2025 is bringing a fundamental shift in how enterprises use artificial intelligence. According to Gartner, systems capable of autonomous decision-making and task execution are the number one trend in technology. The AI agents market has reached a value of 7.38 billion dollars, with forecasts pointing to 103.6 billion dollars by 2032.
Crucially: 74% of organisations achieve a return on investment within the first year of deployment, and leaders report a fivefold increase in revenue and a threefold reduction in costs.
For Polish enterprises facing the mandatory rollout of KSeF in February 2026, AI agents are no longer optional - they are a strategic necessity. This article shows how to turn a regulatory obligation into a competitive advantage.
How does an AI agent differ from existing automation
Traditional RPA automation works on rigid “if-this-then-that” rules. A robot performs a task exactly as it was programmed to - when something changes, it requires manual reconfiguration.
AI agents, by contrast, think. They use large language models to understand context, reason and make decisions in situations they have never seen before.
McKinsey & Company puts it precisely: agents need tools, and those tools will be robots - because an agent must break a complex workflow down into tasks and subtasks, and at the subtask level the only way to achieve deterministic behaviour is through robots.
This is not RPA being replaced by AI, but their orchestration:
-
Agents think and decide
-
Robots execute with precision
-
People supervise and handle escalations
In practice, this means the difference between a robot that can post an invoice according to a rigid pattern, and an agent that can analyse an unusual situation - for example, a dispute with a supplier over an amount - verify transaction history, consult other agents responsible for collections or customer service, and propose a solution. All within seconds, instead of hours of manual work.
Technology giants are betting on agentic capability
SAP is announcing more than 400 AI features by the end of 2025, with Joule - an intelligent assistant - as the entry point into the entire ecosystem. Joule is not a chatbot, but an omnipresent assistant that understands business context, has access to the SAP Knowledge Graph - fifty years of business-process knowledge - and coordinates specialised agents.
The Cash Management Agent saves up to 70% of the time spent on cash positioning. The Dispute Resolution Agent automatically analyses invoice disputes and proposes resolutions, cutting service costs by 30% and customer churn by 10%. The Performance and Goals Agent in SAP SuccessFactors prepares managers for performance reviews, automatically gathering feedback and generating personalised talking points.
UiPath, RPA leader for seven consecutive years according to Gartner, is entering its “Act Two” - a transformation from a robotic-process platform into an agentic automation platform. The key is Maestro - an orchestration layer that coordinates AI agents, RPA robots and people across long-running business processes.
UiPath Autopilot generates workflows from natural language with a 70% acceptance rate. Integration with Anthropic’s Claude 3.5 Sonnet speeds up processing of medical documents by 70%, while a partnership with Microsoft enables two-way integration with Copilot Studio.
Salesforce is targeting a billion AI agents by the end of fiscal year 2026. Microsoft is introducing 11 specialised agents in Security Copilot. Google Cloud is developing an inter-agent protocol for interoperability. Anthropic is positioning the Model Context Protocol as a universal integration standard - already supported by Microsoft, OpenAI and Amazon.
Concrete applications in areas key to SNOK
Finance and controlling: the end of month-end close as a marathon
AI agents are revolutionising finance and accounting processes. Advanced systems reduce month-end close time by 75% through automation of reconciliations, journal-entry proposals and variance analysis directly within ERP systems.
Agents monitor data in real time and automatically create accounting adjustments. Specialised solutions use artificial intelligence for lease accounting, contract analysis and revenue recognition with built-in accounting logic.
The greatest potential lies in automating procure-to-pay and order-to-cash processes. Agents process invoices from multiple channels:
-
Email
-
Supplier portals
-
EDI
-
PDF documents
They then normalise the data, verify compliance with policies and regulations, detect fraud and route items for payment. Autonomous negotiation agents carry out thousands of negotiations, achieving measurable reductions in category costs. Global corporations generate tens of millions of dollars in savings annually thanks to AI in procurement.
Polish enterprises face a unique opportunity: using the KSeF rollout as a catalyst for broader automation. Rather than treating KSeF as a purely technical obligation, companies can use it to transform the entire process, achieving not only regulatory compliance but real business value.
