📊
Market Size & Adoption
Spend is exploding; value capture is rare. The reliable signal is the gap between the two.
The CEO thesis in one line: Adoption is near-universal (88% of orgs use AI) and spend is soaring (Gartner: $2.59T total AI spend in 2026, +47%; agentic software $206.5B, +139%), yet independent research says most efforts never reach measurable ROI — MIT finds 95% of GenAI pilots deliver no P&L return, McKinsey finds only 39% report any enterprise EBIT impact and ~6% are high performers. And in July 2026 the correction turned visible: 55% of leaders who did AI layoffs now call it a mistake (Orgvue). The opportunity EmpowerLabs sells into is precisely that adoption-to-impact gap — and it is widening.
Total AI Spend 2026 (Gartner)
$2.59T
+47% YoY · raised from $2.52T
Agentic AI software alone: $206.5B (+139% YoY, from $86.4B). Microsoft's AI run-rate crossed $37B (+123%).
Gartner Worldwide AI Spending (May 2026 revision); Microsoft earnings
Enterprise AI Adoption
88%
use AI in ≥1 function
~71% regularly use GenAI; 72% report agentic AI in production. Yet <20% of pilots reach production.
McKinsey State of AI 2025; Agentic AI Institute 2026
Leaders Regretting AI Layoffs
55%
of the 39% who cut for AI
32% rehired the same/similar role. Ford, IBM, Salesforce, CBA all reversed AI-driven cuts — "the Layoff Boomerang."
Orgvue survey; CNBC Jul 1, 2026
Pilots With No ROI
95%
no measurable P&L
S&P Global: 42% abandoned most AI projects in 2025. IBM: only 25% of initiatives hit expected ROI. McKinsey: ~6% high performers.
MIT NANDA "The GenAI Divide"; S&P Global; McKinsey State of AI 2025
The Adoption → Impact Gap
McKinsey State of AI 2025; Agentic AI Institute 2026 (production figure); Gartner/RAND (abandonment). Adoption is wide, governance and scaled impact remain rare.
Independent Failure-Rate Evidence
| Source | Finding |
|---|---|
| MIT NANDA (2025) | 95% of GenAI pilots deliver no measurable P&L return (vendor-built tools succeed ~2× vs DIY) |
| McKinsey (Nov 2025) | 80%+ report no meaningful enterprise-wide EBIT impact; only ~6% reach scaled outcomes |
| Gartner (2025) | ≥50% of GenAI projects abandoned after PoC — poor data quality, unclear value, cost overruns |
| Gartner (2026) | >40% of agentic AI projects at risk of cancellation by 2027; 60% lacking AI-ready data abandoned |
| RAND / GSPANN (2026) | 74% of AI agent deployments rolled back; governance gap is #1 differentiator for surviving deployments |
| Orgvue / CNBC (Jul 2026) | 55% of leaders who did AI layoffs now call it a mistake; 32% rehired the same role — the "Layoff Boomerang" |
Vendor ROI Claims vs. Independent Reality (read with skepticism)
Vendor-sourced surveys claim median agentic-AI ROI of 171% globally (US up to 192%, top-quartile >540% in 18 months) and ~66% of agent users reporting productivity gains. These are self-reported and conflict sharply with the independent MIT/McKinsey/Gartner findings above. Present them as "vendors claim X; independent research finds the opposite." Sources: onereach.ai / Accelirate roundups (vendor) vs. MIT NANDA, McKinsey, Gartner (independent).
⚠️ Confidence note: "Market size" TAM figures vary wildly across research vendors ($300B–$758B for 2025) and are sales-incentivized — treat as directional. Spend (IDC), adoption (McKinsey), and failure rates (MIT/Gartner) are independent and decision-grade. The "95%" stat is one influential MIT study (n≈153 surveyed) — cite as such, not gospel.
🌎
Global / USA / LATAM
Where the money and adoption concentrate — and where the runway is.
Global AI Market 2025
$300–550B
defensible range
~25–35% CAGR. Crosses $1T ~2029. ($750B+ vendor figures are outliers.)
Composite of Fortune BI, Grand View, Precedence (vendor)
USA — Share of Global Spend
~58%
2025, declining to ~49% by 2035
North America ~35.5% of global AI market revenue.
Oxford Economics-type modeling (secondary); Grand View
LATAM AI Market 2025
~$5.8B
22% CAGR → ~$30B by 2033
Brazil #1 (top 50 globally for AI readiness, score 65.89); Microsoft committed $2.7B to Brazil AI/cloud through 2027.
