What separates the deployments that work
The largest gap between AI leaders and everyone else is not technology. It is having decided what to build.
The short answer
What does the research show about What separates the deployments that work?
Evidence
- Cisco: 95% of Pacesetters track the impact of their AI investments, characterized as three times the rate of other organizations.
- BCG: agents were 17% of total AI value in 2025 and are expected to reach 29% by 2028. 70% of AI's potential value sits in core business functions including sales and marketing, against 13% in IT.
- BCG: more than 60% of future-built firms rigorously track AI value, against 17% of stagnating firms.
Supporting pages
- The Signal Stack: What Common Room Buyers Learn After Go-Live analysis
- The AI Revenue Team Org Chart: Which Roles Change, Which Appear, Which Go analysis
- Dictionary plain-language definitions
This is the most useful theme for an operator, and it has the best-quantified evidence in the library. Across Cisco, McKinsey, BCG, Deloitte, Scale VP and one field experiment, the same variables keep separating outcomes: redesigned workflows, finalized use cases, defined metrics, and a human kept in the loop.
McKinsey tested 25 organizational attributes and found fundamentally redesigning workflows had the largest effect on EBIT impact. At the time only 21% had redesigned any workflow. In the 2026 wave, roughly three-quarters of high performers had, against roughly one-quarter of everyone else.
What this page is
The organizational variables that separate AI deployments producing measured financial impact from deployments that do not, drawn from six independent sources.
The argument
The separating variable is not model choice, budget, or vendor. It is whether the organization decided what to build and redesigned the work around it.
How to read it
- Most of these sources are self-reported executive surveys. They are consistent with each other, which is evidence, and they share a common bias, which is a limit.
- High performer is defined differently by each publisher. Read the definitions before comparing shares across sources.
- Correlation between workflow redesign and EBIT impact does not prove direction. Well-run companies redesign workflows and also capture value for other reasons.
The largest gap between AI leaders and everyone else is not technology. It is having decided what to build.
Pacesetters are 13% of the sample, and the classification is the publisher's.
- AI use cases already finalizedStrategy+59 pts
18% → 77%
- Change-management planProcess+56 pts
35% → 91%
- Networks fully flexible, scale instantlyInfrastructure+56 pts
15% → 71%
- AI is top investment priorityStrategy+55 pts
24% → 79%
- Short- and long-term AI funding strategiesStrategy+53 pts
43% → 96%
- Networks designed for AI growth and scaleInfrastructure+52 pts
46% → 98%
- Mature repeatable process for scaling use casesProcess+49 pts
13% → 62%
- Highly aware of AI-specific threatsSecurity+45 pts
42% → 87%
- Fully equipped to control and secure AI agentsSecurity+44 pts
31% → 75%
- Defined AI roadmapStrategy+41 pts
58% → 99%
- Investing in new data-center capacity within 12 monthsInfrastructure+34 pts
43% → 77%
- AI integrated into security and identity systemsSecurity+33 pts
29% → 62%
What this does not say
Pacesetter status is a readiness score, not a measured revenue result. The two are correlated in this sample and neither is shown to cause the other.
- Publisher
- Cisco AI Readiness Index 2025 · Vendor research
- Sample and method
- n=8,000+ senior business leaders. Pacesetters are 13% of the sample
- Field dates
- Not published by the source.
What is actually quantified to move the outcome.
The units differ by row, so each row states its own unit rather than sharing one axis.
Fundamentally redesign the workflow
Held by 21% of organizations using generative AI. McKinsey, n=1,491, fielded July 16 to 31, 2024
Highest-impact attribute of 25 tested
High confidence
Keep a human in the loop, novice workers
Field experiment, staggered rollout to 5,179 support agents. +14% on average. QJE 2025, 140(2), 889 to 942
+34% issues resolved per hour
High confidence
Buy from specialist vendors
MIT NANDA. Preliminary, not peer-reviewed, primary PDF reachable only via a third-party mirror
~67% success against internal builds at about one-third the rate
Low confidence
Build, inside GTM specifically
Scale VP, n=278. This points the opposite way to MIT NANDA and both are shown
~35% more likely to see high near-term impact
Medium confidence
Move to second-phase use cases, marketing
Scale VP, n=278 GTM leaders
3 to 5x more likely to improve pipeline and conversion
Medium confidence
Move to second-phase use cases, sales
Scale VP, n=278. 0% of teams not pursuing the specified SDR use cases increased quota achievement
3x more likely to lift win rate
Medium confidence
Hire a GTM engineer
Scale VP, n=278 GTM leaders
30% more likely to realize high impact
Medium confidence
Involve RevOps
Scale VP, n=278 GTM leaders
20% more likely to achieve meaningful impact
Medium confidence
CEO owns AI governance
Held by 28%. Board oversight at 17%. McKinsey, n=1,491
Among the highest-correlated elements with EBIT impact
Medium confidence
What this does not say
Rows are not comparable to each other. Units, samples, and field windows differ by row.
