In conversations with founders and CEOs of European Credit & Lending software businesses, I keep noticing the same shift. A year ago, AI was mostly something to think about. In 2026, it has become part of everyday work. Founders are no longer asking whether AI matters. They are asking which parts of their product deserve real investment, and what is still hype.

What strikes me is how grounded these conversations have become. Founders are not chasing the latest model. They are looking carefully at the work their customers do every day and asking where AI can remove friction without adding regulatory risk or losing customer trust.

I subscribe to the theory put forward by the economist Carlota Perez, among others. While AI is clearly profound in terms of its potential business impact, it is also the latest phase in a much longer wave of technological change that began decades ago with the emergence of the information technology age. What history shows us is that software companies have consistently played an important role in interpreting new technologies and translating them into practical solutions to real business problems within specific markets. That was true twenty years ago, it is true today, and in my view, it will remain true five to ten years from now.

The implication for Credit & Lending software founders is clear. The leaders who understand their market better than anyone else, who know how a mutual guarantee institution processes a guarantee request or how a bank’s back office handles a credit file, are uniquely well placed to put AI to work for their customers. The tools may be new, but the dynamics are familiar. Those who actively explore and adopt AI will continue to be successful. Those who do not will increasingly struggle to remain relevant.

At Upliift, we partner with businesses who are actively adopting AI. We are a Permanent Equity investor with no fixed exit timetable, which means we think in decades rather than quarters. What is happening in European Credit & Lending software in 2026 is not simply one AI trend, but five connected shifts:

1. AI is no longer a pilot project; it is a baseline expectation

The first shift is the clearest. The European Banking Authority reports that almost all EU banks are now using AI in some form. The strongest production use cases are not autonomous credit decisions. They are the practical, document-heavy work that surrounds a credit decision: onboarding, KYC, anti-money-laundering checks, fraud alerts, customer support, and help for the analysts who still make the final call.

The value is also starting to show up in disclosures. Lloyds Banking Group reported around £50 million of value from generative AI in 2025 and expects more than £100 million in 2026, after rolling out more than fifty use cases. That is not a pilot budget. It is a major European bank publicly committing to AI as a measurable contributor to results, and it is a signal that other large European lenders are watching closely.

For Credit & Lending software vendors, the implication is direct. Buyers are no longer impressed by the fact that a platform has AI features. They want to know how those features behave in their real workflows, with their real volumes, and under the eyes of their real auditors.

Andrea Gelfi, CEO of Galileo Network, the Italian financial services specialist serving banks, guarantee institutions, and confidi, describes the shift from inside an operator.

“Artificial intelligence has now become an indispensable element, and Galileo has been studying it for some time now and has begun integrating it into its internal production cycles. The aim is to develop new services for the benefit of our market and our customers.”

2. Responsible AI is becoming part of the buying decision, not just a compliance task

The second shift is the most distinctly European. The EU AI Act treats systems that decide the creditworthiness of a private individual as high-risk, and the main rules of the Act start applying from 2 August 2026. GDPR continues to apply alongside it.

What matters for Credit & Lending software vendors is how this is showing up in procurement. Your customers are now asking harder questions about how a decision can be explained, how a human can intervene, and how the whole process can be reviewed afterwards. The FCA’s 2025 research on credit-decision explainability adds a useful twist. Giving consumers more information about how an algorithm works can, in some cases, make outcomes worse rather than better. Explanation has to be designed, not just disclosed.

Andrea Gelfi puts it plainly.

“As we operate in Financial Services, which is a regulated sector, we believe AI must be used carefully and responsibly in full compliance with regulations. It must not involve any unlawful data processing to protect our customers.”

Compliance is no longer something the legal team handles after the product is built. It is becoming a product feature in its own right, and increasingly part of how customers choose between vendors.

3. Document-heavy work is where the economics are clearest today

The third shift is the one with the clearest near-term return. Credit & Lending workflows are, at heart, document workflows. Bank statements, payslips, bureau reports, KYC and KYB files, sanctions hits, policy manuals, call notes, exception memos. Each is manual work that AI can now meaningfully reduce.

Deloitte estimates that AI-supported document analysis can cut manual processing by up to half per application, and that better monitoring can spot signs of borrower stress more than a year earlier than manual review. These are consultancy figures rather than audited benchmarks, but they line up with what we see across the European market. Document work is attracting AI investment first because the value is easy to measure, the risk is contained, and the work can still be checked by a human.

