training by role and a usage policy.
- Documentation for data protection and the AI Act.
- Monthly governance optional: usage review, new features assessed, policy kept current.
Isn’t the free version enough?
Consumer and free versions may use inputs for training, lack administration and audit, and are tied to private accounts. For company data the business and enterprise plans are the appropriate level — and they are inexpensive per user compared with the time involved.
We have Microsoft 365. Do we need anything else?
Often not for staff productivity: Copilot covers it. For solutions that reach customers or process documents at volume, a model behind the scenes is chosen per task — sometimes Azure OpenAI in your own tenant, sometimes another provider.
Can we switch platforms later?
Yes. Prompts, templates, evaluation sets and knowledge sources are kept portable; switching is a project of weeks, not a rebuild.
Do you build agents on these platforms?
Yes — on Copilot Studio, Gemini Enterprise, OpenAI’s and Anthropic’s agent tooling, or custom on your cloud. See Agents for multi-step work.
How is this priced?
A fixed rollout fee by company size; licences at the providers’ list prices, passed through; an optional monthly governance retainer.
AI solutions catalogue
Twelve solutions working in companies of your size: customer-service and knowledge assistants, invoice processing, inbox triage, sales drafts, reporting — with pilot times and what you get.
Read more →ServicesManaged IT and workplace
Service desk, Microsoft 365 / Google Workspace, identities, devices and licences — run for you.
Read more →AI for businessAssistants and automation
How assistants and automations are built safely: retrieval, permissions, guardrails, evaluation — and how a pilot runs.
Read more →Tell us which workplace suite you use and how many people you are.
We reply within one business day with the platform that fits, the rollout plan and its fixed fee.