The AI Transformation Gap Index (AITG): An Empirical Framework for Measuring AI Transformation Opportunity, Disruption Risk, and Value Creation at the Industry and Firm Level

Dean Barr

econ.GNFeb 27, 2026

The AI Transformation Gap Index (AITG): An Empirical Framework for Measuring AI Transformation Opportunity, Disruption Risk, and Value Creation at the Industry and Firm Level

Authors: Dean Barr

Organizations: Applied AI Researcher

Abstract

Despite the scale of capital being deployed toward AI initiatives, no empirical framework currently exists for benchmarking where a firm stands relative to competitors in AI readiness and deployment, or for translating that position into auditable financial outcomes. In practice, private equity deal teams, management consultants, and corporate strategists have relied on qualitative judgment and ad-hoc maturity labels; tools that are neither comparable across industries nor grounded in observable economic data. This paper introduces the AI Transformation Gap Index (AITG), a composite empirical framework that measures the distance between a firm's current AI deployment and a time varying, industry constrained capability frontier, then maps that distance to dollar denominated value creation, execution feasibility under uncertainty, and competitive disruption risk. Five linked modules address this gap: cross industry normalization (IASS), a dynamic capability ceiling that evolves with frontier capabilities (AFC), trajectory based firm scoring with integrated execution risk (IFS), a CES bottleneck value decomposition mapping gap scores to enterprise value (VCB), and a competitive hazard measure for inaction (ADRI). I calibrate the framework for 22 industry verticals and apply it to 14 public companies using public filings. A retrospective construct validity exercise correlating AITG scores with observed EBITDA margin expansion yields Spearman rho_s = 0.818 (n = 10), directionally consistent with predictions though insufficient for causal identification. A counterintuitive result emerges: the largest AI transformation gaps do not produce the highest value density, because implementation friction, CES bottlenecks, and timing lags erode the theoretical upside of wide gaps.

Figures & tables

Benchmark	Domain	2020	2024	2025–26	Source
MMLU-Pro	General knowledge	43.9 %	76.2 %	87.0 %	Stanford HAI 2025
SWE-Bench Verified	Software engineering	4.4 %	49.0 %	80.0 %	OpenAI 2025b
SWE-Bench Pro	Multilang. coding	n/a	n/a	55.6 %	OpenAI 2025b
AIME 2025	Competition math	n/a	71.0 %	100.0 %	OpenAI 2025b
FrontierMath (T1–3)	Research math	n/a	n/a	40.3 %	OpenAI 2025b
GPQA Diamond	Graduate science	n/a	63.0 %	93.2 %	OpenAI 2025b

Table 2: AI Capability Benchmark Progression (Selected Milestones)

Dimension	Abbrev.	Score 1	Score 3	Score 5	Score 7	Score 9
Data Infrastructure Maturity	DIM	Siloed legacy databases; no cloud data platform	Initial cloud migration; partial data catalog	Unified data lake/ warehouse; governed data catalog	Real time streaming pipelines; ML feature store	Enterprise knowledge graph; automated data quality ${>}$ 99%
Process Automation Coverage	PAC	${<}$ 5% processes automated; manual workflows	10–20% RPA deployment; limited scope	30–50% process automation; cross functional	50–70% intelligent automation; AI augmented	${>}$ 70% end to end automation; self optimizing
Workforce AI Augmentation Rate	WAR	${<}$ 5% employees using any AI tool	10–20% with basic AI tools (copilots)	30–50% augmented; tracked adoption metrics	50–70% AI augmented; integrated into workflows	${>}$ 70% augmented; org wide AI fluency measured
Decision Automation Rate	DAR	Fully human recurring decisions	${<}$ 10% decisions AI assisted	20–40% decisions AI recommended	40–60% automated with human override	${>}$ 60% fully automated with exception handling
AI Product/ Revenue Integration	APR	No AI revenue attribution	${<}$ 5% revenue from AI enhanced features	10–20% revenue AI attributable	20–40% revenue AI driven products	${>}$ 40% revenue from AI native products
Organizational AI Capability	OAC	No AI leadership; ad hoc experiments	Dedicated AI team; early governance	C suite AI sponsor; published AI strategy	Chief AI Officer; model risk framework	C suite AI ownership; published governance & risk controls

