Pruthvik
AI / Cloud Engineering Lead
Pruthvik leads AI and cloud engineering, including the cloud infrastructure the teams build and deploy on.
AI features for apps and web products
We add AI to products that need it: assistants that answer from your own content, and photo and text analysis, with a person checking the result where it matters. We use third-party models through their APIs, built into the app and back end by our team in Rajkot, India.
- We build with
- Third-party models through APIs
- Published
- Our own WooCommerce plugin
- Lead
- Pruthvik, AI / Cloud Engineering Lead
Who this suits
- Products with their own documents, catalogue or data that users keep asking about.
- Apps that could read a photo, a label or a block of text for the user.
- Teams that want to try an AI feature without rebuilding their product.
What we take on
The work around the model is most of the job.
Is AI the right tool?
Some requests are better solved with search, rules or a good form. We say so, because that is cheaper and more reliable for you.
Answers from your own content
Preparing your documents, products or policies, finding the right part for each question, and testing that the model says so when the answer is not there.
A safe back end
API keys stay on the server, never in the app. Requests are limited, logged where you agree to it, and protected against abuse.
Privacy decisions
What is sent to the provider, what is removed first and what is kept, written down and explained to your users.
Store rules
Apple's guidelines require apps to disclose when personal data goes to third-party AI and to get explicit permission first. We build that in, and check the current store rules for your app.
Cost control
Model choice, length limits and caching, so a popular feature does not produce a surprising bill. Provider usage is normally billed to you.
When it fails
Models get things wrong and services go down. We plan what the user sees, and where a person must check the result.
Technologies: OpenAI API · Python · PHP / Laravel · Swift · Firebase
Case studies: We have not published an AI client case study. Our plugin can be installed and tried.
Our AI work
What we have built.
We do not quote accuracy figures: results depend on your content, so we test against your own questions.
Keshav Infotech AI Store Assistant
WordPress · WooCommerce · OpenAI (optional)
Our own free WooCommerce plugin, released in November 2026. Local Only mode answers from store data with no external AI calls and no key. Smart AI and AI First use the merchant's own OpenAI key, billed by OpenAI, and in those modes a visitor's question may be sent to OpenAI, including a typed name or postal address; removal is not guaranteed.
Photo and text analysis in mobile apps
Mobile app · our back end · third-party models
We have connected mobile apps and back ends to third-party AI models for photo and text analysis. The models are the provider's; our part is the app, the endpoints and the handling around them.
The work on our record does not include training our own models. If your project needs that, tell us and we will say whether we are the right team.
Who you work with
Who would lead this work
Rajesh
Backend Development Lead
Rajesh leads backend development: the APIs, databases and business logic that web and mobile apps depend on.
Dharmesh
iOS Development Lead
Dharmesh leads native iOS development: iPhone and Apple Watch apps.
Related
Useful next reads
How an engagement runs
- You tell us what you need and what exists today.
- We agree a written scope, or a dedicated team, with you.
- We build in short cycles you review, then release and support.
Questions
Before you get in touch
Do you build AI chatbots for businesses?
Yes, when a chatbot is the right answer. For WooCommerce shops we have our own free plugin, AI Store Assistant. For anything else we would build the assistant into your product, connected to your content.
Do you train your own AI models?
The work on our record uses third-party models through their APIs. Most business products do not need a custom model, and we will tell you if yours looks like an exception.
Will our data be used to train the provider's model?
That depends on the provider and the plan you use, so we check the provider's current terms with you before anything is sent. We also decide together what should be removed first.
How accurate will it be?
We cannot promise a figure, because models can be wrong and results depend on your content. We test against real questions from your users, and plan for wrong answers, including when a person should check the result.
What does it cost to run?
There is the build, then usage charges from the AI provider, which grow with use. We estimate usage from your expected traffic before you commit, and we build in limits.
Can you add AI to an app we already have?
Usually, yes, and the app itself changes little: typically one screen and one server endpoint. If it answers from your own content, there is also the work of preparing that content and keeping its index current.
Start a project
Tell us what the AI should do.
What people would ask or do, what content or data it should use, and what exists today.
What happens next
Your message goes to Ghanshyam Gadhiya, partner and CEO, who replies by email. If it is easier, you can also reach us on WhatsApp.