AI software development case studies
AI software development at CML Team covers generative features, RAG chatbots, speech-to-text, live translation and computer vision, built into web and mobile products. The case studies below include Levee, a hotel inspection platform where we rebuilt the backend and added real-time object recognition, and RSI.Live, a conference platform with AI-assisted captions that we built from the ground up.
7 case studies in this industry
Case studies: AI
Generative AI for presentations and video clips
Two of our generative AI case studies turn rough input into finished content, one for slide decks and one for video. For Oria, a US-based startup, we delivered Oria AI, a PowerPoint plugin for one-click slide generation, and a scalable backend. Its AI pipeline for slide content is built with prompt engineering, and automated formatting keeps every deck on brand. Oria AI is in pilot with its first enterprise users across multiple countries, and early adopters report faster preparation cycles.
For Arbor, a New York-based startup, we built AI-driven generation of short clips from long-form video, with stylized captions and social media publishing. Users of the upgraded platform create and share ready-to-post clips in seconds, and Arbor scaled its user base rapidly.
RAG chatbots and AI assistants
A RAG chatbot looks up the relevant documents first and then answers from what it found. For Nestwise, a US-based family-focused AI startup, we built a parenting copilot on Retrieval Augmented Generation (RAG). Its engine recalls family context across sessions and gives evidence-based advice with citations to trusted sources. It also turns a parenting question into a step-by-step action plan.
Nestwise was delivered as a working proof of concept, which let the client validate product-market fit and collect real user feedback. For AI agents that run multi-step workflows, see our post on how we build AI systems that think, act and orchestrate.
Speech-to-text, live translation and computer vision
Levee is a US-based startup whose platform automates hotel housekeeping, employee training and brand standard audits. We rebuilt its backend from scratch and upgraded its iOS, Android and web apps. We also added computer vision that recognizes objects in real time during room inspections, and AI agents that turn what it recognizes into automated reports. Speech-to-text and translation cover 4 languages.
According to CML Team project records, hotel teams gained over 40% operational efficiency on inspections and reporting. The platform also helped Levee secure its first B2B contract.
For TRANSLIT, an Ireland-based language and interpretation services company, we built RSI.Live from the ground up. The conference platform combines real-time interpretation with AI-assisted translated captions and multilingual streaming. OpenAI and Agora RTC integrations handle speech and video synchronization. The platform helped TRANSLIT reach new enterprise clients.
Arbor relies on speech-to-text as well. Its transcript is synchronized to the video timeline, so trimming a clip trims its transcript with it.
AI features inside business software
Persona is CML Team Limited's own product, a business operating system for IT agencies and service companies. It keeps sales, delivery and hiring in one platform on web, iOS and Android.
Its AI agents parse resumes and match candidates, analyze recorded interviews, validate and enrich company data, and draft emails and proposals. It also gives RAG-based insights on profiles and deals. Per CML Team Limited project records, Persona brings 70% faster candidate screening and 50% quicker campaign setup.
AI and data for healthcare brands
Health Influence helps healthcare brands find and vet patient influencers and monitor their campaigns in real time. We built the platform end to end, from MVP to production. The discovery and analytics run on AI and machine learning, including automated sentiment analysis and real-time campaign monitoring, with patient-privacy safeguards built in. Healthcare brands get 35% faster campaign optimization on the platform, according to CML Team project records.
How an AI project with us starts and runs
An AI project with us can start small, with consultations. For project work we give a free estimate before you decide. In project-based development, our own project manager runs the project from requirements to launch. With a dedicated team, our engineers join your in-house team and report to your project manager or tech lead.
Billing is by the hour or, for a project, at a fixed price. We sign an NDA, and the rights to the code pass to you after payment. If a team member is not the right fit, we replace them without stopping the work. The services behind these cases are on our page about AI and LLM integration services, and how to get started lists the first steps.
Planning a product in AI? Tell us about it and get a free estimate.
Questions about AI software development
Can you start with a proof of concept?
Yes. Nestwise was built as a proof of concept to validate the business model and gather early user feedback. Oria AI moved from idea to a market-ready pilot with its first enterprise users. Health Influence began as an MVP and reached production, with new features added from real-world feedback.
Can you add AI to a product we already have?
Yes. Levee came to us with a basic prototype and an unstable backend. We rebuilt the backend, upgraded the apps and integrated computer vision and speech-to-text into them. For Arbor we built automated clip generation into its platform and did backend performance work for faster uploads and processing. If the product itself is unstable, our software project rescue page shows how we take it over first.
Have you built AI agents?
Yes. In Persona, AI agents parse resumes, match candidates, analyze recorded interviews and draft emails and proposals. In Levee, AI agents turn what the computer vision recognizes during a room inspection into the inspection report. Both work inside products in daily use, as part of the wider software around them.
What does custom AI software development include?
In our cases it includes the whole product around the AI. For Levee that meant a backend rebuilt from scratch, upgraded apps for iOS, Android and web, and App Store and Google Play releases. It also meant a QA pipeline for object-recognition accuracy. RSI.Live needed role-based access control, event scheduling, usage-based billing and credit management with Stripe. For Oria AI we also supported the deployment to the cloud.
Can an AI assistant answer from our own documents?
Yes, that is what Retrieval Augmented Generation is for. The Nestwise engine indexes pre-installed and user-uploaded document libraries and cites trusted sources. Persona gives RAG-based insights on the profiles and deals kept in the platform. The answers depend on the documents behind them, so plan time to clean up and update the library.
Do you work with companies in the US and Europe?
Yes. Arbor (New York), Levee, Oria and Nestwise are US startups. TRANSLIT, the client behind RSI.Live, is based in Ireland. Our working hours overlap fully with Europe. With the US, the overlap depends on the state your team is in.
Have a project in this industry?
Tell us about your product and we will suggest the right team and approach.
Get in Touch