- Publisher
- McKinsey, Scale VP, MIT NANDA, and Brynjolfsson, Li and Raymond
- Sample and method
- Per-row publisher, sample and field dates stated on each row. Units differ by row
- Field dates
- Not published by the source.
Leaders say they are ready on strategy and unready on the workforce.
Share saying prepared or highly prepared, by dimension.
Share saying prepared or highly prepared
- Vision and strategy52%
- Technology infrastructure48%
- Data foundation42%
- Risk, security, and governance39%
- Ecosystem partnerships34%
- Workforce25%
What this does not say
Readiness is self-assessed. The chart measures what leaders believe about their own organizations.
- Publisher
- Deloitte
- Sample and method
- n=501 US leaders
- Field dates
- April to June 2026
Share saying prepared or highly prepared, by dimension.
| Share saying prepared or highly prepared | Value (%) | Note |
|---|---|---|
| Vision and strategy | 52 | |
| Technology infrastructure | 48 | |
| Data foundation | 42 | |
| Risk, security, and governance | 39 | |
| Ecosystem partnerships | 34 | |
| Workforce | 25 |
Source: Deloitte. n=501 US leaders Fielded April to June 2026. Confidence: Medium.
What this does not say: Readiness is self-assessed. The chart measures what leaders believe about their own organizations.
5% of companies are capturing the outcome. 60% are not.
Assessed against 41 foundational capabilities across nine industries.
Share of the sample
- Future-built: 1.7x revenue growth, 3.6x three-year TSR5%
- Scalers: beginning to generate value35%
- Laggards: minimal revenue and cost gains60%
What this does not say
The 1.7x and 3.6x figures are associations between capability maturity and financial performance. Neither is a measured causal effect of AI.
- Publisher
- BCG, The Widening AI Value Gap: Build for the Future 2025
- Sample and method
- n=1,250 senior executives and AI decision makers
- Field dates
- Not published by the source.
Assessed against 41 foundational capabilities across nine industries.
| Share of the sample | Value (%) | Note |
|---|---|---|
| Future-built: 1.7x revenue growth, 3.6x three-year TSR | 5 | |
| Scalers: beginning to generate value | 35 | |
| Laggards: minimal revenue and cost gains | 60 |
Source: BCG, The Widening AI Value Gap: Build for the Future 2025. n=1,250 senior executives and AI decision makers Confidence: Medium.
What this does not say: The 1.7x and 3.6x figures are associations between capability maturity and financial performance. Neither is a measured causal effect of AI.
Also in the record
Figures that sit alongside these charts.
- Cisco: 95% of Pacesetters track the impact of their AI investments, characterized as three times the rate of other organizations.
- BCG: agents were 17% of total AI value in 2025 and are expected to reach 29% by 2028. 70% of AI's potential value sits in core business functions including sales and marketing, against 13% in IT.
- BCG: more than 60% of future-built firms rigorously track AI value, against 17% of stagnating firms.
- Deloitte: 5% say business processes are highly prepared for AI agents, while 74% expect nearly half of business processes to be redesigned around agents within four years.
- MIT NANDA found the largest measured returns in back-office work while more than half of generative AI budgets went to sales and marketing.
The brief
What is going on here, and why it matters.
The charts above are the evidence. This is the read: what the data describes, the mechanism behind it, where the argument could be wrong, and what a revenue team does about it.
The one variable that keeps winning
McKinsey tested 25 organizational attributes against EBIT impact. Fundamentally redesigning workflows had the largest effect. At the time, 21 percent had redesigned any workflow. In the 2026 wave, roughly three-quarters of high performers had done so against roughly one-quarter of everyone else. That is the widest and most repeated separation in this library.