Human resources: from administration to strategic partnership
In HR departments, AI agents are taking over tasks that traditionally consumed 90% of the time. Agents handle questions about pay, benefits, leave and payroll calculations, while manager assistants support them in managing teams.
The Performance and Goals Agent - available from November 2025 - flags upcoming reviews, automatically gathers feedback from colleagues, generates personalised talking points and suggests follow-up actions.
Recruitment is accelerating dramatically:
-
Generating job descriptions takes 85% less time
-
Candidate assessment takes 80% less time
-
Time-to-hire shortens by 75%
According to LinkedIn, 73% of organisations use AI in recruitment, and by 2025, 80% of Global 2000 companies will be using algorithmic hiring systems.
Payroll automation cuts processing time by 70% and errors by 90%. Multi-agent systems use specialised agents for data integration, compliance checking and real-time support. The best solutions offer a 100% payroll-accuracy guarantee.
Security: from reactive response to proactive defence
AI agents are transforming security operations centres from an “operator versus threat” model into an “orchestrator managing a fleet of agents” model. Analysts manage multiple agents the way a manager manages a fleet.
Multi-agent collaboration includes:
-
An alert-triage agent
-
An investigation agent
-
A malware-analysis agent
-
A threat-hunting agent
Each specialised, all working together.
The effects are dramatic: an 80% reduction in alert noise, 52% faster response times, 99.6% threat-detection accuracy. In one university case study, 74,826 out of 75,000 alerts were resolved automatically, leaving only 174 for manual analysis.
Advanced systems autonomously analyse millions of events, identifying genuine threats and blocking thousands of malicious messages that older filters missed.
For SAP security, SecurityBridge is of particular importance - a leading platform securing more than 5,000 SAP systems globally. A new AI-powered code vulnerability analyser explains complex ABAP code, describes vulnerabilities and recommends fixes. Integration with Microsoft Sentinel provides unified visibility for the security operations centre.
Clients use SecurityBridge for:
-
Unified threat detection
-
Vulnerability management
-
Automated monitoring
Automation of regulatory compliance is becoming a necessity in the context of:
-
NIS2 - mandatory since October 2024, fines up to 10 million euros
-
DORA - in force since 17 January 2025, fines up to 2% of global turnover
AI agents automate risk assessments, vulnerability scanning, security updates, anomaly monitoring and incident reporting. Platforms offer automatic compliance across multiple standards simultaneously.
KSeF as a catalyst for AI adoption in Poland: an opportunity, not just an obligation
The Polish AI market is at a stage of cautious but accelerating adoption. Only 5.9% of Polish enterprises have fully adopted AI, compared with an EU average of 13.5%, but year-on-year growth stands at 36%. Poland ranks 24th out of 27 on the digital economy and society index.
Interestingly, the defence sector achieves 71% adoption, manufacturing 47%, and finance 40% - showing that when a clear business case emerges, Polish companies can move quickly.
KSeF - the National e-Invoicing System - is a turning point. From 1 February 2026, large enterprises with sales turnover exceeding 200 million złoty must mandatorily issue and receive invoices through KSeF. From 1 April 2026, all other businesses except micro-enterprises must comply. This affects more than 1.8 million VAT taxpayers.
The problem is that, according to a 2024 survey, only 42% of entrepreneurs are aware of KSeF, and only 11% have begun preparations.
KSeF requires:
-
A structured XML format (FA(3))
-
Integration via API 2.0
-
Certificate-based authentication
-
Validation against more than 80 compliance checkpoints
For companies with legacy systems, this represents significant complexity. But it is also an opportunity. Rather than a minimal integration “just to meet the requirement”, enterprises can use KSeF as an entry point into comprehensive automation of accounts payable and accounts receivable.
How AI agents support KSeF compliance
Data extraction and validation: AI-assisted text recognition extracts data from PDF files, scans and EDI, transforming it into XML format. Pre-validation checks more than 80 compliance points before submission to KSeF, reducing errors and rejections.
ERP integration: Seamless connection to SAP, Oracle, Microsoft Dynamics and NetSuite. Intelligent routing and approval workflows. Real-time status tracking. SAP modules offer dedicated KSeF handling with the option to extend for local requirements.
