Rethinking financial resilience in the AI era

This report combines FRP proprietary analysis with public market data and recent transaction benchmarks to provide a grounded view of how SaaS businesses are being valued today.

Rethinking financial resilience in the AI era Restructuring insight

AI meets financial reality

Investment is outpacing measurable returns We interviewed 500 business leaders and lenders to understand how AI adoption is affecting financial performance and access to funding. We explore how a restructuring mindset can help businesses identify vulnerabilities, control spending, strengthen financial resilience and protect value. Rapid artificial intelligence (AI) adoption forces mid-market businesses to balance steep infrastructure and software costs against uncertain short-term gains. Many are investing heavily in AI before they have resolved weaknesses in their data, systems and controls, leaving boards exposed to uncertain returns, committed cash and costs that extend well beyond the technology itself. Pressure is building as businesses must fund stronger cybersecurity, governance, compliance and human oversight whilst responding to changing competitor capabilities and customer demands. Lenders, shareholders and boards are also expected to provide a clear view of the full cost, expected returns and report on the effect on margin and cash flow. As investment returns, the challenge is no longer whether to invest in AI, but how to prove its value There is a widening gap between the huge investments being poured into artificial intelligence infrastructure and the lack of proportional, near-term financial returns. This ‘AI expectation squeeze’ is primarily driven by rising customer demands, surging infrastructure costs, complex implementation and lenders who now expect financial evidence to support sustained investment. Successful businesses will be defined by their understanding of the full cost, ability to evidence the benefits and the credibility of that evidence under lender scrutiny. Businesses that can report a clear case for their AI investment are starting to pull ahead because the benefits have been quantified and the costs have been captured. How businesses can manage the cost of AI adoption while protecting value, cash flow and financial resilience

Our findings at a glance

The commercial consequences of AI pressure are already emerging

85% Of lenders say AI is widening the gap between stronger and weaker businesses 66% Report more pressure to deliver more for the same price 78% Of lenders say AI readiness now matters to credit risk assessment 45% Of lenders say weak AI readiness could mean less favourable lending margins

63% Of leaders say AI has raised customer expectations for faster response 26% Name a gap in AI’s cost or margin return as their single biggest issue 5% Of lenders say businesses consistently evidence AI’s financial impact 40% Of leaders would want advice on measuring AI’s impact on margin

Rising expectations are reshaping customer demands Rising customer expectations are increasing pressure on businesses to deliver more for less Artificial intelligence is accelerating expectations, as consumers look for instant and hyper personalised responses. Around two in three leaders (63%) report higher expectations from customers for responsiveness, 60% for quality and accuracy, and 51% for personalisation.

Customer expectations

51% Higher expectations for personalisation

63% Higher expectations for responsiveness

60% Higher expectations for quality & accuracy

Financial and commercial pressures

50% Pressure to cut prices outright

45% Risk less favourable lending margins

66% Pressure to deliver more for the same price

Companies must find ways to deliver high value without raising prices, or they risk losing buyers to cheaper competitors. Customers now expect more, but many will not pay more for it. 66% of leaders face pressure to deliver more for the same price, and half (50%) face pressure to cut prices outright, putting margin under direct pressure. The race to meet rising expectations

The gap between expectation and delivery

Operating profit is shaped by a few critical factors:

As investment accelerates, many organisations are finding that converting AI adoption into measurable commercial value is harder than expected. 92% of leaders feel prepared to convert AI adoption into commercial value without eroding margin or increasing risk, yet in practice, integrating AI into existing systems and managing cyber, data and regulatory risk remain the two biggest barriers, each cited by 34% of leaders, with a further third (33%) still struggling to decide which tools are actually worth the investment. Asked where the gap between expectation and delivery is widest in their own organisation, leaders point most often to governance and money. 15% cite AI strategy running ahead of governance and accountability, an equal share cite technology capability lagging behind legacy systems, and another 15% cite cost savings that have not matched what was expected. Add margin improvement (12%) and productivity gains that have not translated into real operational change (10%), and the financial return gaps alone, cost and margin combined, are named more often than any single operational or workforce issue, while just 5% see no significant gap at all.

Where leaders say the biggest gap is between AI expectation and delivery

AI strategy vs governance and accountability

Technology capability vs legacy systems

Cost savings expected vs actually delivered

Margin improvement expected vs actually delivered

Leadership ambition vs workforce readiness

Productivity gains vs real operational change

Investment required vs available capital

Customer expectations vs ability to deliver

In terms of benefits, half of leaders (50%) report faster delivery, 44% report improved margins and 42% report more output without extra headcount. These are useful improvements, but they are largely self-reported perceptions and lenders are starting to press on AI reporting precision. AI is also forcing organisations to rethink the definition of value. The conversation is shifting from individual productivity to business level capability. Our research highlighted two areas of focus for boards and lenders alike. Forecasts need to show how expected efficiencies translate into revenue, margin, working capital and cash flow, and the gap between expected and actual returns needs treating as a live financial risk.

