How does PNC use AI for fraud detection and credit?
PNC's public record on AI in fraud is mostly patents and executive statements, not results. Its published patent applications describe machine-learning models that score unauthorized deposit activity from streaming and batch data (most recently published February 5, 2026), a decision-tree 'Transaction Monitoring Index' for financial-crime alerts (November 2025), gradient-boosted scoring of card payments likely to be reversed (June 2025) and machine-learning risk signals that speed fraud analysis (April 2025). In June 2026 CEO Bill Demchak said PNC is building its own large and small language models for problems such as blocking fraud, and in September president Mark Wiedman named fraud among the bank's five AI focus areas. Gasan Awad, who runs enterprise fraud product management, said in 2025 that AI has become 'readily available, relatively inexpensive' for fraudsters working 'at an industrialized scale', and that the bank's own AI can automate steps and reorder investigation queues while people stay in the loop. On credit, the head of corporate and institutional banking said on March 10, 2026 that every loan credit action is assessed for impacts from AI, with about $5.3 billion of recurring-revenue software exposure under watch. PNC's 10-K says its AI and machine-learning models are increasingly used with customers but 'often less interpretable'. The bank has published no fraud-loss figures, fraud-model counts or results for AI in credit decisions.[1][2][3][4][5][6][7][8][9]
Data as of Oct 31, 2026 · PNC profile updated Oct 13, 2026 · 21 sources cited on this page · 141 banks profiled on this site · Markdown
Which fraud and financial-crime AI has PNC filed, built or announced?
Patent applications are assigned to PNC Financial Services Group per Google Patents; an application is not a granted right and does not prove the model is in production.
| Item | What it does | Status | Date |
|---|---|---|---|
| US 2026/0037979 A1 | Trains and applies machine-learning systems in fraud detection: a model that scores unauthorized activity from streaming and batch data | Application published | February 5, 2026[1] |
| US 2025/0104080 A1 | Machine-learning fraud detection for deposits | Application published | March 27, 2025[10] |
| US 2025/0348880 A1 | Monitors customer transactions for potential financial crimes with a decision-tree model that computes a 'Transaction Monitoring Index' to improve AML and fraud alerting | Application published | November 13, 2025[2] |
| US 2025/0190992 A1 | Scores credit-account payments by likelihood of reversal, using gradient-boosted decision trees to produce fraud prediction scores | Application published | June 12, 2025[3] |
| US 2025/0124444 A1 | Continuous event monitoring that identifies risk signals with machine learning and speeds fraud risk analysis | Application published | April 17, 2025[4] |
| In-house language models | Large and small language models PNC builds itself for banking problems 'like blocking fraud and running call centers', on its own data centers and Nvidia chips | Announced | June 2026[5] |
| Fraud as an AI focus area | One of five areas PNC's AI initiatives concentrate on, with retail operations and commercial loan servicing | Stated | September 2026[6][11] |
| Authentication and cyber | The 2026 proxy credits head of technology Deborah Guild with password-free authentication and remediation timelines 50% faster than peers, alongside generative AI platforms | Disclosed | March 11, 2026[12] |
What has PNC's fraud leadership said about AI?
Awad, senior vice president of enterprise fraud product management, said in March 2025 that fraudsters are 'using it for deepfakes, impersonations and automated attacks', across channels that now include social media apps, and that scammers 'have an ROI in mind' and run segment tests to find institutions' weak links.[13]
In a July 2025 American Banker panel he said AI is now 'readily available, relatively inexpensive for the fraudsters to use', producing attacks at 'an industrialized scale' and removing old red flags such as spelling and grammar errors.[7]
On defense, he described AI automating parts of the work and prioritizing investigation queues (for example, moving the most important of a thousand queued accounts to the top), alongside behavioral analytics, while arguing that some victim conversations 'can't be done through an algorithm'.[7]
He favors data-sharing consortiums, says customers want safe transactions that are not 'totally frictionless', and argues banks must keep running proofs of concept because a new attack vector appears every three to twelve months.[13][7]
How does AI figure in PNC's credit decisions and risk disclosures?
PNC has not described AI underwriting models for consumer or commercial loans; its statements on credit concern AI as a risk to borrowers and to the bank.
| Item | What PNC says | Date |
|---|---|---|
| AI in credit reviews | 'Every single time we have some loan credit action, we assess it not just for the current performance, but we also assess it for impacts from AI', per Mike Lyons, then head of corporate and institutional banking | March 10, 2026[8] |
| Software and data-center exposure | About $5.3 billion of recurring-revenue software lending exposure managed with a focus on AI impacts; about $2 billion of exposure to large data-center companies | March 10, 2026[8] |
| AI-enhanced attacks | The 2024 10-K warns that attempts to breach PNC's data security may become more effective: 'The effectiveness of these efforts may be enhanced using AI' | February 21, 2025[14] |
| Generative AI and deepfakes | The 2025 10-K says cyber attacks are increasingly sophisticated 'including through the use of generative AI and deepfakes', and that humans may sometimes spot anomalous data better than highly automated tools | February 20, 2026[9] |
| Model interpretability | AI and machine-learning models are increasingly used with customers and for process automation, and their results are 'often less interpretable' than traditional statistical models | February 20, 2026[9] |
What has PNC not disclosed about AI in fraud and credit?