Multi-system orchestration: KSeF does not operate in isolation. Connecting to accounting, banking and procurement systems requires orchestration. Platforms such as UiPath Maestro coordinate workflows between systems, automating reconciliations, payment-batch creation and exception handling.
Compliance management: Continuous monitoring of regulatory changes, multi-country compliance - Poland, Italy, Germany, France, Belgium - comprehensive audit trails, automatic QR-code generation for offline invoices.
Polish companies that move quickly gain incentives: VAT refund times shortened from 60 to 40 days, exemption from submitting JPK_FA, and better cash-flow management. That is real business value beyond mere regulatory compliance.
SNOK and KSeF: automation through the UiPath platform
At SNOK we approach KSeF implementation as a comprehensive digital transformation project, not merely a technical integration. We use the UiPath platform - a leader in automation - as the foundation for the entire process.
Our strategy includes:
Intelligent document processing: UiPath IXP (Intelligent Xtraction & Processing), using advanced AI models to extract data from invoices in any format - PDF, scans, photos, electronic documents. The system automatically recognises line items, validates the data and transforms it into the required XML format.
Process orchestration: UiPath Maestro coordinates the entire invoice pathway - from receipt, through data extraction, KSeF compliance validation, integration with the accounting system in SAP or another ERP, all the way to submission to the National e-Invoicing System. Every step is monitored, with exceptions automatically escalated to the right people.
Systems integration: Seamless connection to SAP via dedicated connectors, using Document and Reporting Compliance and the eDocument Framework. For clients on other platforms - Oracle, Microsoft Dynamics, NetSuite - we provide analogous integrations with full support for Polish legal requirements.
Security and compliance: Every step of the process is secured - from authentication using qualified certificates, through encrypted API connections, to complete audit trails. We integrate SecurityBridge for SAP environments, ensuring the highest level of protection against cyber threats.
Competitive advantage: Our experience in business process automation, knowledge of the specifics of the Polish market and regulations, and status as a UiPath Agentic Automation Fast Track Partner allow us to deliver solutions that not only meet KSeF requirements, but become a platform for further automation across the entire organisation.
SNOK clients gain not only compliance with the February 2026 deadline, but a foundation for building an intelligent enterprise powered by AI agents.
The global context: why now is a critical moment
Global market data is unambiguous. Google Cloud’s 2025 report shows: 52% of organisations have AI agents in production, 74% achieve a return on investment within the first year, and 39% report at least a doubling of productivity. PagerDuty research indicates that 62% of experts expect more than 100% return on investment, with an average expectation of 171%, and 192% in the US.
But there is a catch: MIT: 95% of generative AI pilots fail to achieve significant value. Only 5% of organisations - “future-built companies” - achieve dramatic results: a fivefold increase in revenue, a threefold reduction in costs. The difference lies not in the technology, but in execution excellence.
Leaders do three things differently:
First, focus instead of dispersion: an average of 3.5 use cases compared with 6.1 for others. Choosing one problem and executing it flawlessly outperforms scattered experimentation.
Second, they buy specialised solutions: 67% success rate compared with 33% for in-house builds.
Third, they allocate 70% of effort to people, processes and culture, not just technology.
McKinsey’s internal system - 12,000 agents used by 70% of its 40,000 employees - saves 30% of time on routine work. BCG generates 2.7 billion dollars from AI services, representing 20% of total revenue - up from zero two years ago. Deloitte’s internal system cuts costs by 25% and raises productivity by 40%.
The competitive gap is widening. Future-built companies invest 26% more in IT, dedicate 64% of their budget to AI, and achieve compounding returns. The window is narrowing.
Gartner warns: more than 40% of agentic AI projects will be cancelled by the end of 2027 due to rising costs, unclear value and inadequate risk controls.
How SNOK sees the Polish market: a realistic path to transformation
At SNOK we observe the Polish market from two perspectives: a realistic one, grounded in data, and an optimistic one, based on potential. Our 25-year presence in the Polish IT ecosystem gives us a unique vantage point on how enterprises can effectively navigate a transformation towards organisations built on AI agents.
Phase 1: Foundation - KSeF as a starting point (Q4 2025)
Our observation: Most Polish companies treat KSeF as a regulatory burden, not an opportunity. Only 11% have begun preparations, even though the deadline for the largest organisations is February 2026. This creates time pressure that could paradoxically become a catalyst for broader change.
