An applied AI firm

Intelligence, put to work.

Plyxzy Solutions designs, builds, and operates applied-AI systems for classrooms, restaurants, and the enterprise data estate. A working system in weeks. Then years of keeping it running.

Explore our solutions Start a conversation

EducationHospitality & ConsumerEnterprise DataHuman-in-the-loop by design

What we build

Software, not slideware.

Some of it is ours to run. Some of it we build inside your organisation. All of it has to hold up on an ordinary Tuesday, not just in a demo.

Products we build

Our own software, in the open.

We build these ourselves rather than to order. DoubtBoard is live and open to anyone; AI Campaigns is close behind. They are where our own opinions get tested before we bring them to anyone else.

Live

DoubtBoard explaining why sin(x) makes a wave on an animated whiteboard

01 · Education

DoubtBoard: every doubt, explained at the board.

Students ask anything. The AI works it out the way a great teacher would: step by step on an animated blackboard, narrated aloud, instead of returning a wall of text.

Animated whiteboard explanations, drawn and narrated in real time
Twenty languages natively, with voice narration in each
Every graph computed in the browser, never drawn freehand
Free to start, paid tiers for papers and mock exams, bulk seats for schools

Try DoubtBoard

Launching soon Elegant restaurant dining room set for service

02 · Hospitality & consumer

AI Campaigns: marketing that runs itself.

For restaurants and consumer businesses that don't have a marketing department. The AI plans the occasions, writes the creative, and personalises every send. Nothing goes out without your approval.

Occasion-aware campaign calendars, planned months ahead
On-brand creative generated for you: copy, offers, and visuals
Guest segmentation that learns what brings people back
One-tap human approval before anything ships

Request early access for your venue

Work we take on

Engagements, not products.

These are capability lines we deliver inside your organisation. They are not something you can download; they are something we build with you.

In development Business documents and notes spread across a desk during analysis

01 · Enterprise data

Unstructured intelligence: your documents, finally answerable.

Contracts, SOPs, wikis, meeting transcripts. Eighty per cent of enterprise knowledge lives in prose no dashboard can reach. We turn it into a governed knowledge layer that answers with citations, not guesses.

Domain-organised knowledge, curated like an internal encyclopedia
Cited answers, every claim traceable to its source
Continuous ingestion of documents, meetings, and mail
Deploys alongside your existing stack, not instead of it

Join the early-access list

In development Rows of servers inside a modern data center

02 · Enterprise data

Structured foundations: records you can finally trust.

Master data, mastered by agents: matched, deduplicated, and mapped across systems with deterministic logic, with every consequential change routed through a human gate.

Deterministic entity resolution that is auditable, not fuzzy
Agent-drafted mappings and transformations, human-approved
Golden records that survive contact with your source systems
Warehouse-native: lands where your analytics already live

Join the early-access list

We build for ourselves first.

Building our own software is how we keep our judgement honest. The retries, the migrations, and the 3am pages are ours to answer for, and what we learn there is what we bring to your problem, rather than a methodology deck.

How we work

Weeks to working software.
Not quarters to a deck.

We earn conviction with running systems, not roadmaps. Four disciplined moves, in order.

Diagnose

We sit with your operators, not just your org chart, and find the one workflow where intelligence pays for itself fastest.

Week 0–1

Prototype

A working system in front of your team within weeks, on your real data. If it doesn't survive contact with reality, we find out early, and cheaply.

Week 2–5

Pilot

Measured against a baseline you agree to up front. Humans approve every consequential action while the system earns trust.

Week 6–10

Scale

Integration, governance, and operations: the unglamorous work that separates a demo from a dependable system.

Ongoing

The firm

We are a firm built on shipped software. The positions we hold about AI, we intend to earn the same way: in production, not in a workshop.

Applied over abstract

We don't sell AI strategy. We build the system, then the strategy writes itself.

Evidence over hype

Every pilot has a baseline and a number to beat. If we can't measure it, we won't claim it.

Humans in the loop

Nothing consequential happens without a person choosing it: an invitation you accept, a change you approve. Built into the architecture, not promised in a policy document.

Who you'll work with

Sunkkalp Kataria

Founder & Director

Building software professionally since 2017, almost all of it on systems other people depend on daily: dental laboratories filling orders, an audience watching live feeds in fifteen languages. Plyxzy Solutions is where that turns into applied AI: the same engineering discipline, pointed at problems where intelligence earns its place.

Product: Full-stack engineering across web platforms and APIs
Platform: Docker, Kubernetes, CI/CD, Elasticsearch on ECK, and observability with Prometheus, Loki, and Grafana
AI: Agent systems, and the plumbing around them

Independent contract engagements

DentHubCore platform team · 6 years · ongoing

Swedish cloud platform for dental laboratories, clinics, and chains: case management, intraoral scanner automations, clinic-to-lab communication, logistics and billing integrations. One of the leads across the platform, which has been in production for eight years.

WebcamCollectionsBackend & frontend · 15 months, concurrent

Live-webcam platform serving audiences in fifteen languages. Backend features, a full Angular to Vue.js migration on a running product, and the admin system its editors use to manage cameras and streams.

Enterprise consent platformBackend, for a former employer · 18 months

Built the backend from zero on GraphQL, Node.js, and MongoDB. That is the governed-record and audit-trail problem Structured Foundations is being built to solve properly.

Work with us

Put intelligence to work in your organisation.

Tell us the problem. We'll tell you honestly whether AI is the answer, and show you working software if it is.

Prefer email? Write to hello@plyxzy.com and it reaches us directly, not a sales desk.