Procurement teams evaluating an on-prem analytics platform usually arrive with two questions that look simple but rarely are: what does it cost, and what will it actually cost over the life of the deployment. Those are different questions. The first is a line item. The second is total cost of ownership — the figure that determines whether a platform is a good decision three years from now. This article explains how AI Insights 365 On-Prem Edition is licensed, why the per-user, multi-year model is structured the way it is, how to size the number of users you need, and how to think about TCO when comparing a self-hosted platform against a cloud subscription.
We will not quote dollar figures here. Final pricing is generated by sales through a payment link, because it depends on your specific user count and term. What we can do is explain the model precisely enough that you can estimate confidently and request an accurate quote.
The Licensing Model in One Paragraph
AI Insights 365 On-Prem is licensed per user on a multi-year term — you choose 1, 2, or 3 years. Pricing scales with the number of users multiplied by the term length, and longer terms lower the effective per-year cost per user. You submit a license request specifying users and term, sales generates a payment link, and you receive a signed license file that you activate on your own Windows servers. That file encodes your seat count and expiry; the platform enforces both locally, with no runtime phone-home.
That is the whole commercial shape. The rest of this article is about using it well.
Why Per-User Licensing
Per-user (named-user) licensing aligns cost with the people who actually derive value from the platform. For an analytics product, that alignment is natural: value accrues to the analysts, engineers, and decision-makers who build and consume reports, not to the number of CPU cores or the volume of rows scanned.
Per-user licensing has practical advantages for buyers:
- Predictable. Your cost is a function of headcount, which you can forecast, rather than consumption, which spikes unpredictably.
- Budget-friendly for data-heavy workloads. Because you are not billed per query, per gigabyte, or per refresh, heavy use by a fixed team does not inflate the bill.
- Easy to govern. Seat count maps directly to your role-based access controls. Each named user is a real person you can audit.
Because the platform supports RBAC with Admin, Editor, and Viewer roles, you can give a large reader population access without every person needing to be a heavy author. That distinction matters when you size your seats, which we cover below.
Why Multi-Year Lowers Effective Cost
The platform is sold on 1-, 2-, or 3-year terms, and the longer the term, the lower the effective per-year cost of each seat. The mechanism is straightforward: committing to a longer term reduces renewal overhead and rewards the commitment with better per-year economics.
Think of it the way you would any committed-term purchase. A one-year term carries the highest per-year rate and the most frequent renewal cycle. A three-year term carries the lowest effective per-year rate and a single, predictable renewal horizon. The two-year term sits in between.
You can reason about it with a simple identity, without needing any actual numbers:
effective_per_year = total_license_cost / term_years
For a fixed user count:
3-year effective_per_year < 2-year effective_per_year < 1-year effective_per_year
Multi-year terms also reduce a cost that rarely shows up on a quote: the internal cost of re-procuring. Every renewal consumes time from procurement, finance, security review, and IT. A three-year term collapses three annual cycles into one, and that administrative saving is real even though it does not appear as a discount line.
For air-gapped and regulated environments there is an additional benefit. Because activation and renewal use a signed license file applied live in System Settings — with no restart and no outbound call — a longer term simply means fewer of those controlled change events to schedule and document.
What's Included
The license entitles your named users to the platform's full analytics capability set, deployed on your infrastructure. Depending on your edition, that includes:
- 30+ data connectors — SQL Server, PostgreSQL, Snowflake, Power BI, flat files, and more.
- Direct in-network connectivity for reaching databases behind your firewall without exposing them.
- AI engine powered by Azure AI Foundry or, for fully isolated sites, local air-gapped Ollama.
- Natural-language-to-SQL and AI auto-reports so users can ask questions in plain English.
- 72+ chart types and a no-code transform layer for shaping data without scripts.
- Scheduled report email for recurring distribution.
- Multi-tenant workspaces with RBAC (Admin / Editor / Viewer).
- Active Directory + Entra ID SSO, audit logs, AES-256 encryption at rest, and org/workspace backup & restore.
The license governs how many named users can access this capability; it does not meter your queries, your data volume, or your refresh frequency.
How to Size the Number of Users You Need
Sizing seats is the single most important input to your quote, and it is easy to get wrong in both directions. Over-buy and you pay for empty seats; under-buy and you hit the seat ceiling and have to amend mid-term. Here is a practical method.
1. Count by role, not by department
Map your prospective users to the three RBAC roles:
- Admin — platform and workspace administrators, security owners. Usually a small number.