The reason is mechanical. A model inserted into an unchanged process can only make an existing step faster. Value shows up when a step is removed, a handoff is deleted, or a decision moves earlier. None of those are purchases. All of them are org design.
Measurement is the second variable
Cisco found 95 percent of Pacesetters track the impact of their AI investments, characterized as three times the rate of other organizations. BCG found more than 60 percent of future-built firms rigorously track AI value against 17 percent of stagnating firms. Two different publishers, two different definitions, the same direction and roughly the same multiple.
Measurement is not a reporting nicety here. It is the mechanism that lets a team kill the things that do not work fast enough to fund the things that do. Organizations without it fund everything at a low level forever, which is the observed default.
Where the value actually sits
BCG puts 70 percent of AI's potential value in core business functions including sales and marketing, against 13 percent in IT, and expects agents to move from 17 percent of total AI value in 2025 to 29 percent by 2028. MIT NANDA found the largest measured returns in back-office work while more than half of generative AI budgets went to sales and marketing.
Those two statements are in tension and both are usable. Potential value in revenue functions is large. Realized value so far is concentrated where the work is structured, repetitive, and cheap to verify. The gap between potential and realized in GTM is the whole subject of this publication.
Readiness is a workforce problem, not a strategy problem
Leaders consistently report themselves ready on strategy and unready on the workforce. Deloitte found 5 percent say business processes are highly prepared for AI agents, while 74 percent expect nearly half of business processes to be redesigned around agents within four years. That is a four-year redesign expectation sitting on a 5 percent readiness base.
The practical consequence is sequencing. Workflow redesign takes longer than procurement and it is the variable with the largest measured effect, so it has to start before the tool arrives rather than after the rollout stalls.
What to do with it
The move, by seat.
- CRO
- Fund one redesigned workflow end to end rather than five tool pilots. The evidence gap between those two strategies is the largest in this library.
- RevOps
- Stand up impact tracking before rollout. Tracking is what converts a portfolio of pilots into a decision.
- HR and enablement
- Start the workforce work early. It is the slowest input and it is the declared readiness gap.
Questions this page answers
What the data says, in plain language.
- What does the research show about What separates the deployments that work?
- The largest gap between AI leaders and everyone else is not technology. It is having decided what to build. This is the most useful theme for an operator, and it has the best-quantified evidence in the library. Across Cisco, McKinsey, BCG, Deloitte, Scale VP and one field experiment, the same variables keep separating outcomes: redesigned workflows, finalized use cases, defined metrics, and a human kept in the loop.
- What does the figure "The largest gap between AI leaders and everyone else is not technology. It is having decided what to build" show?
- Pacesetters are 13% of the sample, and the classification is the publisher's. Source: Cisco AI Readiness Index 2025. n=8,000+ senior business leaders. Pacesetters are 13% of the sample Confidence: Medium. This figure comes from vendor research, so read it as a vendor claim.
- What does the figure "What is actually quantified to move the outcome" show?
- The units differ by row, so each row states its own unit rather than sharing one axis. Source: McKinsey, Scale VP, MIT NANDA, and Brynjolfsson, Li and Raymond. Per-row publisher, sample and field dates stated on each row. Units differ by row Confidence: Medium.
- What does the figure "Leaders say they are ready on strategy and unready on the workforce" show?
- Share saying prepared or highly prepared, by dimension. Source: Deloitte. n=501 US leaders Fielded April to June 2026. Confidence: Medium.
- What does the figure "5% of companies are capturing the outcome. 60% are not" show?
- Assessed against 41 foundational capabilities across nine industries. Source: BCG, The Widening AI Value Gap: Build for the Future 2025. n=1,250 senior executives and AI decision makers Confidence: Medium.
- What else sits alongside these figures?
- Cisco: 95% of Pacesetters track the impact of their AI investments, characterized as three times the rate of other organizations. BCG: agents were 17% of total AI value in 2025 and are expected to reach 29% by 2028. 70% of AI's potential value sits in core business functions including sales and marketing, against 13% in IT. BCG: more than 60% of future-built firms rigorously track AI value, against 17% of stagnating firms. Deloitte: 5% say business processes are highly prepared for AI agents, while 74% expect nearly half of business processes to be redesigned around agents within four years.
