Where investment
conviction meets
irreversible change

Ask a management team if AI is transforming their commercial engine and the answer is always yes. Ask what specifically changed in the workflow, and most cannot say. That gap, not the technology, is what should worry anyone writing a check on the strength of a confident pitch.

McKinsey's 2026 B2B Pulse Survey, covering nearly 4,000 buyers and sellers across 13 countries, confirms the divide is real and widening. High-growth companies are three times more likely than laggards to have raised AI investment by double digits this year, 71 percent versus 25 percent.¹ But investment volume is the wrong thing to measure. IFC and the World Bank Group's own survey of more than 2,200 companies across 85 countries found that only about 40 percent report any AI adoption at all, and fewer than a quarter have integrated it across both internal operations and customer-facing work.² That population is particularly relevant to private-market investors: IFC surveyed companies it backs directly or through private equity and venture capital funds across emerging markets. It suggests adoption alone tells an investor almost nothing about what is actually happening inside the business.

What separates growth champions has little to do with the model. High performers are nearly three times more likely to have redesigned a commercial workflow around AI rather than adding it to the one that already existed.¹ Visa's own research into where this is heading is instructive: 53 percent of US business decision-makers now say they would let an AI agent negotiate price and terms directly with a counterpart's AI agent, and 88 percent would hand that agent live pricing and inventory data to make it possible.³ Handing an agent that kind of authority only makes sense once the underlying workflow, and the data behind it, has actually been rebuilt to support it. Most companies are nowhere close.

The real constraint is not technical, and this is the part boards resist hearing. Agents cannot generate trusted recommendations when customer, product, and pricing data sit in systems that cannot agree on what counts as a qualified lead. The fix McKinsey names is one accountable owner per data domain, someone who can say with certainty what a closed deal includes.¹ That is a structure due diligence can verify in an afternoon. Either a company can produce that owner, or the AI story is decoration.

Bain & Company, working the same problem from the private equity side rather than the technology side, reaches an identical conclusion in its own push into portfolio company AI deployment: strategic insight and technical capability only compound value when change management carries equal weight in the plan.⁴ Skip that step and the tool sits on top of the same broken workflow, generating the same false confidence, just faster.

For anyone structuring capital around milestones, this changes what a milestone should measure. A revenue target attributed loosely to AI adoption is not a milestone. It is a story told to a board. A named data owner and a documented workflow redesign are not.

The companies worth backing will not be the ones with the most AI in their stack. They will be the ones that can name, without a deck, who owns the pricing data and what changed because of it.

There is a number that tells the real story of enterprise AI in 2026, and it is not the headline adoption figure. It is this: 88% of organizations globally now use AI in at least one business function. Yet only about one-third have begun scaling AI across the enterprise, and just 25% have moved 40% or more of their pilots into production.⁵

The gap between adoption and deployment is not a technology problem. Deloitte's 2026 State of AI in the Enterprise report, drawn from 3,235 leaders across 24 countries, finds that governance readiness sits at just 30%, data management readiness at 40%, and talent readiness at only 20%.¹ The tools are in the building. The infrastructure to use them at scale is not.

This matters most in markets where the adoption curve is steepest. Forecasts place Southeast Asia AI growth at roughly 30 to 37% annually through the end of the decade, with Indonesia, Malaysia, and Vietnam each expanding at between 29% and 31% annually from a base that is still early.² BCG's July 2025 survey of over 4,500 APAC employees finds that 92% of Indonesian knowledge workers use generative AI, among the highest rates in Asia-Pacific.³ Vietnam's AI law, passed in December 2025, took effect in March 2026, making it one of the first countries in Southeast Asia to establish a comprehensive regulatory framework for AI.⁴ The demand signal is real and accelerating. Yet across the region, the gap between individual AI use and enterprise-level deployment mirrors the global pattern, compounded by data localisation requirements, regulatory fragmentation across borders, and a shortage of implementation infrastructure that converts pilots into production.

One dimension of this gap is particularly underappreciated: governance. McKinsey and the National Association of Corporate Directors find that despite AI's rapid spread across business functions, only 39% of Fortune 100 companies have disclosed any form of board-level oversight of AI. Sixty-six percent of directors globally report their boards have limited to no knowledge or experience with AI. Nearly one in three say AI does not appear on their board agenda at all.⁶ Organizations are scaling AI faster than their governance models can follow. McKinsey's 2026 AI Trust Maturity Survey confirms this is a globally consistent pattern: governance and agentic AI controls lag behind data and technology readiness across every region surveyed, including Asia-Pacific.⁷ The consequence is not merely reputational. A 2025 MIT study found that organizations with AI-fluent boards outperform peers by 10.9 percentage points in return on equity. Those without are 3.8% below their industry average.⁶

For technology platforms entering Southeast Asia and the Middle East, this governance gap creates both a risk and a structural advantage. Markets in this region are actively building AI regulatory frameworks, and doing so quickly. Vietnam's law, Malaysia's evolving data governance standards, and the Gulf states' sovereign AI frameworks are not obstacles to entry. They are institutional signals that the markets where governance infrastructure is being built earliest are also the ones where trusted, well-structured partners will hold the most durable positions.