Galileo’s own situation is a good example. Many credit workflows in Italy and across southern Europe still depend on paper, manual signatures, and slow back-office handovers between institutions. The opportunity is not to replace people. It is to remove the friction around them so the people can do better work.

“We operate in a sector that is fundamentally based on digital processes. However, the relationship between our customers and their markets is still influenced by habits that rely heavily on paper documentation and manual signatures. For several years, Galileo has been pursuing an innovation strategy aimed at moving towards a fully digital process.”

-Andrea Gelfi, CEO of Galileo Network

This is where Credit & Lending software vendors should look first. Faster application intake, cleaner document checks, quicker case summaries for the people making decisions, and better handovers between the front office and the back office. None of this requires the most ambitious AI. It requires the right AI, applied to the right step, with discipline.

4. Specialist advantage must be turned into something durable, or it will be lost

The fourth shift is the most important strategically. Specialist Credit & Lending software businesses already have what AI-native challengers do not. You understand your market in detail. You hold years of proprietary data. Your customers trust you with regulated work. You know the regulations as well as they do. These are real advantages, and they are difficult to replicate from outside.

But they do not automatically turn into an AI advantage. McKinsey’s recent work on generative AI in the credit business makes a similar point: many institutions are still struggling to scale beyond pilots because the value of AI depends on workflow redesign, not feature deployment. BCG describes 2026 as the year when banks must move from experimenting with AI to running it at scale, with cost and revenue effects that show up on the P&L.

The investor Hg has framed this most directly. The head start vertical software businesses enjoy is real, but it is only a head start. The companies that turn it into something durable are the ones that put AI inside the workflows their customers care about most, then build the discipline to know whether the AI version is actually better at the work the customer pays for.

The risk of moving too slowly is concrete. If your customers cannot get the help they need from you, they will start trying other tools. Sometimes, they build general-purpose AI tools themselves. Sometimes, newer AI-native vendors. Once a customer has built a habit around a different way of working, that habit is much harder to win back than it was to lose.

5. Speed, control, and trust are what will decide the winners

The fifth shift will probably decide which Credit & Lending platforms scale and which stall. Buyers are moving from “do you have AI features” to “can you run AI safely inside our regulated work, prove what it is doing, and let us see inside it.” Two-thirds of UK financial services firms already get AI from third parties, and the European Central Bank has warned about the concentration risk this creates. Procurement teams are starting to pay attention.

In practice, this means the credible 2026 approach is not a single clever model. It is a layered one. Traditional scoring models continue to do the hard quantitative work of deciding who can borrow and on what terms. AI sits on top of that, handling the surrounding work: pulling the right document, drafting the case summary, flagging the unusual transaction, explaining the decision in plain language. A clear record runs through both layers, so the whole process can be reviewed.

The unglamorous work is what compounds over time. The founders who put the ‘boring’ foundations in place now, including clean data, clear records, and a sensible way of checking whether each AI feature is getting better, will find that every new feature ships more easily. The ones who skip this step will find that each new feature is harder than the last.

“We aim to offer more efficient services and to shorten the entire guaranteed credit value chain. Our focus must be on increasing the speed of services that enable our customers, particularly the confidi, to receive applications from their own clients more quickly. Digitalisation, combined in the future with AI-driven services, will allow us to achieve this goal.”

— Andrea Gelfi, CEO, Galileo Network

That is what good looks like in 2026: shorter cycle times for the customer, fewer manual steps in the back office, and AI that can be trusted by the people who depend on it.

A founder self-check

The companies that scale and the ones that stall are rarely separated by awareness of these trends. Most founders we speak to can describe them well. The difference is execution: whether the business has done the work of preparing its product, its data, and its people to absorb the change.

A few honest questions worth asking yourself:

• If a regulator asked tomorrow how an AI feature in your platform influenced a credit or compliance decision, could you explain the logic and the human override path clearly enough for an auditor?
• For the two or three customer workflows where you create the most value, do you know exactly where AI could remove friction, and how you would measure the improvement?
• When you ship an AI feature, what is the evaluation method that tells you whether it is getting better at the specific work your customer pays you for? Hg Catalyst calls this the question that separates durable AI advantage from a thin wrapper over the latest model.
• How much of your data sits in a single, organised place, and how much is spread across systems you do not control?
• If you imagine the same product in 2028, under the full AI Act high-risk regime, does your current setup make compliance easier or harder?

