Questions about AI Studio

Yes. Business Automation covers repetitive manual workflows — intake, routing, data entry between systems, notifications and reporting.

Yes. WhatsApp Automation is one of our services — automated conversations for enquiries, support responses, order updates and follow-ups.

Yes. AI Agents & Chatbots is one of our services — assistants that answer questions, qualify enquiries and guide users using your own approved content.

Yes. AI Websites & Web Apps is one of our six services — modern, fast websites and web applications with AI features built in where they add value.

Yes. Python, Data & AI covers data analysis, machine learning models, prediction and AI-driven insight work.

Yes. AI Social Media covers AI-assisted content production — drafting, reviewing, scheduling and publishing on a consistent rhythm.

Use the Contact page enquiry form. Share your requirement, and the team reviews it and replies directly.

Tell us what you want to build, improve or automate through the enquiry form on the Contact page, or describe it here and I can help you shape the enquiry.

We don't publish fixed prices. Engagements are structured as Starter, Growth or Custom and priced from the actual requirements. Share your brief and the team will respond.

We work across six services: AI Websites & Web Apps, AI Agents & Chatbots, WhatsApp Automation, Business Automation, Python/Data & AI, and AI Social Media.

Discovery, definition, design, build, review and handover. The Process page walks through each step.

Yes. We can ground answers in your approved content or documents so responses stay accurate.

Yes. Our portfolio includes sentiment classification and price prediction work built with Python.

Yes — dashboards, internal tools, portals and customer-facing applications with authentication and data behind them.

Secrets stay server-side, access is enforced with row-level security policies, and input is validated before it reaches the database.

We restrict it to approved knowledge, instruct it to say when it doesn't know, and route unknown questions to a human contact path.

AI drafts content from your brief and brand context, a person reviews and edits it, then it is scheduled and published.

Timelines depend on scope. We give an indicative timeline after discovery, once the requirements are clear.

Starter, Growth and Custom. They differ by scope and depth, not by a fixed feature list — the Pricing page explains the difference.

Answer common questions, capture enquiry details, send order or status updates, and follow up — with escalation to a person when needed.

We are an AI business solutions studio. We design and build AI-assisted websites, assistants, automations and data products for businesses.

Rule-based, repetitive work: moving data between tools, generating documents, sending notifications and preparing recurring reports.

The Work section lists our portfolio projects, and Case Studies cover selected projects in more depth.

No. Provider keys are read server-side only and never sent to the browser.

Yes. It can ask discovery questions about the problem, the desired outcome and timeline, then guide the visitor to submit an enquiry.

Yes. The same knowledge base can serve both the website assistant and WhatsApp conversations.

If a tool has an API or webhook, it can usually be connected. We confirm feasibility during discovery.

Yes. Describe the problem in plain language and we help translate it into a solution and the right service.

Yes, where the API is documented and accessible. Feasibility is confirmed during discovery.

Yes. Automated messaging runs through an official WhatsApp Business setup that your business owns.

Working software, source code, configuration, documentation of how it runs, and a handover walkthrough.

We clarify the business problem, users, constraints, existing systems and success criteria before proposing a solution.

Describe the business problem, what you want to build or automate, any systems it must connect to, and your rough timeline.

A website where AI supports the visitor journey — for example an assistant that answers questions, guides visitors to the right service and helps them start an enquiry.

Retrieval-augmented generation: relevant content is retrieved from your knowledge base first, then the model answers using that context instead of guessing.

A chatbot mainly answers questions. An agent also takes actions — looking things up, routing a request or triggering a workflow.

React and TypeScript on the front end, Python for data and AI, managed Postgres for data, and hosted AI models through providers.

Third-party services such as hosting, model providers or messaging platforms bill separately based on your usage.

Yes — that is a core use case. A structured FAQ knowledge base is retrieved and used as the basis for answers.

Yes. When intent or complexity requires it, the assistant guides the visitor to an enquiry or contact route.

Yes. Validated enquiry forms that store submissions securely and can notify your team are a standard part of the build.

Yes. A focused first phase is often the fastest way to validate the idea before expanding.

Yes. We can add an admin area so your team can edit content without touching code.

Yes. The Used Cars Price Prediction project demonstrates regression-based prediction from structured data.

Yes. We can rebuild or redesign an existing site and add AI features where they support the business goal.

Yes. Email/password and provider-based sign-in with role-based access can be part of a build.