- Editor — analysts, data engineers, and report authors who build content.
- Viewer — the larger population that consumes dashboards and scheduled reports.
Every named user counts as a seat regardless of role, so the goal of this exercise is accuracy, not role-based discounting. The role breakdown matters because it tells you how your population will actually grow.
2. Separate "active now" from "next 12 months"
List the users who need access at go-live, then add realistic growth for the coming year. Because seats can be increased mid-term by applying a new signed license live, you do not need to over-provision wildly — but matching your term-length headcount to your hiring plan avoids frequent amendments.
3. Account for shared and service scenarios
Decide how you will handle dashboards consumed by many occasional viewers. If a broad audience needs read access, count those viewers as seats. Keep service and integration accounts to a deliberate minimum and document them, since each named account consumes a seat.
4. Pick a term that matches your planning horizon
If your headcount is stable and your budget favors a single commitment, a 3-year term gives the lowest effective per-year cost. If you expect significant change, a shorter term preserves flexibility at a higher per-year rate. You can always increase seats within a term.
A simple sizing worksheet:
| Input | Your number |
|---|---|
| Admin users at go-live | |
| Editor users at go-live | |
| Viewer users at go-live | |
| Expected new users over the term | |
| Service / shared accounts | |
| Total seats to license |
Carry the total into the pricing calculator, choose your term, and request a quote.
Subscription-Cloud vs. On-Prem TCO
The headline price is only part of the comparison. To compare a cloud BI subscription against AI Insights 365 On-Prem fairly, model total cost of ownership across the same multi-year window you would commit to for the on-prem term.
What a cloud subscription tends to include and cost
- A recurring per-user (or per-capacity) fee that typically renews annually and can escalate.
- Consumption-style charges in some models — per query, per refresh, per gigabyte — that scale with usage.
- Egress and integration costs for moving your data to and from the vendor's cloud.
- Limited control over where data physically resides, which can carry compliance overhead.
What on-prem shifts
- Licensing becomes a committed per-user, multi-year cost with a lower effective per-year rate on longer terms — and no usage metering.
- Infrastructure is yours: you run it on existing Windows servers, so you reuse capacity you may already own and operate.
- Data stays in place. Because the platform runs inside your network — and can use local air-gapped Ollama for AI — you avoid egress charges and reduce data-residency risk.
- Operations are your responsibility, but the features that usually drive cloud "ops premiums" (backup & restore, SSO, audit logs, encryption at rest) are included.
A fair multi-year comparison looks like this:
Cloud TCO (over N years) ≈ Σ (annual per-user fees, often escalating)
+ Σ (consumption / query / refresh charges)
+ Σ (egress + data movement)
+ compliance overhead of off-prem data
On-Prem TCO (over N years) ≈ committed per-user license (term = N, lower effective/year)
+ infrastructure you operate (often already owned)
+ internal ops time
- no per-usage metering, no egress
The qualitative result for most data-heavy, security-conscious organizations: a cloud subscription is attractive when usage is light and headcount is volatile, while an on-prem, per-user, multi-year license tends to win on TCO when usage is heavy, headcount is stable, and data residency matters. The absence of consumption metering is frequently the deciding factor — power users do not generate surprise charges.
How to Request a License or Quote
You do not need to negotiate blind. The AI Insights 365 site provides two self-service tools:
- A pricing calculator — enter your user count and chosen term (1, 2, or 3 years) to see how the model scales and how a longer term lowers the effective per-year cost.
- A license request form — submit your sizing and term to start the quote.
From there the flow is simple:
- Use the calculator to model a few combinations of seats and term.
- Submit the license request form with your chosen configuration.
- Sales generates a payment link priced to your users × term.
- On payment, you receive a signed license file.
- Activate it on your servers — online or fully air-gapped — and manage renewals live in System Settings.
Bringing It Together
On-prem licensing for AI Insights 365 is built around three ideas that procurement and security teams can both endorse. It is per user, so cost tracks the people who get value. It is multi-year, so committing for two or three years lowers your effective per-year cost and collapses renewal overhead. And it is self-contained, so the signed license you activate carries your entitlement without metering your usage or calling home.
To turn that model into a number, size your seats by role and growth, choose the term that matches your planning horizon, run the figures through the pricing calculator, and submit the license request form. The result is a TCO you can defend — one where you pay for the team that uses the platform, keep your data on your own infrastructure, and avoid the consumption surprises that make cloud analytics bills so hard to forecast.