B2B payments software expert Inez Berkhof-Hollander on the huge potential of applying machine learning and other advanced data techniques to core finance processes.
Despite years of investment in automation and digital transformation, CFOs are facing many of the same order-to-cash (O2C) challenges in 2026 they faced two decades ago. External business conditions are becoming more challenging, while competing priorities leave little room to introduce further inefficiency into existing systems. This is reflected in findings from a survey of over 500 finance leaders across the UK, Spain, France and Germany, examining how organisations are managing B2B payments, invoicing, and broader back-office technology adoption.
So what can finance teams do to improve their payment processes? Increasingly, organisations are seeking to drive greater efficiency by making better use of their existing digital infrastructure and the growing capabilities of AI, rather than embarking on costly technology overhauls.
However, several persistent issues continue to create friction. Incorrect invoices, limited ERP integration, inconsistent invoice formats and approval delays remain particularly challenging for the business leaders we spoke to. For example, when asked what increases their likelihood of making repeat purchases, just over half of respondents cited invoicing-related factors such as invoice customisation, namely the ability to configure invoices to the buyer’s specific organisational requirements, which was rated important by four in five respondents (80%+) overall, rising to 93% in Spain.
Delays in approval workflows and difficulty matching invoices to purchase orders were the two most common invoicing challenges, each cited by 34% of respondents. Both inconsistent invoice formats and limited ERP integration are considered problematic by 31%, while well over half of our 500-plus respondents (57%) encountered some form of payment limitation with at least one supplier during the period studied. In Germany, the percentage rose to 76%.
We also found fast, frictionless onboarding is one of the most effective ways for suppliers to secure repeat business. Efficient digital B2B payment onboarding is a particular priority in the UK, where the speed of credit approval and account set-up emerged as a competitive factor in its own right.
But the technology dominating every conversation today is AI. Our research suggests adoption is already well underway. Around half of respondents reported using AI tools often in their purchasing and payment processes, while a further three in 10 said they use them routinely.
A clear picture is emerging of mainstream B2B organisations embedding AI into everyday finance operations. They are using it to improve decision-making through richer data insights, strengthen fraud prevention and risk management and streamline, or even eliminate, the manual tasks that slow down payment workflows.
AI solutions provide real-time invoice status visibility and automatically match invoices to purchase orders are in high demand as they directly address long-standing operational friction. Evidence is mounting that AI can be of help in a key area of payments: order-to-cash (O2C), the end-to-end business process from customer order through to payment receipt and reconciliation.
O2C, an engine for B2B growth
The B2B sector is already actively ground-testing AI capabilities that are set to transform O2C processes as we know them today. These applications include improving cash application and invoice-matching automation, predicting payment behaviour and identifying potential slow payers earlier in the cycle.
Imagine what this kind of capability could enable. It could help intelligently prioritise collector queues, reduce invoice processing times by as much as 15–35% (based on field trials of a service we plan to release), surface dispute patterns for earlier resolution and even sensitively personalise dunning, namely the systematic accounts receivable process designed to improve cash flow while preserving customer relationships.
In other words, this shifts O2C from a back-office function to a measurable driver of revenue growth, harnessing AI to turn payments into a genuine source of business value. Our trials have surfaced what we term an AI- and data analytics-driven ‘Detect, Act, Grow’ cycle. In this model, machine learning detects accounts that are going quiet, takes timely action through targeted interventions and grows revenue by improving collections performance and reducing risk, supported by AI-driven behavioral insights.
Again, consider the potential of AI in the O2C cycle. By analysing recent spending patterns and the frequency of use of credit, it becomes possible to identify behavioural shifts well before they would typically be visible to field sales teams. In fact, we’ve already built systems for users that gently encourage strong customers to re-engage through mechanisms like an open-to-buy email outreach, targeted rebate offers or a personalised engagement campaign—all with minimal human effort.
AI-optimised rebate structures are used to continuously tune customer incentives, encouraging spend while protecting the underlying commercial relationship. The model can work on organisations’ behalf to identify where there’s room to grow in any commercial relationship, deploying the most appropriate or effective incentive at the right moment.
Using payments tech as the basis for stronger relationships
We are also seeing the emergence of AI-powered O2C support tools, including engagement managers that enable self-service creation of campaigns to reactivate dormant buyers, send credit-utilisation nudges, and run marketing-style communications with minimal or no human intervention.
These systems can also dynamically orchestrate incentive and rebate campaigns tied to volume and spend behaviour. This development points to the emergence of a new era where buyers will be able to benefit from personalised rebates and adjust their spend in ways that suit their needs, while businesses drive continued engagement and reap additional sales through better use of data.
In one case, we worked with a major US retailer to send targeted emails to at-risk clients that otherwise might have gone dark, such as reminders of remaining credit. The firm reported no fewer than 59 dormant customers purchased within eight days of the campaign. In another instance, we worked with a large US manufacturer to reinvigorate buyer spend with similar AI prompts, increasing sales growth by 14%.
B2B organisations can, and should, become less conservative about the contribution of AI to payments and O2C efficiency. Done well, this not only supports profitability, but also strengthens customer relationships and improves the quality of supplier partnerships. In this sense, late payments may soon be very much in the rear mirror.
Inez Berkhof-Hollander is EMEA Vice President at the global B2B payments network, TreviPay.
For more on the themes discussed here, you can access the complete TreviPay EMEA market research report here.
