Accessing AI Tools¶
Last updated: June 10th 2026
This page explains how to get access to DHI Group's approved AI tools, what you must complete before using them, and which tool to choose. For the full AI policy and guidelines, see the AI Guidelines.
How to Get Access¶
Access to DHI's approved AI tools is provided through DHI Core IT, primarily via the ServiceNow portal. To request a seat for GitHub Copilot or Claude Code, raise a request through the DHI ServiceNow portal (Software License Request).
Core IT manages the enterprise agreements, provisions seats, and allocates licence costs back to your department.
Complete Onboarding First¶
It is paramount that you complete the required onboarding before using any AI tool.
Before you start using AI tools, you must:
- Read and acknowledge the DHI AI Guidelines, starting with the AI Guidelines, Getting Started & Best Practices, and AI in Software Development.
- Complete the mandatory training listed on the Getting Started & Best Practices page.
- Work through the additional mandatory and recommended documents listed on the Getting Started & Best Practices page.
When to Choose What¶
DHI offers two approved AI coding tools. In short:
- GitHub Copilot - for in-IDE code completion, snippets, boilerplate, and lighter agentic tasks. Every developer can have a Business seat in the GitHub Copilot Enterprise plan. No extra credits are provided beyond that. This covers everyday completion and snippet work which are both free plus marginal agent use.
- Claude Code - for complex, multi-step agentic work across a repository. Staff start on a Standard seat, with a Premium seat available as an upgrade when justified.
Upgrading to a Claude Code premium seat: Premium seats are not granted automatically. First, please contact the AI CoE and then obtain your line manager's approval through the DHI ServiceNow portal (Software License Request). The AI CoE will review how you are actually using the tool - the skills, workflows, and use cases you run - to confirm the upgrade is justified before the premium seat is provisioned.
Examples of the kinds of things the AI CoE may look at (not an exhaustive list):
- Demand - whether you are regularly hitting the standard seat's usage limits.
- Use case fit - work such as large multi-file refactors, framework or API migrations, long agentic sessions, or building reusable skills and workflows.
- Competency - completed training and sound prompting, review, and data-handling discipline.
- Value - a clear, articulable productivity or delivery gain, ideally one that benefits the wider team.
GitHub Copilot¶
GitHub Copilot accelerates development by generating code, boilerplate, tests, and explanations directly in Integrated Development Environments (IDE) e.g. in Visual Studio, Visual Studio Code and JetBrains IDEs.
GitHub Copilot can also be used as an agentic agent (as coding partner) but not as an autonomous executor as its capabilities are limited compared to Claude Code and the credit cost is prohibitive.
Every developer can request a standard seat. See When to Choose What for guidance on choosing between GitHub Copilot and Claude Code.
Claude Code¶
Claude Code is an advanced agentic coding tool developed by Anthropic, designed to execute multi-step processes via command line (CLI) for complex code generation tasks.
Staff start on a standard seat. A premium seat is available as an upgrade when justified - see When to Choose What for the upgrade path.
Claude Cowork¶
Status: Under Legal and IT Security review - not yet approved
Claude Cowork is Anthropic's agentic desktop assistant, bringing Claude Code-style capability to non-technical staff for everyday knowledge work such as organising files, drafting documents, synthesising research, and running recurring tasks.
Cowork carries a higher risk profile than DHI's other approved office tools: it runs as a local agent with access to files, applications, the browser, and connectors, and Anthropic currently excludes Cowork activity from its Audit Logs, Compliance API, and Data Exports. It is therefore under Legal and IT Security review and not yet approved for use with DHI internal or client data. See General-Purpose AI Tools for details.
Perplexity¶
Status: Trial ongoing
A trial of Perplexity - an AI-powered research and answer engine - is currently ongoing to evaluate its usefulness for research and information discovery.
Note: AI tools and agents are evolving rapidly. The AI CoE will regularly reassess tool selection to ensure DHI can adopt new or improved agents as needed.