- Where does this data come from?
- Every figure is reproduced from a named publisher: Cisco AI Readiness Index 2025, McKinsey, The State of AI: how organizations are rewiring to capture value, Brynjolfsson, Li and Raymond, Generative AI at Work, BCG, The Widening AI Value Gap, Deloitte AI readiness survey. Sample, field date, and confidence are shown on each chart. Sources marked as vendor research are labelled on the page.
Cite this page
Permanent URL and suggested citation.
https://www.therevenueaireport.com/research/what-works
Kvarfordt, Jonathan. "What separates the deployments that work." The Revenue AI Report, Research Library. https://www.therevenueaireport.com/research/what-works
Figures on this page are reproduced from the publishers listed below. Cite the original publisher for the underlying data, and this page for the compilation and framing.
Sources
Every publisher used on this page.
If a metric, model term, or method on this page is unfamiliar, every one of them is defined in The AI and Revenue Dictionary. Sample size, field date, and confidence tags are explained there too.
Cisco AI Readiness Index 2025Vendor research
n=8,000+, vendor research, Pacesetters are 13% of the sample
https://newsroom.cisco.com/c/r/newsroom/en/us/a/y2025/m10/cisco-ai-research-the-most-ai-ready-companies-outpace-peers-in-the-race-to-value.htmlMcKinsey, The State of AI: how organizations are rewiring to capture value
n=1,491 across 101 nations, fielded July 16 to 31, 2024
https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-how-organizations-are-rewiring-to-capture-valueBrynjolfsson, Li and Raymond, Generative AI at Work
5,179 customer support agents, staggered rollout, QJE 2025, 140(2), 889 to 942
https://www.nber.org/papers/w31161BCG, The Widening AI Value Gap
n=1,250 senior executives across nine industries
https://www.bcg.com/press/30september2025-ai-leaders-outpace-laggards-revenue-growth-cost-savingsDeloitte AI readiness survey
n=501 US leaders, fielded April to June 2026
https://www.deloitte.com/us/en/about/press-room/deloitte-survey-examines-ai-readiness-agentic-ai-success.html
Could not confirm
What we looked for and did not find.
Claims found during research and not charted
Nothing on this theme was dropped for sourcing. Every claim we found that met the standards on the Research hub is charted above, and anything that failed them would be listed here by name.
Read the analysis
Issues built on this theme.
Research on this site is the evidence layer. These essays take the numbers above and apply them to real decisions, so you can see how the data reads in practice.
- The Signal Stack: What Common Room Buyers Learn After Go-Live
Signal platforms find intent. They do not hand you contacts, and they do not act. Here is the four-layer signal stack, the two costs buyers underestimate, and what the Zoom acquisition changes for renewals.
- The AI Revenue Team Org Chart: Which Roles Change, Which Appear, Which Go
Agents do not flatten the org chart. They move decision rights. Here is the before and after for seven revenue roles, the three jobs that appear, and the reporting line that decides whether any of it works.
- What Truth Tests Should We Run Before Buying Claudeforce, Agentforce, or an AI SDR?
AI on a dirty CRM is a prettier lie. Five truth tests a Director+ can run before Claudeforce, Agentforce, or AI SDR spend. Definitions and freshness first. Then the pilot.
How to cite this research
Written by Jonathan Kvarfordt, Founder and Principal Analyst, The Revenue AI Report. Published under CC BY 4.0.
APA
Kvarfordt, J. (2026). What separates the deployments that work. The Revenue AI Report. Retrieved from https://www.therevenueaireport.com/research/what-works
MLA
Kvarfordt, Jonathan. "What separates the deployments that work." The Revenue AI Report, 31 Aug. 2026, www.therevenueaireport.com/research/what-works.
BibTeX
@misc{kvarfordt2026whatworks,
author = {Kvarfordt, Jonathan},
title = {What separates the deployments that work},
year = {2026},
publisher = {The Revenue AI Report},
url = {https://www.therevenueaireport.com/research/what-works}
}Next theme
The money does not match the measurement$1.76 trillion went to AI in 2025. $826 million of it went to AI data, which is the most-cited barrier in every survey in this library.
Subscribe