IMARC / Intelligent CIO LATAM; Microsoft press
Regional Read for a CEO
| Region | Signal |
|---|---|
| Global | Spend ~$307B (2025) → $1.3T (2029). Adoption 88%, scaled impact ~6%. The gap is the market. |
| USA | ~58% of global spend concentration; leads realized ROI but also the loudest "AI disillusionment" cycle. Most mature buyer. |
| LATAM | Small base (~$5.8B) but ~22% CAGR. Brazil #1 (~25% share, gov't AI plan); Mexico top conversational-AI growth. Manufacturing is the lead vertical (~17%). |
LATAM — Barriers & Opening
LATAM sits at 47% enterprise AI deployment in 2026 — ahead of many mid-market regions. Brazil and Mexico drive ~70% of regional VC. Trust & governance remain the headline friction: 65% of LATAM consumers use AI tools, but "confidence lags participation" (LatAm Intersect PR). Microsoft's $2.7B Brazil commitment through 2027 signals infrastructure is arriving. Opening for EmpowerLabs: a human-in-the-loop, governed agentic model (WORX + SherpaX) directly answers the trust and governance objections that full-automation vendors trigger in this market.
IMARC, Grand View, Intelligent CIO LATAM, LatAm Intersect PR, VuraOS AI Adoption Map 2026. Note: some vendors cite $29.5B–$40.5B for 2025 LATAM AI — these figures likely reflect broad digital/IT spend; the $5.8B (IMARC) is more defensible as an AI-specific estimate.
🏆
Success Cases
The documented exceptions — what winning AI transformation actually looks like.
JPMorgan Chase — AI at scale, in production
VERIFIED · 2025–26~200,000 employees use its proprietary LLM Suite daily, 450+ agentic AI deployments, $18B annual tech budget. Agentic agents now generate investment banking presentations in 30 seconds (vs. hours for junior analysts). COiN platform: 12,000 credit agreements in seconds vs. ~360,000 manual hours. Reported 171% average ROI across its agent portfolio; industry pattern shows scoped deployments returning ~2.3× within 13 months vs. broad experiments.
Lesson: scale comes from an owned, integrated platform deployed broadly with scoped use cases and defined KPIs — not scattered pilots. Agentic AI augments knowledge workers, not replaces them.
AI Monk / alicelabs.ai (2026 case compilation); COiN company stats (directional).
General Mills — AI supply-chain optimization
VERIFIED · 2026Deployed AI-driven supply chain systems delivering $20M+ in savings since FY2024. Measurable P&L impact, production-grade, not a pilot.
Lesson: AI wins in supply chain come from deploying across existing workflows at scale — not isolated experiments.
alicelabs.ai Enterprise AI Case Studies 2026.
Elevance Health — HealthOS claims management
REPORTED · 2026Its HealthOS AI platform cut claims denials by 68%. In healthcare, AI agents for documentation reduced time per provider by ~42%.
Lesson: sector-specific agentic AI delivering 500% ROI on average vs. horizontal tools — validate the vertical specialization argument.
AI Monk agentic AI case studies 2025–26.
Starbucks — "Back to Starbucks" turnaround working
VERIFIED · Q2 FY2026After six straight quarters of declines, Q2 FY2026 snapped the streak: global comps +6.2% (US +7.1%), revenue up 9% to $9.5B, EPS beat at $0.50 vs. $0.42 expected. All 10 largest international markets turned positive simultaneously. Guidance raised: SSS ≥5% FY26, EPS $2.25–$2.45. Brian Niccol's "Back to Starbucks" labor + operational overhaul (including AI/Azure tools) is gaining traction.
Lesson: turnarounds powered by human-centered re-investment + targeted AI tools outperform full automation bets. (Note: also listed under Failure Cases for the NomadGo inventory AI reversal.)
Starbucks Q2 FY2026 Earnings Release (Apr 28, 2026); Investing.com; CNBC.
⚠️
Failure & Cautionary Cases
The hype-vs-reality gap, where the strongest sales narratives live.