Costs businesses are most likely to miss

The true cost of AI goes beyond the technology

Lenders are looking beyond the headline cost of AI tools; 44% say businesses are most likely to underestimate cybersecurity costs, with data infrastructure and legal and regulatory compliance close behind at 36% each, and integration and licensing costs not far off.

AI-related costs lenders say businesses most often underestimate

Cybersecurity

Data infrastructure

Legal, regulatory and compliance risk

Integration with legacy systems

Software and licensing

Training and workforce trends

The financial impact of failed AI initiatives When artificial intelligence investments fail to meet expectations, the financial fallout has the potential to destabilise a business. Lenders expect under-performing AI investment to show up in core financial measures such as; weaker revenue growth (39%), lower margins (37%) and more cash flow pressure (34%). That pressure can influence the funding decision itself as, 45% of lenders say weak AI readiness could lead to less favourable lending margins, 43% say it would reduce their confidence in forecasts, and 35% say it could change the structure or timing of lending. Why cash flow becomes the pressure point Timing matters as much as total return as businesses typically spend the cash before the benefits arrive, liquidity can tighten even when the long term case still stands up. Cash flow forecasting and scenario planning should sit at the centre of AI governance. Boards need to understand what happens if benefits are delayed or reduced, what headroom exists and what triggers action. Lenders need the same evidence to assess covenants, borrowing needs and resilience.

Financial warning signs of AI under-performance Costs outpacing the business case Implementation and operating costs are rising beyond the assumptions made in the original AI investment case, eroding expected returns and extending payback periods. Benefits not translating into performance Forecast productivity gains, revenue growth or efficiency improvements cannot be clearly linked to improvements in margins, cash flow or working capital performance. Expectations running ahead of reality Customer demand and expectations are increasing faster than the business can adapt its pricing, capacity, service delivery or operating model. Lack of a credible recovery plan Management cannot demonstrate credible scenarios or agreed actions if expected AI benefits are delayed or fall short of expectations. Investment crowding out priorities AI spending is competing with other essential investments, diverting capital from core business needs and creating pressure on strategic priorities. Liquidity & covenant pressure emerging Sustained investment without corresponding returns is placing strain on cash resources, reducing financial flexibility and increasing covenant risk. Lenders are often engaged before an AI investment fails. 40% would recommend advisory support when costs exceed expectations, 33% when scenario planning lacks credibility, and 30% when cash flow comes under pressure.

7 9

Lenders are defining what credible AI investment looks like More than three quarters of lenders (78%) say AI readiness now matters when they assess credit risk, largely as a read on management capability and financial discipline rather than on the technology itself How often do borrowers clearly evidence claimed AI efficiencies in forecasts, margins or cash flow?

12%

42%

41%

Rarely/never

Always

Often

Sometimes

Financial evidence gives them the most confidence, 39% look for AI benefits reflected in forecasts and cash flow, 35% want realistic cost and return assumptions, and 31% look for evidence in improved profit margins. Against that bar, most businesses currently fall short, our survey reported that only 5% of lenders say businesses consistently evidence AI’s financial impact in their forecasts and cash flow, and most see it only regularly (42%) or occasionally (41%).

In practice, that means most mid-market businesses could not yet deliver a lending conversation with financial evidence that meets the bar lenders are looking for.

Winners and losers: the performance gap is already opening The “AI divide” is rapidly expanding, separating companies that successfully identify and scale value. Lenders are watching this shift closely, and most already see its effects Is AI widening the performance gap between stronger and weaker businesses in the sectors you lend to?

Yes, to some extent Yes, significantly

12%

No, but we expect it to No, and we don’t expect it to 85% said yes

21%

64%

85% say AI is widening the performance gap between stronger and weaker businesses in the sectors they lend to, with 21% describing that gap as already significant and only 3% ruling it out altogether. For boards, the real question now is whether they can prove the investment is paying off, with evidence that would hold up under scrutiny.

Your lender is already scoring your AI strategy

43.6% saw gains reach margins, but 34.0% saw them absorbed by workload and complexity, and 26.4% offset by rising costs elsewhere. Checking is the cost nobody budgeted for. The firms that industrialise assurance, meaning error budgets, sampling rather than universal review, and audit trails robust enough for a regulator, will be the only ones able to remove the human check and keep the gain. The unit of competition moves from task to operating model. Choosing between platforms concerns 33.2% of leaders today and will concern almost nobody by 2030, as capability converges and switching costs fall. A harder question replaces it: what proportion of your work runs from end to end without a human in the loop, and what error rate will you accept in exchange? Assisted tasks build no compounding advantage, because the same assistance is available to everyone at the same price. Capital reprices before the market does. 78.1% of lenders already treat AI readiness as a significant or moderately important credit factor, and 96.8% either see the gap between stronger and weaker businesses widening or expect it to. The sequence runs from informal question, to diligence heading, to information requirement, to facility term, and each step is invisible from inside until a refinancing or a sale, when it arrives as a price rather than a conversation. Only 47.4% of lenders say borrowers consistently evidence claimed AI efficiencies in forecasts, margins or cash flow. That evidence is missing because most management teams do not realise it is being read.

