The questions enterprise teams ask us.
Answered plainly.
JAS and generative AI, deployment and security, data, analysis, embedding, projects and support. For step-by-step instructions, see the User Guide.
The four answers we give most often
JAS and generative AI
How the AI assistant built into TURBOARD works, where her answers come from, and how she is kept within your rules.
What is JAS, and how is she different from a chatbot added to a BI tool?
JAS (Just Ask Simply) is the generative-AI assistant built into TURBOARD. A general-purpose chatbot connected to a database has to guess what your tables and numbers mean. JAS works through what your organization has already defined and governed inside TURBOARD: the reports, the business definitions and the user permissions.
When you ask a question, she first searches the trusted reports that already exist and brings the certified answer into the conversation. When no existing report answers it, she can build a new dashlet or dashboard from the same governed model your analysts use — a native, reusable BI asset, not a screenshot.
How can we trust the answers JAS gives?
- The engine calculates, not the language model. Your question is answered with TURBOARD’s governed model, date logic, calculations and permissions — the same ones behind every dashboard.
- Every answer carries its source, so users can open the report or view behind it.
- Context is governed. JAS learns your business on five levels: database fields, governed views, personas, context generated by GenAI, and context curated by your team. Generated context is reviewed and approved by your authorized users.
- Personas change the explanation, never the number. A finance director and an operations manager can get different commentary on the same, identical figure.
Does JAS respect our data permissions?
Yes. JAS uses the same permissions as the rest of TURBOARD, including row- and column-level restrictions. She answers each user from the data that user is authorized to see — a user who cannot see a region’s rows on a dashboard cannot reach them by asking JAS either.
Which AI models does JAS use, and can the AI run fully on-premise?
You choose the model, and you can change it later:
- Commercial models such as GPT (OpenAI) and Gemini (Google).
- Open models such as DeepSeek, Gemma, Mistral and Qwen.
- Your own private LLM endpoint.
- Offline small language models (SLMs) installed on your own hardware — air-gapped if required.
With the offline option, the whole platform and the AI run inside your infrastructure: prompts and data never leave it.
What can JAS do besides answering questions?
- Explain a dashlet or a whole dashboard in plain language, tailored to the reader’s role with built-in or custom personas.
- Build dashlets and complete dashboards from a sentence — “show sales by region as a pie chart” — and edit them with commands such as “change to a bar chart and sort”. Several changes can be requested in one prompt.
- Write and validate SQL for views and calculated measures, finding the right tables and fields even in schemas with thousands of tables.
- Write Pandas functions for advanced calculations.
- Guide users through TURBOARD itself: JAS is trained on our user guide and answers how-to questions in context.
Deployment and security
Where TURBOARD runs, how users sign in, what each user may see, and where your data is kept.
Does TURBOARD run in the cloud or on-premise?
Both, with the same platform and the same features:
- Cloud — public or private cloud, including JAS with commercial or open AI models.
- Fully on-premise, AI included — the whole platform inside your walls, with JAS running on offline small language models where required.
- Hybrid — for example, the platform on-premise with an approved private model endpoint.
Most of our enterprise and public-sector customers run TURBOARD on servers in their own data centres. While much of the industry is moving customers away from on-premise, we continue to support it fully, with no end-of-support countdown.
How do users sign in? Do you support SSO and Active Directory?
TURBOARD integrates with the identity systems enterprises already use:
- LDAP / Active Directory
- Single sign-on (SSO) with your corporate identity infrastructure
- OAuth 2.0 identity providers, including national e-government login services
- Multi-factor authentication (MFA)
- External user tables — hundreds of thousands of users can be imported and kept in sync through a configurable interface
On a national health platform, more than 900 concurrent users sign in through corporate SSO at month-end peaks.
How fine-grained can permissions be?
- Row-level and column-level restrictions, so each user sees only the records and fields they are entitled to.
- Data masking for sensitive fields.
- Permissions synchronized from other applications’ databases, so you do not maintain them twice.
- Flexible grouping at user and group level.
The same rules apply everywhere: dashboards, exports, every scheduled report e-mail and every JAS answer. Single installations serve hundreds of thousands of named users, each seeing only their own data.
Does TURBOARD copy or store our data?
By default, no. TURBOARD connects live to your databases and queries the data where it is. Data is stored by TURBOARD only when you choose to:
- Use the ETL built into TURBOARD — query results are loaded into its columnar store for fast analysis.
- Upload Excel or CSV files — the file data is loaded into the same columnar store.
- Benefit from caching — frequently used results are kept in the cache so popular dashboards open quickly.
In an on-premise installation, all of this stays on your own servers.
How do we move content between test and production environments?
Dashboards, views and related content are moved between environments as .tur packages. Administrators manage the installation from a dedicated system console.
Data and connectivity
Connecting sources, modelling real enterprise data and keeping performance high.
Which data sources can TURBOARD work with?
- Live connections to relational databases — pick tables and TURBOARD writes the SQL, or use your own SQL.
- Excel and CSV files uploaded by users.
- Native Global Temporary Table (GTT) integration for database-side processing patterns.
- SAP — see the dedicated question below.
If you have a specific database or version in mind, ask us and we will confirm the connection method.
Can we load Excel files?
Yes. Users can create data views by uploading .xlsx and .csv files. To refresh the data, upload the new file again: as long as its structure stays the same (the file name and the column names), every view, dashlet and dashboard built on it updates automatically.
Can one analysis combine several tables or several fact tables?
Yes. A data view can combine different tables by relating them to each other. Beyond that, TURBOARD models constellation schemas as they are: several fact tables — sales, returns, stock, targets — sharing the same dimensions such as date, product and region. One dashboard can combine them correctly without flattening everything into a single star schema first.
