Custom product features
Recommendations, personalisation, content generation, or assistive writing built into your app or platform.
AI can power product features that feel native - recommendations, personalisation, content generation, or assistive writing. The key is fitting the AI to your product and your users, not the other way around.
I work with product teams to design and build AI features: recommendation engines, personalised dashboards, content generation (drafts, summaries, variations), or writing assistants that understand your domain. For SaaS companies, agencies, or product-led businesses, this might mean adding AI to an existing product or prototyping a new AI-powered feature. I focus on shipping something useful, then iterating based on real usage.
Example AI integrations
AI services and tools I've integrated for businesses include:
Vercel AI SDK
React streaming hooks for chat UIs and AI completions. For product features, it powers chat UIs and streaming completions in React apps.
Cohere
Embedding and rerank APIs for semantic search in products. For product features, it adds semantic search and reranking to apps.
Together AI
Open model inference API for embedding AI in apps. For product features, it provides inference for embeddings and completions.
OpenAI
GPT APIs for chat, embeddings, and assistive features. For product features, it powers chat, recommendations, and assistive writing.
Anthropic
Claude API for long-context and tool-use in products. For product features, it enables long-context chat and tool-calling in products.
Stability AI
Image generation and multimodal AI for product features. For product features, it adds image generation and visual AI to apps.
Try these free tools
Types of businesses I work with
- E-commerce and retail - Product search, customer support, inventory and order triage, and personalised recommendations.
- Manufacturing and engineering - Process documentation, quality checks, supplier comms, and internal knowledge bases. Often starting with one high-friction workflow.
- Agencies and service providers - White-label AI solutions, client delivery, and internal efficiency tools for teams that bill by the hour.
Frequently asked questions
- AI features can be added to existing apps without rebuilding. I integrate via APIs and SDKs so recommendations, search, content generation, or assistive writing slot into your current product. The AI layer sits alongside your existing codebase.
- Recommendation engines analyse user behaviour and content to suggest relevant items - products, articles, features, or actions. They learn from usage patterns and improve over time. The AI considers context like what the user has viewed, purchased, or searched for.
- Costs depend on the feature complexity and scale. A simple content generation or assistive writing feature might take two to four weeks to build. Running costs come from the AI model API usage, which scales with your user base. I scope and estimate before starting.
- I focus on shipping a minimal useful version first, then iterating based on real usage data. The AI should solve an actual user problem - saving time, improving discovery, or reducing effort. If users don't engage with it, we adjust or remove it.
Can AI features be added to an existing app or does it need rebuilding?
How do AI recommendations work in a product?
What does it cost to add AI features to a SaaS product?
How do you ensure AI features are useful rather than gimmicky?
Want to discuss AI for your business?
I help businesses integrate AI into their workflows. Get in touch to talk through your specific situation.