Table 3: AITG Dimension Scoring Rubric: Anchor Points and Evidentiary Requirements

$w$	Wave	Scope	$L_{w}$	$k_{w}^{\text{base}}$ (mo -1 )	$t_{0,w}^{\text{base}}$ (mo)
1	Foundation AI	Supervised ML, RPA, structured analytics, BI automation	4.0	0.38	18
2	Generative & Agentic AI	LLM workflows, copilots, Retrieval Augmented Generation (RAG) pipelines, agentic task execution	3.5	0.42	36
3	Autonomous AI	Multiagent orchestration, self improving systems, AI native products	2.5	0.32	60
Total asymptotic ceiling			10.0

Table 4: Cascading S Curve: Three Capability Waves

ADRI	Level	Competitive Implication	Action
$<2.5$	Low	Wide gap but protected by structural moat or slow diffusion	Monitor annually
$2.5$ – $4.9$	Moderate	Gradual margin compression; window for orderly transformation	Plan 3–5 year horizon
$5.0$ – $6.9$	High	Active competitive displacement; first mover advantages crystallizing	Immediate investment
$\geq 7.0$	Critical	Structural impairment risk; M&A or repositioning may be required	Acute / escalate

Table 6: ADRI Interpretation Grid

Factor	Type	Definition	Weight	$\delta$ Range
Org. Change Capacity	Trajectory	Leadership alignment, cultural readiness, prior transformation record	0.30	[0.60, 1.00]
Data Readiness	Trajectory	Quality, completeness, accessibility, governance; data pipeline health	0.30	[0.55, 1.00]
Vendor/Tech Risk	Residual	AI vendor ecosystem maturity; contract flexibility; switching costs	0.35*	[0.75, 1.00]
Competitive Response	Residual	Speed at which AI advantages erode due to competitor adoption (CADR)	0.35*	[0.70, 1.00]
Regulatory Exposure	Residual	Risk of regulatory action or compliance cost escalation during hold	0.30*	[0.55, 1.00]

Table 7: IFS Factor Specification (Revised: Trajectory vs. Residual)

Dimension	Score	Tier	Evidence
IASS Anchor: Commercial Banking (Large Cap), IASS ${}^{*}=9.38$ (AFC adjusted, $C_{2026}=1.90$ )
Data Infrastructure (DIM)	8.5	A	$18B tech budget; 65% workloads on cloud; near zero legacy debt
Process Automation (PAC)	8.2	A	450+ AI use cases in production; covenant extraction; pitch deck generation
Workforce Augment. (WAR)	8.5	A	250K LLM Suite users; $\sim$ 4 hrs/week productivity gain per employee
Decision Automation (DAR)	7.5	A/B	Agentic AI deployment begun; multistep autonomous task handling
AI Revenue Integration (APR)	7.8	A/B	AI enabled client concierge; IndexGPT personalization; ChatCFO

Table 8: JPMorgan Chase: AITG Dimensional Scores (Tier 1)

Dimension	Score	Tier	Evidence
IASS Anchor: Commercial Banking (Large Cap), IASS ${}^{*}=9.38$ (AFC adjusted, $C_{2026}=1.90$ )
Data Infrastructure (DIM)	4.5	C	Multibrand legacy systems; nCino migration in early stages; no data lake disclosed
Process Automation (PAC)	4.0	C	nCino deployment begun 2025; primarily manual lending workflow currently
Workforce Augment. (WAR)	3.5	C/D	No enterprise Generative AI (GenAI) platform disclosed; limited AI skill demand in job postings
Decision Automation (DAR)	3.8	C	nCino pricing tools initial; no autonomous credit decisioning at scale
AI Revenue Integration (APR)	2.8	D	No AI enabled product differentiation disclosed in public filings

Table 9: Zions Bancorporation: AITG Dimensional Scores (Tier 1)

