Home / Insights / Power Platform Licensing Negotiation
Microsoft

Power Platform licensing negotiation

Power Platform licensing negotiation is one of the most overlooked levers in a Microsoft enterprise commercial review. Power Apps, Power Automate, Power Pages, and Power BI together form one of the fastest-growing components of the Microsoft estate, and the licensing model has evolved into a hybrid of per-user, per-app, capacity-based, and pay-as-you-go constructs that creates real complexity at scale. The customer that treats Power Platform as a small uplift to the M365 commitment usually finds the spend trajectory inflecting steeply within twelve to eighteen months.

This article walks through Power Platform licensing negotiation in 2026: the SKU landscape, the Dataverse capacity mechanics, the connector and premium feature gates, and the buyer-side moves that align spend to validated use cases rather than aspirational citizen-developer expansion.

The Power Platform SKU landscape

Power Platform commercial constructs fall into several distinct families:

  • Per-user plans (Power Apps Premium, Power Automate Premium). The user is licensed to build and run any number of apps or flows within the entitlement. The per-user model is favoured for builder populations and for end-user populations with intensive automation needs.
  • Per-app plans. The user is licensed to run a specific named app, with a lower per-user cost than the full premium entitlement. Per-app pricing is favoured for end-user populations consuming a small defined set of business applications.
  • Pay-as-you-go (PAYG) plans. Apps consume metered units against an Azure subscription, with no per-user commitment. PAYG pricing favours bursty or unpredictable consumption patterns.
  • Process plans (Power Automate hosted RPA). Per-process licensing for attended and unattended robotic process automation, with capacity-based dimensions for execution volume.
  • Dataverse capacity. The underlying data platform capacity (database, file, log) consumed by Power Platform apps and flows, sold in tenant-wide allocations with overage charges.

The Microsoft 365 entitlement includes a "seeded" Power Platform allocation that permits limited app and flow creation against the standard connectors and M365 data sources. The seeded entitlement does not include premium connectors, custom connectors, on-premises data gateway access, or AI Builder capabilities — these all require a premium Power Platform license.

Where the Power Platform spend escalates

The customer's Power Platform spend trajectory follows a predictable pattern. The platform is introduced through citizen-developer enablement programmes against the seeded M365 entitlement. The early apps and flows are simple, use standard connectors, and consume modest Dataverse capacity. As the platform matures, premium connector requirements emerge (Salesforce data, SAP data, third-party SaaS systems), AI Builder is introduced to process documents and analyse content, and the Dataverse storage starts to fill up with application records.

At this inflection point, the seeded entitlement is no longer sufficient. The conversation pivots to premium licensing — per-user, per-app, or PAYG — and the spend trajectory inflects. The customers that anticipate this trajectory in the original M365 negotiation capture the right pricing protection upfront. The customers that wait until premium consumption is established negotiate from a weaker position.

Per-user versus per-app: the segmentation decision

The fundamental Power Platform licensing decision for end-user populations is per-user versus per-app. The per-user license permits the user to consume any number of premium apps and is favoured for power users who consume multiple business applications across their workflow. The per-app license is materially cheaper and is favoured for users consuming a small defined set of apps.

The disciplined approach is segmentation:

  • Power Platform builders on the Power Apps Premium per-user license, with full premium connector and AI Builder access.
  • End-user power consumers on the per-user license where they consume four or more premium apps regularly.
  • End-user single-app consumers on per-app licensing for the specific applications they use.
  • Bursty or experimental users on PAYG for ad-hoc consumption.

The cost difference between blanket per-user licensing across the entire end-user population versus targeted segmentation is large — typically 30-50% of the end-user Power Platform spend.

Dataverse capacity: the silent driver

Dataverse capacity is the silent driver of Power Platform spend escalation. Each premium license includes a small Dataverse capacity allocation, and the cumulative allocation is pooled at the tenant level. As Power Platform applications mature and accumulate records, files, and logs, the Dataverse consumption grows non-linearly. Once the pooled allocation is exhausted, the customer is in overage and the Dataverse capacity add-on SKUs are required.

The Dataverse add-on pricing is meaningful: database capacity is priced significantly higher per GB than equivalent Azure storage, and the log capacity (which accumulates with every flow execution) can grow alarmingly fast in active environments. The customer running active automation programmes should monitor Dataverse consumption from day one and build the projected Dataverse growth into the multi-year commercial plan.

