One thing I see people getting confused with when they start with Claude is understanding what’s what. There’s Chat, Cowork and Code, then you’ve got the desktop app, browser and phone. The easiest way to make sense of it is to start with the job you want done: - Chat: asking questions, working through ideas and handling standalone tasks. - Cowork: working across your project documents and completing tasks with outputs you can save, review and update. - Code: building software, updating websites and creating more advanced automations. For most people working in construction, I’d encourage you to get working in Cowork as soon as possible. In the video, I show the difference between asking “What are liquidated damages?” and “What are the liquidated damages for this project?” That second question needs access to the actual contract. Start with one project folder and a question you already know how to check. Confirm Claude can see the relevant documents, ask your question, then check its answer against the source. I walk through how the different parts fit together, including working from your laptop, browser and phone in the video here.

@Tim Fairley or @Josh Turner have either of you (or anyone else for that matter) found AI or general resources for effectively managing worker competency on projects. We have a very long and wide excel file that has all staff and a wide range of skills that is not that useful when onsite managers/supervisors are trying to show staff are up to date on their skills and certifications (first aid and inductions being a classic one). I've looked at some online tools like Damstra and Skills-Base.com but I know one tool doesn't necessarily mean a good system. Love to hear if anyone has insight/experience on approach or tools that have worked.

Google’s AI strategy makes a lot of sense, even though I don’t think Gemini is the best model yet. The ecosystem has a few layers: Workspace: Gmail, Drive, Calendar, Chat, Docs, Sheets, Slides, Forms and Sites. This is where your files, communication and project data live. Gemini and Workspace Intelligence put AI inside those tools and let it work across the same context. NotebookLM retrieves answers from a defined set of project information. Spark and Workspace Studio handle agents and automations. Canvas and Antigravity cover app prototypes and software development. For a contractor, Forms can collect site data, Sheets can hold registers, Drive can store the documents and Sites can display a dashboard. The AI sits across all of it. The limitation is model quality. In my experience, Gemini can be too brief and hallucinate more than Claude. I wouldn’t migrate just for the AI today. If you already use Google Workspace, though, the integration is very useful, provided you check every important output. Full walkthrough: https://youtu.be/E95eQFsaa9s

I've been working on an AI-assisted subcontractor management system. It starts with the signed subcontract and agreed scope, budget and programme. Basically after procurement is done and you have locked in the sub-contractor. From there, we have each step in the process and some small, discrete Claude skills that you run to support each process. We also have a library of business templates and project information that sits behind them. Setup and mobilisation: summarise obligations, build a checklist for documents and site readiness, and agree how progress will be reported. Delivery: check performance against the plan, review payment claims against verified progress, assess variations and draft notices. For variations, check entitlement first, then price and time. This Close-out: check the final account and handover requirements, then record lessons learned and subcontractor performance for future projects. The system combines organised project records, shared business information and reusable AI workflows that you customise to your processes. Check out the walk-through: https://youtu.be/04Vw2WqI2dE

How to Use AI in Construction Estimating Without Losing Control of the Bid A construction estimate is a commitment: we can build this scope for this price, in this time, from documents we’ve often only had for a couple of weeks. That’s why handing the whole bid to AI and hoping for the best is a terrible idea. I think there are three sensible levels: - Start with risk reduction. Use AI to check your estimate for missed requirements, gaps, double-counting and formula errors. - Use it to structure information. Build requirements and clarifications registers, then review them before they feed into the next step. - Once your data is organised, use it for more of the heavy lifting like assemblies, line-item pricing, subcontractor quote comparisons and submission documents. The important part is keeping human review between each workflow so errors don’t compound. Where are you currently using AI in your estimating process?