A guide to the tools, what they can help with, and what to check before using them on a job.
Find your starting point↓Choose your role for a suggested first step and a few tools to explore.
Learn what the systems can do, set the operating boundary, then sponsor one measured workflow—not a broad AI programme.
Learn the difference between a model, assistant, agent, automation and construction-native platform before comparing brands.
Start with ChatGPT, Claude, Gemini or Copilot. Practise context, source material, clear instructions and review.
Move from generic prompting to controlled work with approved documents, visible sources and a defined decision owner.
Choose a repeatable task, record the current baseline and compare quality, time and adoption before expanding.
Short introductions from the teams behind the tools.
A short official introduction to opening ChatGPT and beginning a conversation.
▶V.02Anthropic / YouTubeAn official walkthrough of the everyday Claude experience and its basic working model.
▶V.03Anthropic / YouTubeA deeper official session for technical users moving from Claude into agentic coding work.
▶V.04Mistral AI / YouTubeAn official introduction to the assistant previously called Le Chat and now part of Vibe.
▶Assistants help you think. Research systems help you find. Agents can act. Coding tools build. Construction-native systems work inside industry data and workflows.
39 systems in view
Everyday assistants
A general AI workspace for drafting, analysis, files, spreadsheets and structured problem-solving.
Everyday assistants
A general assistant well suited to careful document work, explanation and long-form writing.
Everyday assistants
Google’s general assistant, with a natural fit for teams already working across Google Workspace.
Everyday assistants
AI embedded across Microsoft 365, grounded in the work environment your organisation permits it to access.
Everyday assistants
A general assistant for conversation, web information, file analysis and media tasks.
Everyday assistants
A family of general and reasoning models available through consumer and developer products.
Everyday assistants
Mistral’s combined work, chat and coding environment, previously known as Le Chat.
Everyday assistants
A consumer assistant closely connected to Meta’s messaging and social products.
Research & knowledge
A source-grounded notebook that answers from the material you provide and keeps citations visible.
Research & knowledge
A web research assistant designed to return answers with links to the underlying sources.
Work agents
Outcome-based work in ChatGPT for producing multi-step deliverables from files, tools and instructions.
Work agents
Claude’s agentic environment for longer-running knowledge work beyond a standard chat.
Work agents
Agentic work across permitted Microsoft 365 information and applications.
Work agents
An autonomous general agent that plans and completes work inside its own cloud computer environment.
Work agents
A business agent platform for communication, meetings, follow-up and connected application workflows.
Coding & app building
A coding agent that can understand a codebase, make changes and verify technical work.
Coding & app building
An agentic coding tool for understanding, changing and operating software projects.
Coding & app building
Coding assistance and agentic development work integrated with GitHub and common editors.
Coding & app building
An AI-first code editor that combines direct editing with agent-led implementation.
Coding & app building
Google’s agent-first development environment for coordinating coding work across an IDE and agents.
Coding & app building
A conversational builder that turns an application idea into working, hosted software.
Coding & app building
A natural-language platform for building full-stack websites and applications with editable code.
Coding & app building
A browser-based AI builder for creating, changing and publishing web applications through conversation.
Coding & app building
An AI environment for generating interfaces and developing web applications from natural-language instructions.
Automation
Agents and automations connected to a broad catalogue of common business applications.
Automation
Visual automation that combines fixed workflow steps with agentic judgement where needed.
Automation
A technical workflow automation platform combining API integrations, business logic and AI agents.
Automation
A low-code environment for building governed agents and flows around Microsoft and enterprise data.
Automation
A developer framework for designing specialist agents that cooperate inside a defined technical workflow.
Construction-native
Construction-native search, drafting and agents grounded in permitted Procore project records.
Construction-native
An assistant for finding and understanding information held within supported Autodesk construction products.
Construction-native
Analytics and predictive intelligence across Oracle construction, programme and safety information.
Construction-native
A visual intelligence platform using site capture to create records and track visible progress.
Construction-native
Construction progress intelligence comparing captured site conditions with models and schedules.
Construction-native
Construction-specific contract intelligence that identifies and cites risk across project documents.
Construction-native
Construction knowledge tools that answer across project documents and retain links to the source.
Construction-native
AI-assisted measurement and comparison tools for construction drawings and take-offs.
Construction-native
Generative construction scheduling for comparing methods, resources and delivery scenarios.
Construction-native
AI schedule-risk analysis informed by a large collection of historical construction programmes.
Choose a task you can check against existing project records. Agree who will review the result.
Tender review, scope extraction, exclusions, bid comparisons, drawing interrogation and take-offs.
Begin with one completed tender and compare the AI output against the estimator’s final record.Meeting preparation, RFIs, submittals, daily records, lookaheads and progress reporting.
Begin with drafting and retrieval. Keep issue, approval and contractual actions with the authorised person.Contract comparison, obligation registers, notices, variations and supporting evidence.
Use source citations and legal or commercial review. Never treat a plausible summary as contractual advice.Site diaries, visual records, observations, coordination and toolbox material.
AI can prepare information; accountable people retain control of safety decisions and work methods.Source-grounded search across drawings, specifications, correspondence and project records.
Establish permissions, revision logic and a clear source hierarchy before trusting retrieval at scale.Inbox triage, CRM updates, reporting, document movement and recurring administrative workflows.
Use fixed automation for known steps. Introduce agent judgement only where the inputs genuinely vary.The more a system can see or do, the stronger its permissions, evidence, review and recovery controls must become.
Know what the system receives, where it is processed, how long it is retained and who can retrieve it.
Contracts, safety information, cost and programme decisions require a traceable basis—not an answer that merely sounds right.
Email, system updates, notices, commitments and external communication stay behind a deliberate human approval point.
Record quality, cycle time, omissions and adoption against the existing process before expanding the workflow.