AI platform that takes care of data entry, update, query and analytics. Your data stays accurate, organised and ready .
Tarsk is our AI Assisted Data Management platform, based on a purpose-built retrieval-augmented generation (RAG) pipeline, architected specifically for managing the complexity of project data in a team environment.
It handles all data entry, data query, analytics, semantic search and data administration duties, connected to live project data, including schedules, costs, BOQs, risk registers, issue logs, inventory, and more.
Tarsk works alongside you. You can talk to it, message it, or send it an email. Alternatively, you can do it yourself the old-fashioned way on the web interface, whatever is easier in the moment. It’s a collaboration.
You can talk to Tarsk on a phone or tablet when you are away from your desk. Add updates, notes, photos or issues while you are on the move.
You can submit information by email, report, message, or voice. Tarsk reads it, picks out what matters and updates the database. Anything it is unsure about is held in a review queue for approval.
Tarsk can run queries across your data, apply calculations and aggregations, and present the results in natural language alongside data, visualisations or formal reports.
Want to edit or query your project database (Gantt Chart, BOQ, risk register etc..)? Just ask, or perform the action yourself directly in the web app, whatever is easier in the moment.
We’re redesigning the UI from the ground up. The updated experience will include a familiar chat interface, seamlessly integrated with a flexible, reconfigurable dashboard and data view.
Your project database can be shared with your team, with full role-based access control.
Initial data upload supports structured formats like Excel and CSV. In future versions, we plan to introduce direct integrations with third-party databases and tools.
Not ready to fully trust the AI just yet? Updates can run automatically or be queued for manual review.
Performance indicators (KPIs) are calculated in real time, no spreadsheet wrangling needed.
Log updates via mobile or web app, using voice or messaging. You can even email the AI agent.
All changes are logged and traceable, so you always know who did what and when.
We’re targeting a version 1.0 release in Q4 2025.
We’re targeting a version 1.0 release in Q4 2025.
Version 1.0 will support essential project management workflow logic. We’re also developing a configurator to enable customizable, rule-based logic pipelines for version 2.0 in 2026.
We plan to launch a suite of advanced analytics tools in version 2.0, including risk assessment, critical path analysis, extension of time, window analysis, and more. We may prioritize operational, risk, or commercial modules based on user feedback.
While version 1.0 supports structured imports (e.g. Excel, CSV), we will introduce integration with third-party platforms, enabling live data synchronisation and interoperability.
We’re building this platform for real users. we have already solicited much feedback from industry practitioners. If you have suggestions, we’d love to hear them. We are ready to roll up our sleeves!
For us, project data management is just the beginning. Our broader vision is to redefine how data, in general, is structured and managed. Let’s talk!
Project data is stored in a multi-tenant PostgreSQL database hosted on a secure cloud infrastructure (e.g. AWS, Supabase, or equivalent). Enterprise clients may optionally request dedicated database instances for data isolation and compliance purposes.
All data is encrypted in transit (TLS 1.2+) and at rest (AES-256 or equivalent cloud-provider encryption).
Authentication is enforced per request, with user roles and tenant boundaries strictly respected at the application layer.
Automated backups are performed regularly, with recovery options available for enterprise tiers.
We’re looking for a small group of like-minded development partners. While the platform is not yet deployment-ready, we offer the opportunity to co-develop a solution tailored to your needs, shaping the product together, from the ground up.
We’re also seeking investors who share our ambition and vision. If you believe in building the next generation of data platforms, we’d love to hear from you. Please reach out for more information.
We’re launching with a freemium model, so you can explore the platform and start using core features at no cost. We’ll introduce subscription tiers designed to suit different needs:
Our aim is to keep pricing simple and accessible, an intelligent add-on to your workflow than a heavy software commitment.
Tarsk is a quiet nod to Alfred Tarski (1901–1983), a pioneering logician and mathematician whose work still underpins much of modern computing. Tarski is best known for defining a formal notion of truth, a deceptively simple idea that turns out to be foundational for logic, semantics, and ultimately, artificial intelligence.
By separating language from meaning, and syntax from semantics, Tarski gave us the tools to reason rigorously about systems. Our platform builds on that spirit: clarity, structure, and the pursuit of meaningful truth in data.
By submitting, you agree to receive early access updates from Tarsk.ai and PPA.
Recording of interaction with Ai Agent Tarski . PGL version 0.2 . April 2025