Metric	JPMorgan Chase	Zions Bancorporation
Industry IASS ∗	9.38	9.38
AITG (raw)	8.22	3.80
Industry Relative (IR) Score	8.76	4.05
Effective Gap ( $G_{\mathrm{eff}}$ )	1.16	5.58
ADRI (Disruption Risk)	0.5 (Low)	2.6 (Moderate/Elevated)
IFS Residual Multiplier	0.88	0.71

Table 10: JPMorgan Chase vs. Zions Bancorporation: AITG Head to Head

Step 1: Raw Score (Cross Industry Comparison Without Normalization)
Metric	Zions Bancorp.	UPS
Industry	Comm. Banking	Logistics
AITG (raw, 0–10)	3.80	4.08
Naïve conclusion: UPS is “more transformed.”
Step 2: Apply IASS Ceiling (Cross Industry Normalization)
Industry IASS ∗ (AFC adjusted ceiling)	9.38	7.68

Table 11: Cross Industry Normalization: Zions Bancorporation vs. UPS

Panel of Figure 5 — Figure 5: The Value Density Paradox: Wider Gap $\neq$ Higher Return. Across 14 companies spanning 8 industries, the cross industry correlation between effective gap ( $G_{\mathrm{eff}}$ ) and Value Density midpoint is $r=0.22$ ( $p=0.45$ ), weakly positive and statistically non significant—far below the naïve expectation that wider gaps mechanically produce proportionally higher returns. The shaded green zone ( $G_{\mathrm{eff}}\in[1.3,\,3.0]$ ) is the sweet spot where S curve inflection proximity and manageable IFS produce maximum VD. The orange dashed curve shows the theoretical VD envelope derived from the AITG framework’s mathematical structure; points with the highest VD (WFC, CRM, PANW, GS) reflect high IASS ∗ ceilings or large revenue bases; points with the lowest VD (HCA, CVS) reflect near zero $G_{\mathrm{eff}}$ and regulatory ceiling IFS suppression. Vertical lines are Monte Carlo P10–P90 VD ranges. This chart illustrates a mathematical property of the framework, not an empirically established causal finding. See Section 11 for the endogeneity caveat and IV research agenda.

Parameter	Base	Range Tested	Output Metric	$\Delta$ Range
$\rho$ (CES)	5	$[3,8]$	$\Delta$ EV	$\pm 12\%$
$\theta_{i}$ (AFC)	industry	$\pm 50\%$	IASS ∗	$\pm 0.13$ pts
$\lambda$ (capture)	3.5	$[2.0,5.0]$	$\eta(g)$ at $g=0.5$	$[0.29,0.39]$
$\psi_{\mathrm{floor}}$	0.15	$[0.10,0.25]$	Min. capture	$\pm 0.05$
$k$ (wave steepness)	0.38	$[0.30,0.46]$	$\hat{t}_{f}$	$\pm 4.2$ mo
Exit multiple	$10\times$	$[8\times,16\times]$	AITG-VD	$\pm 42\%$

Table 13: Consolidated Sensitivity Analysis

Appendix figures & tables14 assets

Supplementary material from the paper’s appendix.

Appendix

Industry level constructs
Symbol	Definition	Range / Units
$\mathrm{IASS}_{i}$	Industry AI Susceptibility Score (base, before AFC)	$[0,10]$
$\mathrm{IASS}^{*}_{i}$	AFC adjusted ceiling: $\min\!\bigl(\mathrm{IASS}_{i}\cdot(1+\theta_{i}(C_{t}-C_{0})),\ \alpha_{\max}\,\mathrm{IASS}_{i}\bigr)$	$[0,\;\alpha_{\max}\!\cdot\!\mathrm{IASS}_{i}]$
$\theta_{i}$	Industry sensitivity to capability growth; anchored to $\Delta\mathrm{ATD}_{i}$	$[0.05,\,1.50]$
$\alpha_{\max}$	Cap on AFC multiplier (default 1.35)	$[1.0,\infty)$
$C_{t}$	AI Capability Index at time $t$ (externally observed)	$[1,\infty)$