The "Layoff Boomerang" — Ford, IBM & others rehire after AI cuts
NEW · VERIFIED · Jul 1, 2026The week's biggest signal: firms that cut staff for AI are reversing. Workforce-planning firm Orgvue found 39% of leaders laid off employees because of AI, but 55% later concluded it was a mistake, and 32% rehired the same or a similar role. Named cases: Ford rehired/promoted 350 experienced engineers and then topped JD Power's 2026 Initial Quality Study for the first time since 2010. IBM's AI handled ~94% of routine HR requests but couldn't manage the 6% needing judgment/ethics — IBM is now tripling US entry-level hiring. CBA (call center) reversed similarly earlier in the year.
Lesson: "replace headcount with agents" is the failure pattern; "augment humans with governed agents" is the win pattern. The rehiring wave is direct, public proof of the human-in-the-loop thesis EmpowerLabs sells.
CNBC (Jul 1, 2026); Orgvue survey; IBTimes/Medium roundups; JD Power 2026 IQS.
Salesforce — cuts its own Agentforce team; AI chief departs
NEW · VERIFIED · 2026The loudest "agentic workforce" incumbent stumbled: Salesforce laid off <1,000 employees across four core teams, including the Agentforce division, and Adam Evans, EVP/GM of AI, left the company — the 4th senior executive to depart within three months. A cautionary signal that even the best-capitalized agentic-platform push is hitting execution and org friction.
Lesson: selling agents is not the same as operating them at scale. Method + governance + change management (what EmpowerLabs productizes) is what turns an agent platform into results — and its absence shows up as reorg churn.
TechCrunch running-layoffs list (Jun 22, 2026); CNBC; multiple.
Microsoft Copilot — 20M seats, but 74% can't show ROI
VERIFIED · Q4 FY26M365 Copilot crossed 20M paid enterprise seats (up from 15M last quarter), with 90%+ of Fortune 500 using it. But real usage is only 20–30% weekly active among licensed seats, and only 3.3% of all M365 commercial users have bought the paid add-on. Critically, 74% of companies cannot yet show measurable AI ROI. Even Forrester's optimistic model shows 116% ROI — contingent on structured adoption programs that most firms skip.
Lesson: seats ≠ adoption ≠ ROI. Per-seat AI add-ons face a brutal measurement bar. The 70%+ weekly-active ceiling is where real returns appear — requires the governance and change management work that EmpowerLabs productizes.
Windows News/windowsforum.com (Jun 2026); valueaddvc.com; Forrester TEI study; getpanto.ai.
Commonwealth Bank of Australia — call center AI reversal
VERIFIED · 2025–26CBA replaced its 45-person call center with AI voicebots targeting 2,000 calls/week — but performance fell short. After one month, the bank apologized and rehired the displaced workers. The contrast: CBA's AI-powered fraud detection (processing 20M transactions/day, ~30% fraud loss reduction) is a major success. Same company, wildly different outcomes based on use case fit.
Lesson: AI works where pattern-matching dominates (fraud); fails in nuanced human interactions (call centers). Use-case selection before deployment is the highest-leverage decision.
IntuitionLabs AI enterprise failure roundup; CIO.inc (CBA fraud AI success cases).
Klarna — the canonical AI reversal
VERIFIED · May 2025After claiming its AI did the work of 700 agents, CEO Siemiatkowski publicly reversed course in May 2025 and resumed hiring humans, admitting AI-only support "led to lower quality." Moved to a hybrid model: AI for routine, humans for escalations.
Lesson: in accuracy-sensitive, customer-facing work, full automation underperforms. The winning pattern is hybrid — AI triage + human escalation. (Validates WORX's human-in-the-loop stance.)
FinTech Weekly (May 12, 2025), Reworked.
Starbucks NomadGo — inventory AI abandoned mid-turnaround
VERIFIED · 2025 (see also: Success Cases)During its "Back to Starbucks" pivot, Starbucks retired its NomadGo inventory AI across ~11,000 stores after workers had to recount every scan — the system's error rate made manual verification mandatory, defeating the purpose. Separately, the overall turnaround is now working (Q2 FY26 global comps +6.2%, US +7.1%) using human + Azure OpenAI tools. This is the cautionary thread: AI tool choice matters even within a successful transformation.
Lesson: individual AI tools can fail inside overall turnaround success. Vetting each use case for accuracy requirements before deploying is non-negotiable.
CNBC; Restaurant Dive; TechTimes (NomadGo, single source); Starbucks Q2 FY26 earnings.
McDonald's — AI drive-thru cancelled
REPORTEDEnded its IBM-partnered AI voice-ordering pilot at 100+ locations after persistent order errors (misheard orders, unwanted items added).