This is precisely where structured, locally embedded partnerships create asymmetric value. An AI-capable technology platform entering these markets alongside a partner with dedicated operational presence, local governance fluency, and milestone-based execution accountability is not simply faster to deploy. It is structurally better positioned to cross the threshold from experimentation to enterprise-wide impact.

Seventy-four percent of organizations globally say they want AI to grow revenue. Only 20% are already doing so.¹ The gap between ambition and outcome is not closed by better models, or by better intentions at the board level. It is closed by better execution infrastructure, built into the market from the start.

The tools exist. The governance frameworks are being written. The question is whether the operational structure to act on both is in place.

Southeast Asia's AI market is growing at a 37% compound annual growth rate, faster than any other region in Asia.¹ Indonesia, Malaysia, and Vietnam are expanding at between 29% and 31% annually from a base that is still early.² Over 4,600 megawatts of new data center capacity is planned across the region, with total infrastructure set to grow by 180%.³ The demand signal is unambiguous. What is equally clear is that international enterprise AI platforms have not followed it.

Over 50% of APAC digital businesses remain at the early stages of AI workflow maturity, compared with 87% of North American enterprises at an optimized stage.² Enterprise AI deployment in most Southeast Asian markets is running well behind individual adoption. Indonesia's worker AI usage rate is among the highest in the world. Vietnam passed a dedicated AI law in March 2026. Yet the structured, enterprise-grade AI platforms serving these markets at depth remain scarce.³ That gap is the opportunity.

The first-mover case is not abstract. Enterprises that embed AI into core workflows, compliance processes, supply chain operations, or financial services infrastructure become structurally difficult to displace. Seventy-one percent of Southeast Asian businesses report AI investment ROI within twelve months, the fastest realization rate globally.³ Early entrants capture that value and set the standard against which every subsequent platform is judged. Customers calibrate expectations to the first system that works well. Competitors entering later face those expectations while operating with the learning curve of beginners.

The structural case is equally strong. Markets growing at 30%+ annually from an underpenetrated base reward platforms that arrive with genuine local commitment: product adaptation, local partnerships, and capital structures designed for the market rather than retrofitted from a headquarters playbook.

The window is open. Platforms that move now, with the right architecture, will not simply be first. They will be hardest to replace.

When technology platforms expand into Asia, the default playbook concentrates on a familiar cluster: markets with well-established legal infrastructure, mature cloud ecosystems, and minimal product adaptation requirements. These are genuinely strong markets. Singapore alone hosts the regional headquarters of Salesforce, Microsoft, ServiceNow, and Amazon Web Services, and its SaaS market is growing at 13.5% annually.¹ The problem is that this quality and accessibility attracts every other platform simultaneously. Australia, New Zealand, and Singapore reached SaaS maturity earlier than any other markets in the region and today draw the sustained focus of the world's largest vendors.² Winning a position and holding it against well-capitalised incumbents requires sustained investment and tolerance for margin compression.

Meanwhile, Southeast Asia's broader enterprise software market is projected to grow from $3.2 billion in 2024 to $8.6 billion by 2029, at nearly 22% annually, within a digital economy that surpassed $300 billion in GMV in 2025.^3,4 The growth is structural, driven by digital transformation and an enterprise base still early in its adoption curve. Vertical SaaS platforms across the region attracted over $4.54 billion in cumulative funding over the past decade, with investor conviction accelerating.³

The platforms capturing the most defensible positions are not the ones that arrived in the most accessible markets. They are the ones that built inside the most complex ones.

Markets with higher regulatory friction, localisation requirements, and deeper product integration demands act as natural filters. They discourage generalist entrants and reward operators who invest in genuine product adaptation and local partnership infrastructure. Once embedded in a customer's compliance workflows or supply chain operations, these platforms are functionally difficult to displace. Complexity is the moat, and it compounds. Vertical SaaS companies that own the core workflow in their market consistently achieve net revenue retention of 108 to 120%.⁵ Global incumbents face the same entry friction as everyone else in these markets, and they rarely prioritise it.

Capturing this opportunity requires dedicated local structure, milestone-based capital, and partnerships built on genuine operational presence rather than remote coverage. The organisations best positioned to help technology platforms navigate this are those that understand both the capital architecture and the market-entry execution required to convert local complexity into lasting competitive advantage.

Difficulty, entered deliberately, is a strategy.

It is worth acknowledging something plainly: private equity in Asia has operated for much of the past decade in unusually favorable conditions. Cheap debt amplified returns. Rising multiples did the heavy lifting. These were real structural tailwinds. They are now gone, and what replaces them is harder to manufacture: the operational ability to grow a business during the hold period, in markets that do not reward generalist approaches.

Bain's 2026 Global Private Equity Report frames this shift with precision. The phrase "12 is the new 5" captures the new return math directly. Where five percent annual EBITDA growth was once sufficient to generate target returns, today's deals, underwritten at higher entry multiples and financed at elevated interest rates, require closer to twelve percent annual EBITDA growth to justify the same outcome.¹ Operational value creation is all that remains.

