Title

AI spend management and governance diagnostic

Section
Insight
Summary

A practical executive discussion tool for finance and technology leaders

Is your AI spend growing faster than your ability to control it?

As AI adoption accelerates, organizations face a growing challenge: understanding where AI costs are coming from, what is driving consumption, and how to keep spending under control without slowing innovation.

But how confident are you that your organization has the right visibility, forecasting, and governance in place?

Use our 15-question AI spend diagnostic to uncover potential gaps across four critical areas: spend visibility, demand management, budgeting and forecasting, and commercial governance.

Identify where your organization may be exposed to unexpected costs and discover practical steps to improve control over AI spending.

Complete the form below to access the diagnostic and start assessing your organization's AI spend management.

Spend visibility

AI costs can grow rapidly and emerge across multiple providers, applications, and business functions. Without a consolidated view of spend and its underlying drivers, organisations risk unexpected budget overruns and may struggle to identify where intervention is needed.

  • Do we have a consolidated view of AI spend across our major providers (e.g. OpenAI, Google, Anthropic)?
  • Can we attribute material AI consumption to the teams, departments, or use cases driving it?
  • Can we identify our current run rate, significant changes in consumption, and emerging sources of cost?
  • Do we have a central tracker of all third-party applications or consumption-based services that introduce additional AI costs?

Demand management

AI consumption can vary significantly between users, teams, and use cases, and higher usage does not necessarily indicate waste. Without understanding these differences, organisations risk allowing inefficient consumption to continue or imposing restrictions that undermine productivity and innovation.

  • Do we understand how AI usage varies across different user groups and use cases? 
  • Have we defined expected usage or spend ranges for materially different user profiles? 
  • Do we have a process to investigate unusual consumption and provide targeted guidance before imposing restrictions?
  • Have we established preferred AI models for different user groups or use cases, with a process for reviewing these choices as new models become available?

Budgeting and forecasting

Adoption and consumption patterns can change faster than traditional budgeting cycles can accommodate. Forecasts based on initial assumptions rather than observed demand risk understating future spend, leaving organisations unprepared for growth and potentially making supplier commitments that do not reflect actual needs.

  • Are our AI budgets and forecasts informed by observed usage rather than relying primarily on generic per-user assumptions?
  • Do we regularly compare actual spend against budgets or expected ranges and investigate material variances?
  • Do we incorporate expected adoption growth and planned AI initiatives when updating our forecasts?
  • Do budget owners have visibility into actual AI spend against their allocated budgets?

Governance and commercial management

AI spend spans multiple functions, each with only part of the picture. Without clear accountability and coordinated decision-making, organisations risk fragmented oversight, duplicated investments, and supplier agreements that offer insufficient flexibility as demand and technology evolve.

  • Are responsibilities for AI cost governance clear across Finance, IT, business owners, and Procurement?
  • Do we use demand and usage information to inform supplier commitments, contract terms, and renewal decisions?
  • Do we provide appropriate training when users gain access to AI tools, with additional guidance for those whose consumption exceeds expected levels?

What should you do next?

Use your responses to identify where your organisation has limited visibility or unclear processes. Prioritise gaps that could create unexpected spend or poorly informed supplier commitments. This is a qualitative discussion guide, not a scored maturity assessment.

  1. Establish a reliable view of spend: Bring together recent spend and usage data, budgets, and supplier commitments. Identify the current run rate, material variances, and areas of accelerating demand.
  2. Segment demand and establish expected ranges: Group users, teams, or use cases according to their consumption patterns, including the models and tools they use. Use observed demand to establish practical spend ranges and identify exceptions.
  3. Intervene and educate before restricting usage: Investigate unusually high consumption, understand the work being performed, and help users adopt more efficient tools or behaviours. Protect productive power users where higher demand is justified.
  4. Build a more realistic AI forecast: Use actual consumption, user profiles, and planned adoption to forecast demand. Monitor variances and revise assumptions regularly.
  5. Align supplier commitments with demand: Use the improved baseline to review commitment volumes, rollover provisions, overage protection, contract duration, and flexibility.

Questions for your next AI governance meeting

  • Where are our greatest gaps in visibility, forecasting, or accountability?
  • Does our reporting provide enough insight to recommend practical changes to how AI is being used?
  • Which areas of AI consumption warrant investigation, and who will lead it?