How I use AI at Made Grand
AI tools have become a substantial part of how I run Made Grand, develop products and explore new business ideas. I use it to research, plan, write, code, create visuals and connect information across different systems.
Some of that work is straightforward e.g. preparing an email draft or turning notes into a useful document. Other work involves building applications, in-depth research or giving an AI assistant controlled access to project information.
That practical experience is the foundation of Made Grand’s AI training and consultancy. This article brings the main examples together, explains where AI fits and shows the judgement that still sits around it.
Starting with the work that needs doing
I’m a web developer and creative technologist with more than 25 years of experience. My current work includes building websites, membership platforms, booking systems, e-commerce and custom business tools.
Using AI is helping me build on that experience. I still need to understand the problem, choose an approach and decide whether the result is suitable for real use. This part is based on my experience, professional judgement and expertise in my field.
I work across AI assistants, development platforms such as Lovable, image-generation tools and connected workflows. The choice depends on the task, the information involved and the systems a business already uses.
Often, the useful starting point is a repeated job, like a report somebody prepares each month, information copied between applications or a document that always needs the same structure and checks.
Researching business opportunities and markets
I use AI to explore potential services, compare business models, review competitors and organise market research. That includes work on Made Grand’s own services and products such as Eat With Friends, our group dining app.
It helps me turn broad questions into something I can investigate. Who is the customer? What problem are they trying to solve? What alternatives exist? What would a useful first version include?
For Eat With Friends, that work has included exploring the needs of meal organisers and venues, considering possible revenue models and preparing venue outreach material. AI helps me compare options and challenge assumptions before committing development time.
I also use it to review company websites before meetings and turn findings into briefing notes, questions and reports.
Researching suppliers and checking information with n8n
For Planet Preserve (a personal e-commerce project still in development), I’ve been developing supplier and product research workflows using n8n.
n8n is a tool for joining applications and steps together. A workflow can receive a structured brief, pass information to an AI research step, check the returned records and save the results for human review.
The supplier work brings together sourcing requirements, potential suppliers, supporting information and questions that need further investigation. Related product-intake work has passed candidate records back into a review system.
An important part of the build has been deciding what happens when information is incomplete or inconsistent. Testing has included catching a supplier appearing in both a shortlist and a rejected list, and flagging product candidates that still need evidence.
AI can help find and compare claims made by suppliers. Confirming those claims requires checking the underlying sources. A supplier’s sustainability statement, for example, does not become independently verified because an AI says it is.
This work has taught me to keep unknown information visible and include a review before supplier approval or product publication.
Planning and building complex website functionality
AI tools support both the thinking and the coding in my development work.
I use it to break down requirements, compare technical approaches, investigate errors, review code and prepare focused implementation instructions. That includes WordPress work and applications built with React, TypeScript and Lovable.
A useful example is working through membership and learning-platform rules. Questions about which content a member can access, when it becomes available, and how parent and child accounts interact, all need careful rule definitions before changing code.
AI helps me define those rules, identify awkward cases and investigate why the current behaviour differs from the intended behaviour. I then review and test the change against the actual system before implementing.
On Eat With Friends, I use AI-assisted development alongside work on user journeys, invitations, guest responses and venue discovery. It helps turn an idea into a clear specification and manageable development steps, keeping track of the roadmap, features and documentation.
AI helping to build a website is separate from AI being a feature inside that website. A customer may use a booking or invitation journey without ever interacting with an AI assistant.
As a Certified Lovable Expert Partner, I also bring this experience into development support and guided sessions for people building their own applications.
Connecting AI to business information through MCP
An AI conversation becomes more useful when it can work with relevant information from an authorised business system.
MCP stands for Model Context Protocol. In everyday terms, it provides a common way for an AI assistant to use tools supplied by another application. An MCP server makes specific operations available, such as finding a project or adding a note.
Those operations can be narrow. A connection might allow the assistant to read project records without giving it permission to delete them or access every other part of the business.
Made Grand’s Project Command Centre is an example. It brings projects, ideas, priorities, notes and next actions into a private portfolio dashboard. Its MCP connection lets supported AI tools work with the project information made available to them.
