Questions Your IT Team Will Ask
Security, data residency, permissions and rollout — answered plainly.
Can Beacon modify or corrupt our production data?
No. Beacon is read-only by default. It never connects to your production systems directly. It reads from a replica of your data, and credentials are scoped to that replica only. Writes are only possible if you enable the action module, and even then, every write goes through your own approved APIs with validation and a full audit trail. Beacon never writes directly to a source database.
Where does our data live?
On your premises by default. Your production systems, your replica, and your identity server all stay inside your network. Specific AI processing runs in a pinned cloud region, encrypted in transit, and every component that touches your data stays in that region. Nothing inbound: the connector dials out from your network. No inbound ports are opened.
Does Treel see our user passwords or credentials?
No. Sign-in runs through your own company login. Your OAuth2 server stays inside your network, and credentials are verified there. They are never sent to Treel. What Treel receives is a signed session scoped to that user's role, level, and market.
How do you prevent the AI from making up numbers?
Lizzie never computes figures herself. Every number in every answer is calculated programmatically from your data and inserted into the response. You can trace any value to see exactly how it was derived. On top of that, our Verification Suite breaks every answer into individual claims and checks each one against your actual data before it reaches the user.
Can a user see data they're not supposed to see?
No. Role-based permissions are enforced twice: once at the connector and again in the semantic layer. Rank, level, and market decide which rows each user sees. Asking Lizzie a question differently does not data outside a user's approved scope. There is no prompt that bypasses permissions.
What about our external partners and distributors?
They log into the same platform with their own scoped access. They see only what you approve. Corporate can lock their interface, curate their widgets, and define exactly which data points and actions are exposed. Nothing unapproved ever reaches an outside audience.
How long does integration take, and what do you need from our IT team?
Full integration takes under two months, with Beacon typically live sooner than that. Your team provisions the on-premise infrastructure and grants API access to your data sources. We handle the rest: configuration, integration, fine-tuning, and ongoing maintenance.
And those weeks are not just technical work. During integration, we sit with your teams, learn what each role actually needs, and build the initial views for them. Your CFO, your ops lead, and your partners each start with a working back office tailored to their job, not a blank screen. We teach your people how to adjust and extend it along the way. When we leave, you have a running solution and the knowledge to reshape it yourselves. Nobody spends their first month building Beacon instead of using it.
Does Beacon replace our existing systems?
No. Beacon sits on top of your existing stack. Your ERP, e-commerce platform, CRM, and finance tools keep running exactly as they do today. Beacon connects to them, unifies the data, and gives your people one place to see and query everything.
How is Beacon different from a BI dashboard?
A BI dashboard answers the questions someone predicted months ago. A developer builds it, and from that day it shows the same charts to everyone, whether they need them or not. New question? New ticket. New dashboard. More waiting.
Beacon has no fixed set of questions. You ask anything about your business in plain language and get an answer in seconds. Useful answers become widgets. Widgets become pages. Your back office grows out of the questions your people actually ask, not the ones a developer guessed at. And Beacon works on its own: it monitors your data continuously and surfaces anomalies, trends, and risks before anyone thinks to ask. No BI dashboard does that.
How is Beacon different from ChatGPT, Claude, or Gemini?
Those are general-purpose models. They don't know your business, and connecting them to your data yourself means heavy manual work to maintain, tune, and scale, with no practical way to enforce permissions across hundreds of internal and external users. And when they answer data-specific questions, they compute the figures themselves, confidently, and sometimes wrong.
Beacon is a complete back office built around Lizzie, our proprietary AI, fine-tuned to your business during integration. Every number is computed programmatically, never generated by the model. Every answer passes the Verification Suite before a user sees it. Every user gets role-based access, including outside partners. And it's proactive: it watches your business instead of waiting for a prompt. A general-purpose LLM gives you a chat window. Beacon gives you an operating layer.
Do we need to train our employees to use it?
No formal training required. Users talk to Lizzie in plain language. They ask questions, get answers, and pin the ones worth keeping as widgets. Building dashboards, pages, and custom views happens through conversation. If your team can ask a question, they can use Beacon.
What happens to our dashboards and data if we leave?
Your data never stopped being yours. It lives in your systems and your replica, on your infrastructure. Beacon reads it. If you leave, you keep everything you had before Beacon, and your data can be exported in standard formats at any time.