Working alongside founders for the long term

Upliift is a Permanent Equity partner for specialised European software businesses in regulated industries. We have no fixed exit timetable, which shapes how we work. We invest alongside founders for the long term, focusing on the fundamentals that compound over time: the strength of the team, the depth of customer relationships, the quality of the culture, and the durability of margins.

Our philosophy has been the same since the very beginning. We back and support founders and CEOs to run their businesses, while preserving a strong sense of ownership and accountability for their success. Our role is to augment the leaders we work with: giving them more clarity and structure on the drivers of value in their business and offering hands-on support where we are invited in to help close knowledge gaps around best practice, go-to-market, or, increasingly, AI. We are being asked more often to partner with our founders to explore how AI will play out for their business in their market, to share AI use cases, learn from one another, and build the kind of peer-to-peer relationships that help each business move faster than it could on its own. That is the model we believe in, and it is only the beginning.

If you are a founder or CEO building specialised Credit & Lending software in Europe, and you are thinking about what the next chapter could look like, we would welcome an honest conversation. No pitch, no pressure. Just a direct exchange between people who care about building durable European software businesses.

This blog is written with input from Andrea Gelfi, CEO of Galileo Network. It’s a part of the series exploring the forces shaping European Credit & Lending software. Read the other posts: The Long-Term Partner for Credit & Lending Founders & Five Forces Reshaping European Credit & Lending Software — and What They Mean for Founders.

References

European Banking Authority. ‘Rising application of AI in EU banking and payments sector’. September 2025. https://eba.europa.eu/sites/default/files/2025-09/146b3558-d026-47bf-a872-f05e93ed30d2/Rising%20application%20of%20AI%20in%20EU%20banking%20and%20payments%20sector.pdf

European Banking Authority. ‘AI Act: implications for the EU banking and payments sector’. November 2025. https://eba.europa.eu/sites/default/files/2025-11/d8b999ce-a1d9-4964-9606-971bbc2aaf89/AI%20Act%20implications%20for%20the%20EU%20banking%20sector.pdf

European Commission. ‘Annex III of the EU AI Act’. https://ai-act-service-desk.ec.europa.eu/en/ai-act/annex-3

Bank of England and FCA. ‘Artificial intelligence in UK financial services – 2024’. November 2024. https://www.bankofengland.co.uk/report/2024/artificial-intelligence-in-uk-financial-services-2024

FCA. ‘Credit where credit is due: how can AI’s role in credit decisions be explained?’. February 2025. https://www.fca.org.uk/publication/research-notes/how-ai-role-credit-decisions-explained.pdf

European Central Bank. ‘The rise of artificial intelligence: benefits and risks for financial stability’. May 2024. https://www.ecb.europa.eu/press/financial-stability-publications/fsr/special/html/ecb.fsrart202405_02~58c3ce5246.en.html

McKinsey. ‘Banking on gen AI in the credit business: The route to value creation’. July 2025. https://www.mckinsey.com/capabilities/risk-and-resilience/our-insights/banking-on-gen-ai-in-the-credit-business-the-route-to-value-creation

Deloitte. ‘Unlocking the benefits of GenAI for next-generation lending’. March 2025. https://www.deloitte.com/ch/en/Industries/financial-services/blogs/unlocking-benefits-of-genai.html

BCG. ‘For Banks, the AI Reckoning Has Arrived’. May 2025. https://www.bcg.com/publications/2025/for-banks-the-ai-reckoning-has-arrived

Lloyds Banking Group. ‘Lloyds expects over £100 million in value from next-generation AI in 2026’. January 2026. https://www.lloydsbankinggroup.com/media/press-releases/2026/lloyds-banking-group/ai-driven-benefits-2026.html

Hg. Dr Amr Ellabban. ‘The question to ask every software company you’re invested in’. May 2026. https://hgcapital.com/insights/the-question-to-ask-every-software-company-youre-invested-in

Hg. Chris Kindt. ‘The agentic flywheel: building the next decade of product differentiation in B2B Tech’. May 2026. https://hgcapital.com/insights/the-agentic-flywheel-building-the-next-decade-of-product-differentiation-in-b2b-tech