Yes. Analytical dashboards can be delivered as part of a data engagement or as a web application.

Yes. The GenAI RAG Document Assistant project answers questions grounded in a document set.

Yes. The Airline Sentiment AI project in our portfolio classifies customer sentiment from text feedback.

Yes. We set up deployment and can hand over the environment so you own it.

Yes. Titles, descriptions, canonical URLs, social preview metadata, semantic HTML and a sitemap are part of a standard build.

Ongoing support and maintenance can be arranged as part of the engagement. Mention it when you enquire.

Yes. After discovery you receive a written scope and delivery approach to review before work starts.

Yes. The code is yours as part of the handover.

We turn your goals into a written scope: features, integrations, data, and what is explicitly out of scope.

No. A human review step sits between the AI draft and publishing.

Clean and structure it, analyse it, build dashboards, and where it fits, train models that support decisions.

Managed PostgreSQL, with row-level security so data access is enforced at the database level.

A retrieval-augmented assistant that answers questions grounded in a document set rather than generating unsupported answers.

Healthy Kitchen is a portfolio project in our Work section. Its detail page covers the problem, approach and technologies used.

We use hosted models through providers such as OpenRouter or the Lovable AI gateway, chosen per project for cost, quality and latency.

Typically React with TypeScript on the front end, a managed Postgres backend, and server-side functions for anything sensitive.

Yes. Every build is responsive and tested across common mobile, tablet and desktop widths.

WhatsApp conversation charges come from the messaging platform and are billed to your account based on your usage.

Yes. Conversations can be escalated to your team when the request needs a person.

Yes. Workflows can run on a schedule or be triggered by an event such as a new form submission.

Yes. The knowledge base and prompt are maintainable, so answers can be updated as your business changes.

Yes. Common support questions can be answered automatically, with unknown cases handed to your team.

Yes. A model can be served behind an API and used by a web app or automation.

Yes. Enquiry details collected in chat can be stored in your database or sent to your team.

Yes. Recurring reports can be generated from your data and delivered automatically.

Yes. Brand tone and rules are captured in the assistant's context so drafts stay consistent.

Yes. Approved content can be queued and scheduled ahead of time.

Yes, when the ordering or booking system exposes the data we need to trigger the message.

Commercials are agreed per engagement after scope is clear. Discuss the preferred model in your enquiry.

Yes, for selected projects. Case Studies go deeper into the problem, approach and technology.

Yes. Review rounds are part of each delivery phase; larger changes outside the agreed scope are handled as new scope.

Share your requirement through the enquiry form and the team will respond with next steps for a discovery conversation.

We minimise stored personal data, keep access controlled, and can support deletion requests. Legal compliance sign-off stays with you.

Yes. Engagements are scoped to the requirement, so both small businesses and larger teams can work with us.

Only if that is part of the scope. By default a session is ephemeral, and we avoid storing personal data without a reason.

Accuracy depends on the data. We measure it against a held-out set and report honest metrics rather than promising a number upfront.

We look at frequency, time spent and error rate. High-volume, rule-based tasks are automated first.

Scope, number of integrations, data quality, compliance needs, and how much of the workflow must be automated.

Failures are logged and can raise a notification so a person can review and reprocess the item.

That is normal. Cleaning, validation and structuring are part of the work before any modelling.

Changes are handled as scope updates. We review the impact on effort and timeline before agreeing to them.

An NLP project that classifies sentiment in airline customer feedback text.

A machine learning project that predicts used car prices from structured vehicle attributes.

On your website, and where scoped, in messaging channels such as WhatsApp.

Platform support depends on the API access available for your accounts; we confirm this during scoping.

We build on managed infrastructure that scales with usage, and design the data model with growth in mind.

No. Portfolio projects show capability and approach. We don't publish or promise outcome guarantees.

Yes — enquiry forms are standard, and booking links or scheduling flows can be added to the scope.

If your CRM has an API, conversations and enquiry data can be pushed into it.

Yes. Event tracking can be wired into key actions so you can see how visitors use the site.

Yes. Follow-up sequences can be triggered from events such as a new enquiry or an incomplete conversation.

We can build product and ordering experiences. Share what you sell and which payment or fulfilment systems are involved.

Performance is part of the build: lean assets, sensible loading, and testing before handover.

Yes. Semantic markup, keyboard navigation, labels and contrast are checked during the build.