Lesson: same as Klarna and CBA — full AI automation in accuracy-sensitive, real-time customer contexts repeatedly fails. Scope agents narrowly and keep a human backstop.
AOL/Yahoo Tech; widely covered.
📉
SaaS in Decline — the AI repricing
"AI is eating SaaS." Markets are repricing seat-based software as switching costs fall.
The thesis (now being tested by a bounce): SaaS repriced hard through mid-2026 — BVP Emerging Cloud Index ~6.3× EV/revenue (down from 6.6× in early June), private SaaS ~3.7×, vs. the 2021 ~20× peak. But the week's twist is a relief rally: HubSpot bounced off its Jun 25 low of $170.33 to ~$182, and Guggenheim upgraded both Salesforce and ServiceNow to Buy, arguing "AI-disruption fear gutted the sector and pushed valuations too low." The picture stays bifurcated — AI-native platforms hold premiums, legacy seat-based SaaS is cheap — but the trade is no longer one-directional.
Salesforce (CRM)
−40%
YTD · Jul 2 · ~3.7× rev
Recovered modestly off its 52-wk low; Guggenheim upgraded to Buy. FY27 Q1 rev +14%, 19% net margin. Seat-pricing thesis still the overhang.
Motley Fool Jun 30; Yahoo Finance Jul 2, 2026
Monday.com (MNDY)
−23%
past month · strong fundamentals
Revenue +24% in Q1'26, 14% operating margin; shares soared in early May then gave back ~23% on seat-model fears. Consensus PT implies ~47% upside.
Yahoo Finance; StockStory Jun 2026
HubSpot (HUBS)
~$182
−54% YTD · bounced off Jun 25 low
Recovered from its Jun 25 low of $170.33; still ~64% below its May 2025 high of $672. Revenue +23% in Q1. Concern: AI agents vs. per-seat CRM.
Yahoo Finance Jul 2; Motley Fool Jun 30, 2026
ServiceNow (NOW)
−50%
12-month · ~$101 (Jun 16)
Down ~49.6% over 12 months; Guggenheim upgraded to Buy, $125 target, calling the AI-disruption discount overdone.
Yahoo Finance; Guggenheim (John DiFucci) Jun 2026
Year-to-Date Stock Performance (Jul 2, 2026)
Yahoo Finance / Motley Fool / StockStory live quotes Jun 30 – Jul 2, 2026. HUBS ~$182 (bounced off Jun 25 low of $170.33); CRM −40% YTD; both upgraded to Buy by Guggenheim. NOW figure is 12-month; MNDY is trailing-month. Re-pulled this Saturday.
Index & Contrast
| Signal | Reading |
|---|---|
| BVP Emerging Cloud Index (avg) | ~6.3× EV/revenue (late Jun 2026), down from ~6.6× in early June; private SaaS closing ~3.7×. Cloud leaders pull the average up; the "typical" SaaS company is near record-low multiples. |
| Adobe (ADBE) — AI-first ARR | Firefly / AI-first ending ARR >$250M and tripling YoY (~75% QoQ) — but stock −30–33% YTD, CEO Narayen exit announced, $25B buyback, 850M MAU. The AI pivot is real; the market wants proof it offsets declining traditional licenses. |
| Winners by contrast | Agentic platforms — Sierra ~$15.8B ($200M ARR), Cognition ~$26B ($492M ARR, Devin writes 89% of its code) — raise at >60–100× ARR while traditional SaaS trades at 3–6×. AI-native execution is the only re-rating lever. |
| What it means for EmpowerLabs | Even with this week's relief rally, buyers of EmpowerLabs are under SaaS repricing pressure while agent-platform costs rise — budget is being freed for governed transformation work. EmpowerLabs sits on the right side of this trade. |
⚠️ Confidence note: Stock figures re-pulled Jun 30 – Jul 2, 2026 (Yahoo Finance, Motley Fool, StockStory). YTD figures use Jan 1 2026 baselines; different sources cite HUBS at −37% to −54% within the same week depending on intraday quote — the −54% (Motley Fool Jun 30) is the most conservative dated pair with CRM −40%. BVP ~6.3× is the weighted average, not median — headline skewed by Snowflake, ServiceNow, Datadog. Asana (ASAN): FY26 revenue $790.8M (+9.3%), FY27 guide raised. This Saturday's read: repricing intact but a Guggenheim-led relief rally is underway.