Appendix

Table 15: Consolidated Notation Reference

$k$	Domain	Representative Benchmarks	$\omega_{k}$
1	Language Understanding	MMLU, HellaSwag, ARC	0.20
2	Mathematical Reasoning	MATH, GSM8K	0.15
3	Code Generation	HumanEval, MBPP, SWE-bench	0.15
4	Multimodal (Vision)	MMMU, ChartQA	0.10
5	Agentic / Tool Use	WebArena, WorkArena, ToolBench	0.15
6	Domain Specific	MedPerf, FinBench, LegalBench	0.15

Appendix

Table 17: AI Capability Index

C_{t}

: Benchmark Domains and Weights (

\omega_{k}

)

Scenario	Assumption	Annual $\Delta C_{t}$	AFC (24 mo)	$\Delta$ HC-IASS	Weight
Conservative	Capability slowdown; inference costs persist	8–12 %	1.04	+0.5 pts	0.20
Base Case	Observed 2022–24 trajectory continues	18–25 %	1.10	+1.1 pts	0.60
Aggressive	Reasoning models + agents deployed at scale	35–50 %	1.22	+2.2 pts	0.20

Appendix

Table 18: AFC Scenarios (24 Month Horizon)

IFS Specification	Backtest $\rho_{s}$	Mean $\|\Delta\text{EV}\|$ Error (%)	$\Delta$ vs. Full
Full IFS (OCC+DR+VTR+CRS+REG)	0.818	—	Baseline
Drop OCC	0.709	$+$ 18%	$-$ 0.109
Drop DR	0.745	$+$ 14%	$-$ 0.073
Drop VTR	0.800	$+$ 5%	$-$ 0.018
Drop CRS	0.791	$+$ 7%	$-$ 0.027
Drop REG	0.809	$+$ 3%	$-$ 0.009

Appendix

Table 20: IFS Ablation: Impact of Dropping Individual Factors

Company	Est. Cost	Basis	5 yr $\Delta$ EV	VD Range	VD Sensitivity
JPMorgan Chase	$2.3B	Revenue $\times$ 1.2% implementation rate	$20–98B	9.5–46.0 $\times$	–
Zions Bancorp.	$38M	Revenue $\times$ 1.2% implementation rate	$0.6–1.2B	15.0–30.1 $\times$	$\pm 0.4\times$ per $100M
UPS	$1.1B	Revenue $\times$ 1.2% implementation rate	$30–61B	27.3–55.4 $\times$	$\pm 0.3\times$ per $500M
HCA Healthcare	$780M	Revenue $\times$ 1.2% implementation rate	$5.8–20.6B	7.4–26.4 $\times$	$\pm 0.5\times$ per $500M
Salesforce	$420M	Revenue $\times$ 1.2% implementation rate	$10.6–33.7B	25.2–80.4 $\times$	$\pm 2.0\times$ per $500M
Ferguson	$360M	Revenue $\times$ 1.2% implementation rate	$7.8–16.0B	21.8–45.0 $\times$	$\pm 0.4\times$ per $100M

Appendix

Table 22: Implementation Cost Assumptions Underlying VD Estimates

Dimension	Score	Evidence Signal
Data Infrastructure (DIM)	5.0	Cloud investment ongoing; RFID rollout complete
Process Automation (PAC)	4.5	62 % automated facilities; Teamster constraints persist
Workforce Augment. (WAR)	3.5	AI tools in operations; broad workforce augmentation nascent
Decision Auto. (DAR)	4.0	AI routing, pricing, network planning in production
AI Product/Revenue (APR)	3.0	Smart logistics features; no standalone AI revenue line
Org. AI Capability (OAC)	4.5	Dedicated AI team; no public Chief AI Officer

Appendix

Table S7 : UPS: AITG Dimensional Scores (Tier 1)

Dimension	Score	Evidence Signal
Data Infrastructure (DIM)	5.5	Epic EHR; Google Cloud; AWS Bedrock; MLOps emerging
Process Automation (PAC)	4.8	Revenue cycle AI deployed; clinical workflows scaling
Workforce Augment. (WAR)	4.2	78 % physician DAX Copilot satisfaction; broad rollout
Decision Auto. (DAR)	4.0	Staffing AI deployed; sepsis prediction; pricing manual
AI Product/Revenue (APR)	2.5	No external AI product; AI improves service quality
Org. AI Capability (OAC)	5.0	Dedicated AI teams; CFO publicly leading AI agenda

