AI value in brief
The issue. AI investment can pursue lower cost, higher revenue or both. Those routes require different evidence, different operating choices and different tests of success.
What is driving it. Across the collection, fragmented pilots, unchanged workflows, oversight costs and pressure to show economic returns repeatedly appear as reasons value stalls before it reaches the P&L.
How the thinking is developing. The expanded set still shares a strong emphasis on organisational redesign, while differing over the value it should create. Baringa strengthens the efficiency-led case; FTI explicitly combines cost and revenue outcomes; PwC gives greater weight to growth and reinvention; Microsoft provides a narrower workflow-value counterpoint.
What we can learn
Efficiency and growth are not opposite strategies. The same redesign can lower cost, release capacity and create revenue.
A high-value use case is not yet an operating model. Scaling requires ownership, workflow change, governance and repeatable economics.
The business case should name where value will appear. Time saved, lower cost, higher throughput and new revenue are different outcomes.
A cost programme is not necessarily a modest technology upgrade. BCG and Bain connect financial impact to changes in processes and operating models; McKinsey concentrates attention on high-value domains; Baringa similarly links efficiency to redesign of the work rather than tool deployment alone.
The useful signal is that buying tools and changing the business are becoming distinct subjects in the advice. FTI, PwC and Accenture broaden that argument by making revenue expansion, new offerings and business-model value more explicit. This is convergence within this selected set, not a measurement of all advisers or enterprises.
Where the thinking differs
PwC gives greater weight to growth opportunities and business reinvention. FTI balances cost and revenue routes inside a repeatable operating model. Baringa is more disciplined around efficiency and service outcomes. Microsoft is an important counterweight: measurable value can begin with narrower workflow gains rather than enterprise-wide redesign on day one.
These positions differ in scope, sequencing and ambition. They do not establish a simple divide between disciplined investors and ambitious ones: growth ambition can coexist with proof gates, while a cost-led case can still require radical redesign.
Plot’s interpretation: The useful divide is not AI optimists versus sceptics. It is between firms treating AI as a productivity layer and firms redesigning how value is created - and whether that redesign is expected to show up mainly in margin, growth or both.
In practice
A useful leadership test is to choose one business activity and state the economic claim before the technology claim: what changes in cost, throughput, revenue or risk; what workflow has to change; and what evidence releases the next investment.
What we still don’t know
Amazon’s shareholder letter and Microsoft’s earnings material remain context rather than markers on this map. They describe supplier investment and reported company activity, which are different evidence types from an enterprise buying AI or an adviser recommending a value strategy. The current set does not establish a general adviser-versus-operator gap.
Ideas to watch
Realised economics: whether reported time savings become margin, capacity, service improvement or revenue.
Funding discipline: whether organisations keep scaling before proof or move toward staged investment gates.
Business-model evidence: repeated examples of AI creating incremental demand, not only automating existing work.
A larger article set is not evidence that institutional thinking has shifted. Plot 01’s current axes create a baseline; a movement claim needs comparable dated statements under the same definitions.
Where does your organisation sit?
Plot one business activity against the two questions: is the main ambition efficiency, growth or explicitly both - and does the approach improve existing workflows or redesign them? A reader position shows how the organisation compares with published thinking; it is not a maturity score or a recommended destination.
Follow this Plot for new evidence, revised arguments and any substantiated movement. Plot your organisation to compare its assumptions with the published thinking.
Evidence note: 18 thought pieces are plotted at high editorial confidence. Placements are editorial interpretations against the V3 rubric.