Do we need a separate ETL tool?
For most BI needs, no. ETL is built into TURBOARD: run a query against your source and load the results into the columnar store. Delta indexing refreshes only what has changed. There is no separate tool to buy, schedule and monitor.
Can TURBOARD connect to SAP?
SAP HANA: yes — TURBOARD connects to it like any other database.
Older SAP systems running on a traditional database are harder for any BI tool: direct database access usually requires an SAP license, and the data spans tens of thousands of tables whose meaning takes SAP expertise to interpret. We typically recommend one of three routes:
- A live SAP driver (for example, from CData): BI queries are sent directly to SAP.
- Scheduled extraction: a driver transfers data hourly or daily — well suited to routine reporting.
- File exports: data exported from SAP to Excel and loaded into TURBOARD with a simple script.
We help you choose based on how fresh the data must be and on your SAP licensing.
How does TURBOARD stay fast with large data volumes and many users?
TURBOARD does not need to load everything into memory before anyone can ask a question. Each request goes where it is served best:
- Cache for frequently used results;
- Columnar store for fast scans and less frequent requests;
- Live source when data must be current to the second.
Filter values are held in an in-memory cache, so type-ahead search does not load your database. Sorting and aggregation go to the database only when needed: for example, reordering a pivot table that already holds all its dimensions happens in the browser. Hardware grows with real usage, not with data volume alone.
Dashboards and analysis
Filtering, drilling, comparing and calculating — with or without code.
How flexible is filtering?
Filtering has been at the heart of TURBOARD since our first product in 2004, which was built around fast type-ahead search: type the name of a product, a customer or a physician and go straight to that record’s report.
- Filters at dashboard and dashlet level, easily combined.
- Almost any visual can act as a filter for the others (a slicer) — slice and dice without pre-processing.
- Custom filters written in SQL when the field you need does not exist in the database.
- Filter values are cached in memory, so rich filtering does not slow the system down.
Can we drill down to the underlying records?
Yes. Drill-down and drill-through are simple to configure and work on charts, pivot tables and maps. You can go from a grouped total down to record level — for example, from a patient group to the individual patients in it — always within the user’s permissions.
Can we compare periods and scenarios without writing formulas?
- Vis-a-Vis date comparison lines up comparable periods for you — month-to-date against the same days of last month and the same period last year — as a setting, not a formula project.
- Relative dates such as this month or last 12 months keep dashboards current without editing them.
- Patented side-by-side dashboard cloning: clone the dashboard you are looking at, change its filters and read both versions side by side. This technology is protected by a granted US patent.
How are calculations beyond drag-and-drop handled? Do we need a proprietary language?
Most users never write code in TURBOARD: they drag, drop and configure. When a calculation needs more, you use the languages data teams already use — not a vendor-specific expression language:
- SQL expressions create virtual dimensions and measures without changing the source database.
- Python Pandas powers calculated measures for logic that is awkward in SQL.
- JAS can write and validate this SQL and Pandas for you from a plain-language description.
What your team writes stays plain SQL and Python — transferable skills, not lock-in.
Can a team build and edit dashboards together?
Yes. Users with designer (edit) rights can design, build, edit and share reports and dashboards together, while viewers see the published result. Content can be organized into showcases for each audience.
What mapping and GIS options are available?
- Regional maps (SVG) for countries, regions and cities. You can define custom regions — such as sales or production regions — by grouping cities with “Synonyms” under Administration › Maps.
- New regional maps can be produced by converting SHP or SVG files to GeoJSON.
- Spatial data in Oracle stored as
SDO_GEOMETRYcan be visualized directly, without converting to latitude and longitude. - With the GIS-S add-on, data is layered on Esri and other GIS basemaps, with map-optimized visualizations.
Projects, support and the company
What a TURBOARD project looks like, who supports you, and who we are.
What are the steps of a typical TURBOARD project?
- Requirements and prioritiesThe questions the business needs answered, and in what order.
- Data preparationConnecting sources, modelling, and migrating content from a previous tool where needed.
- Deployment planningCloud or on-premise, security, sign-in and permissions.
- Dashboards and reportsBuilt, reviewed and approved with their users.
- Go-liveFollowed by continuous support and monitoring.
Depending on scope, we provide a business analyst, data warehouse consultant, BI developer, project manager and TURBOARD trainer, on an hourly or daily basis.
How does support work?
Your requests are handled by BI professionals who know your setup, together with our QA and business development teams; when needed, they bring in our engineers directly. Our support hours cover Europe, the Middle East and the US East Coast.
Customer requests feed our product roadmap: we ship monthly releases and publish every change.
Will TURBOARD increase our dependency on IT?
It should reduce it. Business users build their own reports and dashboards for ad-hoc needs with self-service tools — and increasingly simply ask JAS. IT keeps control of what matters: the governed data model, security and permissions, instead of a queue of report requests.
Who is behind TURBOARD?
TURBOARD was founded in Chicago in 2004 (as E-Kalite), and its first product was built for US pharmaceutical data. Our first client, PharmaCircle, is still a client today; through PharmaCircle, 79% of large pharma companies use our visualization engine.
- TURBOARD UK, London — commercial headquarters for EMEA.
- TURBOARD Inc., Delaware — the Americas.
- Engineering and R&D in Ankara since 2004.
We are privately held, founder-led and have been profitable for two decades. A team of 79 people serves more than 500,000 business users, from enterprises to UNDP, national ministries and defence organizations. Our technology is patented, including a granted US patent.
How can we evaluate TURBOARD?
Bring one real business question. We demonstrate TURBOARD and JAS against your reports, your definitions and your security context — not a generic sample database. If you are comparing platforms, our comparison page answers the questions evaluators raise most often, including total cost of ownership.