That gives the conversation a firmer starting point. I can discuss priorities using structured project records rather than repeatedly reconstructing the situation from memory.
MCP and workflow automation serve different purposes. MCP gives an assistant access to defined tools. A workflow system such as n8n moves information through an agreed sequence of steps. They can work together where the process calls for it.
Building a system for branded marketing content
I’ve also used AI to build a private marketing dashboard through ChatGPT’s website-building tools: Made Grand Operations.
The work brings content planning, drafts and creative assets into one place for Made Grand, Eat With Friends and Lovable training content. Keeping those workstreams distinct is important because each has its own audience, message and visual identity.
It also exposes the practical challenges of AI-generated marketing. A convincing image can use the wrong font, omit a logo or miss the purpose of the message, or just feel like AI slop. Brand guidance needs to influence the actual output, and I need a clear way to review it and request changes.
Developing that process has involved brand and artwork guidance, feedback and approval requirements. The system continues to evolve, with connected publishing treated as a separate stage. The AI generated visuals are still a way off being ready for any marketing campaigns, in my opinion.
For another business, the same thinking could support a shared process for turning approved service information into draft articles, campaign material and social posts, with somebody responsible for the final decision.
Writing emails, articles and useful business documents
Some of my most frequent AI tasks are the ordinary jobs involved in running a small business.
I use it to prepare client email drafts, explain technical work, shape proposals, draft blog posts and adapt messages for different audiences. It helps me turn rough notes into something another person can understand and act on.
The context i.e. who the message is for, what has already happened, what I want to achieve and how Made Grand should sound, makes a substantial difference to the output.
I use connected access to Google Drive where appropriate, and work with spreadsheets, documents, presentations and PDFs. Recent examples include drafting outreach spreadsheets, business plans, training material, a participant workbook and presentation slides.
This work also includes organising records, comparing lists, identifying missing information and preparing reports from supplied material.
The finished files always need checking. Calculations, names, links, page layouts and the relationship between different documents can all need attention. Producing a spreadsheet or PDF is only useful if somebody can rely on and use it and it represents the quality of the brand behind it.
Creating images, animation and video material
AI is part of my creative work too. I use it to explore visual directions, create image drafts, develop promotional concepts and prepare supporting artwork.
More recently, that has extended into video-related work, including Made Grand logo intro and outro animations I’m considering using, storyboards and promotional scripts for Eat With Friends.
Naturally, the process varies. Sometimes an AI tool generates an image, or helps write the code that animates supplied brand assets. For video, it can also help structure a sequence and prepare the script before editing and production.
The brief still needs creative direction and a clear goal in mind for the business and the customers.
I review generated material for visual quality and accuracy every single time. Product demonstrations need genuine product captures, and illustrative imagery should not imply a real customer or completed project.
Knowing what to check and when to stop
Working across these tasks has made the limitations as useful to understand as the capabilities.
My approach is to define the expected results, provide suitable information, test the output and keep responsibility clear. Where a process writes to another system or produces public material, permissions and approval points belong in my hands.
I also distinguish between experiments, working internal tools and production services, which need different levels of testing and support. I haven’t attached time-saving percentages or commercial returns to these examples because I haven’t measured them consistently enough to make those claims.
What this means for Made Grand’s AI training and consultancy
The service brings together everyday AI use and the development experience needed to build something around it.
Training helps teams give clearer instructions, use appropriate reference material, check answers and develop reusable ways of handling real tasks. People need opportunities to practise, understand mistakes and decide when a task needs further expertise.
Consultancy helps identify which opportunities deserve attention. We can review repeated work, assess the available tools and information, then agree on a practical next step.
Implementation can take that further through a focused bespoke dashboard, application or connected workflow. That includes defining access, testing behaviour, handling failures and making sure somebody owns the result after handover.
My experience comes from applying AI across Made Grand and my own products, alongside established development work. These examples show the practical basis of the service; they are not claims of accredited training or measured client transformation results.
If your team is already experimenting with AI, or you have a repeated task you would like to improve, explore Made Grand’s AI training and consultancy or get in touch.
We can start with the work you need to do and decide what would make it easier.
If you're planning an MVP or early-stage product and want to make sure the foundations are right before you build, we're happy to talk it through.
Start a conversation