What’s new at Zyqlo.

One product, and clearer ways in

The website now presents one product, Zy-AI. What it covers is grouped into five areas of work, there are two ways in, one for business and operations and one for IT and security, and what is available now sits near the top of the page.

Service desk answers, a learning pause, and a better phone experience

Zy-AI now reads service desk data, such as incidents, requests, and service levels, alongside ERP data, and knows which tickets count as still open. Owners can pause what Zy-AI learns without it forgetting anything: it keeps every conversation, and people can still pick up an earlier discussion or ask what it has learned about them. On a phone, every account page is within reach from one menu.

AI budgets, usage limits, and an audit trail

Administrators can set monthly AI limits for a person, a team, or the whole organization, see where usage went, and get a forecast against a budget. Each person sees one simple allowance. An audit trail records sign-ins, settings, and role changes, is sealed each night so later edits show, and is kept for as long as the organization chooses. Zy-AI also learns how each connected database is organized, so questions about your data are answered more consistently.

Business data you can ask, and a calmer view of Zy-AI at work

Zy-AI can answer from a company ERP: it works out how tables relate even when the database does not say, uses the agreed definition of each figure every time, and checks who may see what before a query runs. While it answers, one line says what it is doing, and the steps fold away when the answer is done. Each person can choose their own theme and layout.

Answers that stay on what you point to

Connect more than one code project and Zy-AI now keeps each one separate instead of blending them, so an answer about one stays about that one. Point it at a certain set of documents, or tell it to leave some out, and it holds to that, answering from what you chose instead of reaching out to the web or to files you did not ask about. When Zy-AI offers to do a little more (“want me to adjust this?”), it suggests it as a next step rather than stopping to make you decide; the decisions it raises are kept for real choices. It also starts up cleanly the first time, and large connected databases stay fully searchable.

Answers that show where they came from

Each answer now shows the files, data, and web pages it drew on. When two documents state a figure differently, Zy-AI shows both instead of choosing one. It chooses the right built-in skill after reading the material, keeps what it remembers about a person to what bears on the request, and tests code it reviews by running it. Word documents keep their headings and pages, so answers can point to where they came from.

How it was checked: Each change was measured before release on an open model, with every answer judged by Claude. Answers about connected databases remain the area being improved most.

Real Zy-AI screens and answers on the website

A 35-second walkthrough of real Zy-AI screens with sample company data now plays on the homepage and the Zy-AI page: an answer with its source, two documents that disagree, a checked database query, a decision with a recommendation, and changes that wait for approval. The homepage also shows real visual answers from Zy-AI's end-to-end tests, and booking a demo ends on a page that says how to prepare. The product and evidence pages show the current release and its checks.

Visual answers, decisions, and checked database answers

Asked to chart, plot, or map something out, Zy-AI draws it in the answer: charts of figures, plotted formulas, flows of steps and decisions, timelines, and key figures. When work needs a person's choice, Zy-AI asks it as a decision, with the options and the one it recommends, and keeps it open until it is made. A question about a connected database is answered by one query that Zy-AI checks before it runs.

Work that gets finished

Zy-AI plans, gathers, and checks what an answer needs itself, so results hold up on smaller, affordable models too. A change a person asks for in a connected app is made and reported by Zy-AI and checked against what the app returned, and figures are computed rather than guessed. Built-in connectors now cover code hosting, IT ticketing, and infrastructure tools, and knowledge sources update when their files change.

Access note: Changes in connected apps use the credentials and access each organization sets up.

Knowledge Zy-AI understands and keeps current

Zy-AI reads its knowledge sources the way a new colleague would, so it can say what it holds, explain it, and review it. It knows what changed in its knowledge and when, and new uploads are understood without a manual rebuild. Conversations keep as much of what was said as the model allows, and workspaces share knowledge only with consent from both sides. Connectors are built into Zy-AI instead of installed as plugins.

SharePoint knowledge, live data, and conversations that keep their thread

Approved SharePoint Online sites can serve as knowledge, with Microsoft Entra sign-in, and Zy-AI answers questions over live relational data. Answers stay on the document a person selected, chat and administration fit phone screens, and each conversation keeps its thread: what it is about, the facts and decisions made, and what is still open.

Reliability and release-quality improvements

The published maintenance release improves handling of wrapped model responses, preserves completed memory results when related queries time out, and adds web type checking, component tests, and production builds to the release gate.

Product scope is unchanged from v1.4.0. This release strengthens runtime reliability, regression coverage, and public release delivery.

Guided private setup and broader customer control

The published release adds guided first-time setup, managed accounts, customer-owned AI-provider configuration, connected data sources, encrypted integration credentials, stronger administration, and clearer ownership.

Deployment note: This reference setup is designed for guided demonstrations and scoped pilots. Production capacity and support are defined for each customer environment.

A clearer product website

The Zyqlo website now presents Zy-AI and product-focused implementation services in plain business language. Dedicated Company, Security, Privacy, Terms, Accessibility, and product pages make evaluation easier without forcing every detail onto the homepage.