SNOK’s approach: We use KSeF as a “Trojan horse” for automation. Rather than building a minimal integration just to meet the requirement, we design a solution that is ready for expansion from day one.
Our solution, built on the UiPath platform, offers:
-
Full automation of data extraction from invoices in all formats
-
Built-in validation of more than 80 KSeF compliance checkpoints
-
Ready-made integration with SAP and other ERP systems
-
An architecture prepared to add further processes without rebuilding
Key takeaway: Companies that treat the KSeF rollout strategically - not as an IT project, but as a business transformation - will build a competitive advantage that competitors will take years to close.
Phase 2: Expansion - full automation of finance processes (H1 2026)
Our observation: Once KSeF compliance has stabilised, companies discover that the infrastructure they have built can handle much more. A typical scenario: the finance department requests automation of further processes, having seen the time savings and error reduction.
SNOK’s approach: We expand automation across the entire procure-to-pay and order-to-cash processes. We use AI agents working alongside RPA robots to:
-
Automatically reconcile invoices against orders and delivery documents
-
Intelligently route items to the right approvers
-
Automatically resolve simple disputes and discrepancies
-
Optimise payment terms to maximise discounts
-
Predict cash flows
Expected results, based on benchmarks:
-
75% reduction in month-end close time
-
30% reduction in dispute-handling costs
-
70% time savings on cash positioning
-
50% reduction in days receivables outstanding
Phase 3: Security and regulatory compliance (H2 2026)
Our observation: With NIS2 and DORA in full force, automatic compliance stops being optional - it is a requirement. Particularly for companies that have already deployed automation, since an expanded attack surface demands proportionally stronger protection.
SNOK’s approach: This is where our core expertise comes in - 25 years of specialisation in SAP security and cybersecurity. We implement a protective layer that grows alongside automation:
SecurityBridge for SAP environments:
-
An AI-powered code vulnerability analyser for custom ABAP solutions
-
Continuous real-time threat monitoring
-
Automatic detection of anomalies in user behaviour
-
Integration with the Security Operations Centre
Agentic security operations centre:
-
Automatic alert triage and prioritisation - an 80% reduction in noise
-
Autonomous investigation of low-risk incidents
-
Proactive threat hunting across the entire infrastructure
-
Automatic reporting for NIS2 and DORA compliance
Key value: Security not as an added cost, but as a value enabler - automation without proper protection is a risk no company can afford to take.
Phase 4: Human resources and cross-functional capabilities (2027)
Our observation: Once finance and security are automated, HR naturally becomes the next area of focus. Polish enterprises are grappling with a talent shortage - automating HR processes frees teams to focus on strategic talent management.
SNOK’s approach: We extend automation to HR processes:
-
75% reduction in time-to-hire
-
70% reduction in payroll processing time
-
50% reduction in HR help-desk costs
-
Automatic preparation for performance reviews
-
Intelligent matching of candidates to roles
Cross-functional orchestration: Teams of agents collaborating across functions - finance, procurement, supply chain, HR - coordinated by an orchestration platform. Example: a procurement agent consults a finance agent about available budget, a supply-chain agent about delivery timelines, and a risk-management agent about supplier reliability - all within seconds.
Phase 5: Continuous optimisation (an ongoing process)
Our observation: Transformation does not end with deployment - it is the beginning of a journey of continuous improvement. Companies that treat AI as a project rather than a programme lose momentum and value.
SNOK’s approach: We implement a continuous-optimisation framework:
Governance and control:
-
Automatic discovery of all deployed agents
-
Clear assignment of accountability
-
Comprehensive performance monitoring
-
Role-based access control
-
Complete audit trails
Employee skills development:
-
A role-tailored training programme
-
Practical workshops with real use cases
-
A best-practice sharing platform
-
Using AI to learn about AI
Measurement and optimisation:
-
Tracking key financial metrics, not vanity metrics
-
Productivity gains, cost reductions, revenue growth
-
Time-to-value, customer satisfaction, employee engagement
-
Quarterly reviews and process optimisation
Our unique position in the Polish market
SNOK combines competencies that are essential to successful AI transformation:
SAP security: 25 years of experience securing the most critical enterprise systems. SecurityBridge as a platform protecting more than 5,000 SAP systems globally.