The Asia-Pacific data makes this concrete. Fundraising for the region's focused funds fell to a twelve-year low in 2025, dropping 37% in value and 44% in fund count. The number of portfolio companies held for more than five years rose 18% versus 2024, as 2020-2022 vintage assets struggled to deliver the growth originally underwritten.² These numbers reflect not a market in structural decline, but a market actively separating into two tiers.

The same report documents what is happening on the other side of that divide. Exit value rose 24% in 2025, growing for the second consecutive year. Net distributions to LPs turned positive for the first time since 2021. At the end of 2025, 57% of GPs surveyed expressed positive sentiment about market conditions.² The 20 largest funds accounted for more than 50% of total capital raised, up from an average of 41% between 2020 and 2024, a clear signal that LPs are concentrating capital with managers who have demonstrated repeatable execution.²

Japan illustrates the point most sharply. The only Asia-Pacific market to deliver growth in both deal value and count in 2025, it rewarded investors who arrived with governance reform tailwinds, disciplined operating focus, and the infrastructure to move with conviction.² The lesson is not that Japan is structurally easier. It is that prepared operators outperform in any market.

Southeast Asia and the Middle East present the same dynamic at greater complexity. The six core ASEAN economies recorded over 3,200 transactions in 2025, with combined value exceeding $130 billion, up roughly 18% year-on-year, across markets growing at 4% to 7% GDP.³ Gulf markets are now institutionally sophisticated and actively competitive. The opportunity is large and credible. But Southeast Asia is not a single market. Regulatory environments, payment infrastructure, and consumer behavior shift meaningfully across every border. Expansion initiatives that compete for budget and leadership attention inside an existing portfolio company structure are disadvantaged before they begin.

BCG's analysis of the current PE environment is direct: with multiple expansion slowing, true operating skill is showing through, and it carries across a franchise, from deal to deal.⁴ The firms pulling ahead have built repeatable value creation systems, with plans built before close and executed from day one.

Asia's fundamentals, its demographics, its digital growth, its corporate reform agendas, remain among the most compelling in the world. What has changed is the mechanism by which returns are generated. The operators who understood this early are already compounding.

Hope is not an execution model.

Private equity has declared operational value creation its top priority. The data suggests it does not yet have the people to deliver it.

Simon-Kucher's PE Value Creation Study finds that 78% of deal teams expect operational improvements to become even more important over the next twelve months, a 20% increase over the prior year.¹ Yet KPMG's survey of 500 private equity leaders finds that just 18% of PE professionals are operating partners, with two-thirds of them simultaneously responsible for five or more portfolio companies.² The ambition is stated. The infrastructure to execute it is not there.

This is not a marginal inefficiency. It is the central capability problem in the industry right now. Gain.pro's Private Equity Value Creation Report, drawn from more than 10,000 global PE deals, found that 71% of all value created in 2024 exits came from revenue growth, up from 64% the prior year and higher than any period in the previous five years.² Revenue growth at the rates deals now require does not emerge from a part-time operating resource stretched across a portfolio. It requires dedicated execution capacity, structured from the start of the hold period, with accountability tied to specific milestones.

The return dispersion data makes the stakes visible. The gap between top and bottom quartile buyout fund performance for the 2021 vintage year stretched to nearly 14 percentage points, according to JP Morgan data cited by Moonfare.² That spread is not explained by deal sourcing. It is explained by what happens, or does not happen, inside portfolio companies during the hold.

KPMG estimates that operational expertise headcount across the industry would need to increase threefold to meet current value creation demands.² That is not a gap that closes through better intentions.

The firms that close it will compound. The ones that do not will keep extending hold periods and waiting for conditions to do the work.

The Gulf states have historically been understood as capital exporters: patient, sophisticated allocators into assets built and managed elsewhere. That model is being replaced, deliberately and at speed, by something more consequential.

Saudi Arabia is constructing the physical infrastructure on which AI economies run. HUMAIN, launched in May 2025 under the Public Investment Fund, is building a full-stack AI company covering data centers, cloud infrastructure, and AI models, with plans to develop 1.9 gigawatts of data center capacity by 2030.¹ The company has announced $23 billion in strategic technology partnerships and a $10 billion venture fund targeting AI startups globally.² In a joint venture with AMD, it plans to deliver 500 megawatts of capacity over five years. Google Cloud and the PIF have advanced a $10 billion partnership to build a global AI hub in the Kingdom.³ The Saudi data center market is growing at a 29% compound annual growth rate through 2030.⁴

For technology companies with genuine AI capabilities, the implication is immediate. Saudi Arabia is not waiting for global platforms to arrive on their terms. It is building the infrastructure to set those terms, and actively seeking partners who can operate within sovereign AI governance frameworks, adapt their products to local institutional requirements, and commit to long-term relationships rather than transactional sales. That kind of entry requires dedicated on-the-ground structure and capital committed to the partnership, not subject to reallocation when quarterly targets shift.

Technology spending across MENA is projected to reach $169 billion in 2026.³ The window to enter as a genuine partner in this buildout, rather than as a late-arriving vendor in a mature market, is open now. It will not stay open indefinitely.

The infrastructure is being built. The question is who builds with it.