Appendix

Table S8 : HCA Healthcare: AITG Dimensional Scores (Tier 1)

Dimension	Score	Evidence Signal
Data Infrastructure (DIM)	8.5	50T+ records in Data Cloud; Hyperforce global deployment
Process Automation (PAC)	8.2	Agentforce; automated service flows at scale
Workforce Augment. (WAR)	7.5	Einstein tools across all roles; AI assisted selling
Decision Auto. (DAR)	7.8	Einstein GPT; automated service resolution; ML forecasting
AI Product/Revenue (APR)	8.2	AI is core product; $900M Data & AI ARR
Org. AI Capability (OAC)	8.2	C suite AI mandate; Responsible AI framework published

Appendix

Table S9 : Salesforce: AITG Dimensional Scores (Tier 1)

Test	Null Hypothesis	Proposed Design	Status
AITG predicts subsequent margin improvement	$H_{0}$ : No AITG margin correlation	Retrospective: compute 2020 AITG from 2020 10-K data; regress on 2020–2024 EBITDA margin change; 50+ company panel	Not yet run
ADRI predicts competitive share loss	$H_{0}$ : ADRI uncorrelated with share loss	High ADRI firms ( $>3.0$ ) in Commerce, Logistics lose revenue share within 24 mo; IV: CADR shock via LLM release event	Proposed
IFS predicts transformation failure	$H_{0}$ : IFS uncorrelated with AI project cancellation	Lightcast AI job posting drop as proxy for abandoned programs; low IFS firms show higher cancellation rate	Proposed
Cross industry VD ranking is ordinal consistent	$H_{0}$ : AITG-VD does not predict actual deal multiples paid	PE transaction database: AI thesis deals vs. non thesis deals; AITG scores predicted from pre deal public data	Proposed

Appendix

Table S10 : Predictive Validity Test Battery: Status and Proposed Design

Component	Primary Public Sources	Diligence Gap	Public %	Tier
Company Dimension Scores
DIM (Data Infra.)	Tech budget disclosures (10-K MD&A); cloud provider announcements; IT vendor press releases; EDGAR risk factors mentioning legacy systems	Actual data quality, schema maturity, MLOps pipeline state	50–65 %	B/C
PAC (Process Auto.)	Earnings call AI deployment counts; investor day demos; vendor partnership announcements (e.g., nCino, Salesforce, SAP)	Process level automation depth; proportion of workflows vs. surface coverage	55–70 %	B/C
WAR (Workforce Aug.)	Headcount disclosures; AI platform adoption figures (when disclosed); Lightcast AI skill demand by employer; LinkedIn job postings with AI tool requirements	Actual employee usage rate vs. license provisioning rate	45–60 %	B/C
DAR (Decision Auto.)	SEC filings describing automated underwriting, pricing, or routing systems; vendor AI product deployment announcements; patent filings (Google Patents)	Decision quality metrics; actual automation rate vs. human in loop proportion	40–55 %	C
APR (AI Revenue)	Product documentation; AI feature release notes; investor day product demos; sell side analyst coverage of AI driven pricing; ARR growth in AI segments	Revenue attribution by AI enabled feature; AI product margin vs. legacy	60–75 %	B/C

Appendix

Table S11 : AITG Scoring Component: Public Data Coverage Map

Original cohort
Company	DIM	PAC	WAR	DAR	APR	OAC	OCC	DR	VTR	CRS
JPMorgan Chase	A	A	A	A/B	A/B	A	A	A	A	A
Zions Bancorp.	C	C	C/D	C	D	C	C	C/D	B	B
UPS	B	B	B	B/C	B	B	B	B/C	A	A
HCA Healthcare	B	B	B	B/C	B/C	B	B	C	A	A
Salesforce	A	A	A	A	A	A	A	A	A	A

Appendix

Table S13 : Fourteen Company Evidence Tier Audit (Tier A–D per Appendix S6 )