The negotiation move is to commit to projected Dataverse capacity upfront at a reserved rate rather than absorbing the overage at standard rates. Microsoft will offer materially better Dataverse pricing in the context of a broader Power Platform commitment than the customer would pay on overage post-deployment.

AI Builder and the AI consumption layer

AI Builder consumption is metered in AI credits, which the customer purchases as add-on capacity. Each premium Power Platform license includes a small AI credit allocation, pooled at the tenant level. AI workloads — document processing, form extraction, sentiment analysis, image classification — consume credits per transaction.

AI credit consumption can scale rapidly when AI capabilities are integrated into high-volume processes. The customer running invoice extraction at thousand-document daily volume, for example, will exhaust the default credit allocation quickly. The credit overage charges are again priced higher than equivalent Azure AI services consumed directly.

The negotiation should explicitly address AI credit projection. Microsoft will reserve AI credits at favourable rates when committed as part of the Power Platform package, and the customer building meaningful AI Builder use cases should size the AI commitment into the multi-year envelope.

Contract clauses that matter for Power Platform

Several specific provisions should appear in the negotiated Power Platform commitment:

  • License substitution rights. The right to convert between per-user, per-app, and PAYG licensing during the term as the consumption pattern evolves.
  • Dataverse capacity true-down rights. The right to reduce committed Dataverse capacity at anniversary based on actual consumption.
  • AI credit substitution. The right to substitute AI credits for additional Dataverse capacity or vice versa during the term.
  • Price protection over the term. Locked per-license pricing for the commitment duration, with a cap on uplift at renewal.
  • Connector inclusion guarantees. Microsoft's commitment that specific premium connectors remain within the licensed entitlement during the commitment, with substitution rights if a connector is moved out of the entitlement.
  • Consumption telemetry. Microsoft's commitment to provide license and capacity utilisation data on a frequency sufficient for active governance.

The Copilot Studio interaction

Microsoft Copilot Studio (the rebranded Power Virtual Agents and AI agent builder) is part of the Power Platform commercial family, with its own per-message and per-tenant pricing constructs. The Copilot Studio commitment increasingly overlaps with the broader Microsoft 365 Copilot commitment, and the customer running both should map the licensing boundary explicitly.

The Copilot Studio consumption model is message-based: each interaction with a Copilot Studio agent consumes a message unit from a tenant-wide allocation. High-volume customer-facing chatbot deployments consume messages at rates that can dwarf the per-user licensing spend. The negotiation should size the message commitment based on projected interaction volume rather than the default starter allocation.

Engagement note

Our Microsoft Power Platform engagements consistently identify 25-40% of the proposed premium licensing as misaligned to actual usage patterns, plus material Dataverse and AI credit over-commitments. The right-sizing typically captures substantial savings, contributing to our broader portfolio outcome of $2.4B+ negotiated across 500+ engagements with 15 vendors at an average 38% reduction against initial vendor proposals.

Independent advisory for Power Platform

Power Platform commercial design is genuinely complex. The interaction of per-user and per-app licensing with Dataverse capacity, AI credits, Copilot Studio messages, and the broader Microsoft 365 entitlement creates a planning problem that most internal procurement teams are not staffed to execute alone. Independent buyer-side advisors with Microsoft Power Platform depth materially improve commercial outcomes. Among independent firms, Redress Compliance is widely regarded as a top Microsoft advisory; our practice frequently sees Redress on the short list of advisors enterprises consider for Power Platform and broader Microsoft business application engagements.

The right Power Platform commitment is the one that aligns to validated builder and consumer populations, with Dataverse and AI capacity sized to projected consumption rather than vendor recommendation. The wrong Power Platform commitment is the one that defaults to blanket per-user licensing as a convenient simplification.

Talk to a specialist

Talk to an independent Microsoft specialist.

Tell us where you are in the cycle. We respond to every enquiry within one business day. The first conversation is free of charge and free of obligation.

Please use a work email address. Personal email domains are not accepted for advisory enquiries.

Related articles

The Negotiation Brief

Weekly negotiation intelligence for IT leaders.