Andy Pardoe Founder & Chief AI Officer at KARRIK

Background Two findings in this research sit directly opposite one another, and the gap between them is where the next five years will be decided. Ask mid-market leaders which AI issues most challenge their strategic and investment decisions, and scenario planning for under delivery comes last, at 15.2%. Ask where external advice would be most valuable, and lender and stakeholder conversations come last again, at 20.8%. Now ask the people lending them money. A third (33.5%) name the absence of credible scenario planning as a reason to send a borrower for outside help, and 45.4% say weak AI readiness could affect the pricing and margin on their lending. Management ranks these issues last; capital ranks them close to first. AI performance will show up in the cost of capital before it reaches the profit and loss account, and most boards are watching the wrong one. Three forces that will shape the next three to five years Verification becomes the binding constraint. 42.0% of organisations are increasing oversight of AI-enabled work; only 20.0% are reducing headcount. That is a supervision story, not an automation one, and it explains where the efficiency went.

What failure to adapt looks like Not obsolescence. Mid-market firms are rarely destroyed outright by technology shifts. They are quietly repriced instead: the cost of capital first, then customer contract terms, then talent, then exit multiples. None of these announce themselves in a board pack. And a service that has not deteriorated still becomes a worse service once the baseline moves. liability of autonomous agents, concentration risk among model suppliers, and insurance cover for autonomous action. Self-reported preparedness stays close to today’s 91.9% and stops distinguishing between firms, giving way to measures that are harder to flatter: the share of work running without human intervention, and the error rate the business accepts. The proportion reporting improved margins falls rather than rises, because efficiency reaches customers before it reaches shareholders. 65.6% already report increased pressure to deliver more for the same price. Questions arrive that cannot sensibly be asked today: the authority and What the same survey will find in 2030 As a thought experiment, here is how the same survey might read in four years’ time.

Experiment or advantage One test separates the two. If a competitor bought the identical licence tomorrow, how much of your gain would survive? Whatever survives is strategy; whatever evaporates is learning. Learning is worth funding, but it should be reported as learning rather than presented to a board or a lender as competitive position. Five decisions worth making in the next twelve months Put AI benefits and costs into the forecast rather than the narrative. Create a named oversight cost line, together with a plan to retire it. Model the case where benefits land at half of plan, before anyone external asks for it. Decide explicitly which gains you keep and which you pass to customers. Pricing strategy is now AI strategy. Give one board-level owner clear accountability for AI decisions.

How FRP can help

Companies and stakeholders’ ability to quickly and appropriately respond to business challenges and opportunities is critical to protecting and driving value. A restructuring mindset can help businesses identify vulnerabilities and plan for challenges ahead. Our skilled teams provide advice and help to review financial and operational performance, test investment assumptions and identify what is actually driving the pressure, giving boards clear options. We help manage debt, liquidity and cash flow, including refinancing, debt renegotiation, capital structure reviews and stakeholder negotiations. We work with management and stakeholders to simplify operations, reduce cost and address under-performing parts of the business, and help businesses return to financial health. As AI grows, the companies that are best placed to capture the opportunities will be those that quantify costs, validate returns and respond early to pressure. Alongside huge potential upside, artificial intelligence is also bringing vast challenges to businesses through soaring costs, rising operational risks, a changing competitive landscape and accelerated customer expectations

Get in touch

Whether you are assessing the return on AI investment, responding to increasing lender scrutiny or managing the impact on margins and cash flow, our Restructuring Advisory team can help. Get in touch with our experts to discuss the challenges facing your business and the practical steps you can take to protect value and strengthen resilience.

Luke Wilson Partner Restructuring Advisory London +44(0)7789 615 744 luke.wilson@frpadvisory.com

Shaun Hudson Partner Restructuring Advisory Newcastle +44 (0)7484 040 615 shaun.hudson@frpadvisory.com

The research was conducted by Censuswide on behalf of FRP Advisory across two UK audiences: 1) 250 senior leaders and board members at UK mid-market companies with 50 to 500 employees. Fieldwork took place from 29 June to 6 July 2026. 2) 251 UK lenders and investors involved in lending and investment decision making, including private equity lenders, asset based lenders and other lender and investor types. Fieldwork took place from 26 June to 3 July 2026. Censuswide is a member of the Market Research Society and the British Polling Council, and a signatory of the Global Data Quality Pledge. It adheres to the MRS Code of Conduct and ESOMAR principles. Percentages have been rounded to whole numbers in the narrative. Multi select questions may total more than 100%.

FRP Advisory Trading Limited

110 Cannon Street, London EC4N 6EU

FRP Advisory Trading Limited is a company incorporated in England and Wales registered number 12315855.

Any use of this publication or reliance on it for any purpose or in any context is at your own risk, without any right of recourse against FRP or any of its partners, employees or agents. Copyright © September 2026 FRP. All rights reserved. Published in the UK

For media enquiries, please email press@frpadvisory.com

September 2026

Page 1 Page 2 Page 3 Page 4 Page 5 Page 6 Page 7 Page 8 Page 9 Page 10 Page 11 Page 12 Page 13 Page 14 Page 15 Page 16

Made with FlippingBook - professional solution for displaying marketing and sales documents online