Tier	Source Type	Examples	$\Delta$ UQ
A	Public structured data	EDGAR XBRL financials, BLS OEWS, O*NET	$0.00$
B	Public unstructured / alt. data	Earnings transcripts, Lightcast, job postings	$+0.08$
C	Primary diligence (verified)	Management survey with evidence attachments	$+0.18$
D	Primary diligence (self reported)	Unverified management assertion	$+0.30$

Appendix

Table S14 : Data Evidence Tier Definitions

Exit Multiple	Raw EV ( $\sum V_{p}$ )	Risk Adj. EV	Cost	VD
$8\times$ EBITDA	$38B	$28.5B	$8B	2.6 $\times$
$12\times$ EBITDA	$57B	$42.8B	$8B	4.3 $\times$
$16\times$ EBITDA	$76B	$57.0B	$8B	6.1 $\times$

Appendix

Table S15 : AITG-VD Sensitivity to Exit Multiple (Healthcare, Representative Case)

Component	Excel?	Formula (single cell)
IASS Geometric Mean	✓	=EXP(SUMPRODUCT(weights, LN(scores)))
RFF $\psi$ multiplier	✓	=MIN(1,(RFF/5) ˆ 1.5)
CES Bottleneck ( $\rho=5$ )	✓	=SUMPRODUCT(alpha,scores ˆ -5) ˆ (-1/5)
$\Phi_{f}$ logistic	✓	=1/(1+EXP(-2*LN(rev/S_star)))
Value ramp $R_{f}(t)$	✓	=1/(1+EXP(-0.18*(t-t50)))
$\Delta R_{f}$ correction	✓	=ramp(t_hat+60) - ramp(t_hat)

Appendix

Table S17 : AITG Components: Excel Implementation Formulas

Competitive Analysis

Competitive analysis compares online decisions with an offline optimum to measure performance under worst-case inputs and imperfect information.

See all 17 papers

AI Exposure and AI Resilience: A Two-Dimensional Assessment Framework for Software and Software-Based Business Model
Sep 10, 2026Paul Darius Mandl, Peter Mandl, Martin Häusl
Findustrial GmbH · Munich University of Applied Sciences
The Eticas AI Risk Taxonomy: Open Infrastructure for Operationalizing AI Audits
Jul 2, 2026Gemma Galdon Clavell, Pablo Accuosto, Usman GoharAI Assurance AI Risk Management
The AI Evaluability Gap: The Missing Layer for Managing Risk and Sustaining Value
Jun 19, 2026Vishal Srivastava, Tanmay SahAI Assurance AI Risk Management
Whiting School of Engineering Johns Hopkins University Baltimore, MD 21218 · Harrisburg University of Science and Technology Harrisburg, PA

Sep 10, 2026J/K move · Enter open · S save

AI Exposure and AI Resilience: A Two-Dimensional Assessment Framework for Software and Software-Based Business Model

Paul Darius Mandl, Peter Mandl, Martin Häusl

Findustrial GmbH · Munich University of Applied Sciences

Open paper AskPDF arXiv

Artificial intelligence is changing both software production and the economics of software-based business models. Classical technology due diligence mainly examines technical properties such as architecture, scalability, and technical debt. These criteria do not fully capture how AI can affect a company's value proposition, competitive position, margins, or access to customers. This paper develops Artificial Intelligence Exposure and Resilience (AI-ER) as a two-dimensional assessment framework. AI exposure describes the pressure for change that AI creates for a business model. AI resilience describes the company's ability to absorb that pressure, adapt to changed conditions, and use AI in an economically viable way. Metrics for both dimensions are derived from current AI capabilities, their deployment conditions, and relevant research on business models and organizational adaptability. The model keeps exposure and resilience separate and adds an explicit assessment of evidence quality and confidence. It can be applied first with public information and later refined with internal evidence. The result is a traceable company profile that supports comparison without concealing uncertainty in the underlying evidence. The paper also specifies an initial score logic and a procedure for empirical validation.

Abstract

Competitive Analysis

Explore similar work

AI Exposure and AI Resilience: A Two-Dimensional Assessment Framework for Software and Software-Based Business Model

The Eticas AI Risk Taxonomy: Open Infrastructure for Operationalizing AI Audits

The AI Evaluability Gap: The Missing Layer for Managing Risk and Sustaining Value