Cybersecurity: Essential for NIS2 and DORA compliance, particularly given the expanded attack surface created by automation.
UiPath automation: Agentic Automation Fast Track Partner - orchestration of multi-agent systems, which lies at the heart of transformation.
Knowledge of the Polish market: Understanding of regulatory specifics - KSeF, NIS2, DORA - and the cultural aspects of transformation within Polish organisations.
It is no coincidence that these four areas are exactly what Polish enterprises need. We do not offer technology for technology’s sake - we offer strategic business transformation with technology as the tool.
Challenges specific to the Polish market
Skills shortage
85% of companies are ready to increase salaries by 28% for AI talent. Poland needs 1.5 million ICT specialists by 2035.
Solution: Partnerships with universities, participation in the Digital Poland Association, and use of Polish large language models for local AI development.
The investment gap
-
Poland invests ten times less in AI than Germany or France. But funding exists:
-
12.4 billion euros for the Digital Decade
-
149 billion euros in EU funds, of which 21.3% is earmarked for digital transformation
-
2.5 billion euros for cybersecurity in 2025-2026
-
1 billion złoty for an AI development fund
Regulatory uncertainty
Implementation of the EU AI Act is under way, overseen by the Ministry of Digital Affairs. 32% of defence-sector companies cite a lack of clear regulation as a barrier.
Mitigation: Engagement with the work of the AI Advisory Council, monitoring guidance from the AI Chamber, and early adoption of responsible-AI frameworks.
Cybersecurity threats
Poland was the most cyberattacked country globally in 2023 - 80,000 incidents, a 100% year-on-year increase. Fewer than 43% of companies have incident-response plans.
Imperative: A security-first approach from day one, particularly for SAP landscapes; automation of the security operations centre, mandatory for NIS2 and DORA.
Cultural resistance
55% of micro and small companies see AI as “irrelevant”, and 44% of companies are sceptical of innovation. Change management is essential:
-
Clear communication that AI empowers rather than replaces people
-
Broad communication of pilot successes
-
Psychological safety for employees
-
Engagement from senior leadership
Final thoughts: the future belongs to those who act today
2025 is a turning point. McKinsey puts it plainly: the time for exploration is ending, the time for transformation is now. Gartner predicts that by 2028, 15% of everyday business decisions will be made autonomously by agentic AI, and 33% of enterprise software will include agentic capabilities.
For Polish enterprises, three forces are converging:
-
The mandatory KSeF rollout - February 2026
-
Regulatory pressure - NIS2, DORA
-
Rapid AI maturity
Companies that treat this as an integrated transformation opportunity - not separate compliance projects - will build lasting competitive advantage.
The difference between the top 5% of “future-built companies” and the rest is execution excellence:
-
Strategic focus on 3-5 use cases
-
Purchasing specialised solutions
-
70% of effort devoted to people, processes and culture
-
Rigorous tracking of key metrics
-
Sponsorship from senior leadership
Technology is the foundation. Organisational transformation is the differentiator.
The window for strategic positioning is narrow. Gartner estimates organisations have 3-6 months to define an agentic AI strategy before falling behind competitors. Compounding returns for early adopters are creating widening value gaps.
Time is not on the side of the cautious.
The question is no longer whether to adopt AI agents, but how quickly and how strategically it can be done. For enterprises ready to act, the combination of KSeF compliance, SAP security, UiPath automation and comprehensive cybersecurity creates a proven path to an AI-driven enterprise.
The agentic era is not coming - it is already here. Success belongs to those who act decisively, learn continuously and transform comprehensively. The Polish market, with its unique combination of regulatory catalysts and government support, offers a compelling opportunity for a well-executed AI transformation.
At SNOK we believe technology should serve people, not replace them. AI agents, properly implemented with solid governance and human oversight, empower human capability, free people from repetitive tasks, and enable focus on strategic, creative and valuable work.
This is the future of work worth building consciously and responsibly.
We invite you to a conversation about how we can build your AI-driven future together - starting with KSeF, and extending to comprehensive enterprise transformation.