PNC does not publish fraud losses prevented, false-positive rates, the number of fraud models in production, or which of its patent applications describe systems it actually runs. The in-house language models for fraud blocking were announced in June 2026 without a launch date.[5][1]
On credit, the bank has made no public statement that it uses AI or machine learning to approve or price loans, and it names no third-party fraud or credit-AI vendors. Its 10-K treats AI models chiefly as a model-risk and cyber-risk matter.[9][8]
What does the PNC profile record on this topic?
The rows below are the entries of the PNC AI strategy profile that bear on this page, as of Oct 13, 2026; the profile holds the full record of 26 dated moves, 14 use cases and 37 sources.
| Use case | What the record shows | Status | Source |
|---|---|---|---|
| In-house fraud-blocking models | One of the named targets for the bank's own language models. | Announced | [5] |
| Microsoft Copilot, firm-wide | Deployed across the organisation as one of several generative-AI platforms; the 2026 proxy credits it to the enterprise AI strategy. | In production | [12] |
| AML anomaly detection | One of the Big Five: large data sets fed to models looking for anomalies; 'we're making progress there' (Reilly). | Rolling out | [15] |
| Patented data migration factory | Built for the BBVA conversion, improved and used for FirstBank; granted US patent 12,321,323 with a family of continuations. | In production | [16][17] |
| AI model-risk disclosure (10-K) | AI/machine-learning models are a minority of PNC's models but increasingly used with customers and for internal automation; the 10-K names their complexity and the laws that govern them. | In production | [9] |
| Date | Move | Detail | Source |
|---|---|---|---|
| Sep 14, 2026 | President Wiedman on agent risk | After a weekend of public alarm over autonomous AI agents, Wiedman says keeping agents from being 'a little bit too aggressive in what they're trying to get done' is a top priority. | [18] |
| Jul 15, 2026 | Q2 2026 call: the patented data factory ran the FirstBank integration | Demchak: the data factory, patented in its first form for the BBVA deal, performed even better inside the FirstBank integration; the continuous-improvement programme is on track for $350 million of 2026 savings; AI is affecting hiring 'at the margin'. | [16] |
| Jul 7, 2026 | New mobile app built with agentic development | PNC's agentic development system and patented data-streaming microservices architecture. | [19][20] |
| Apr 27, 2026 | Annual meeting: 'aggressively building defense as AI rolls out' | Demchak frames AI as productivity and client-service upside and a new cyber risk, on an 'aggressive technology investment schedule'. | [21] |
| Mar 11, 2026 | Proxy: Microsoft Copilot firm-wide, AI-enabled search, two new data-centre sites | The 2026 DEF 14A credits the enterprise AI strategy with multiple generative-AI platforms including Microsoft Copilot and AI-enabled search for customer-facing teams; a new Eastern data-centre facility is under construction and a Midwestern hosting site secured; pre-provision revenue rose to $9.3 billion 'supported by continued technology modernization, including AI-driven automation'. | [12] |
| Mar 10, 2026 | RBC conference: AI disruption folded into credit reviews | Head of corporate and institutional banking Mike Lyons: every loan credit action is assessed for AI impact; about $5.3 billion of recurring-revenue software exposure is reviewed quarterly. | [8] |
| Feb 20, 2026 | 2025 10-K: 'risks resulting from the extensive use of models, some of which use AI' | AI/machine-learning models increasingly used with customers and for process automation; AI law named as a constraint. | [9] |
| Feb 10, 2026 | BofA conference: the 'Big Five', $1.5 billion of AI-addressable spend | CFO Rob Reilly names software, retail operations, AML compliance, the client care centre and commercial loan processing (Midland); the new mobile app was '100%' agentic coding; on vendors, 'everybody, and a lot in-house'. | [15] |
Does PNC use AI to detect fraud?
PNC has filed several machine-learning fraud-detection patent applications, including a deposit-fraud model and a payment-reversal score, and in June 2026 said it is building its own language models for problems such as blocking fraud. It has not published results such as fraud losses prevented.
What did PNC say about AI-powered fraud?
Gasan Awad, who leads enterprise fraud product management, said in 2025 that AI has become readily available and inexpensive for fraudsters, enabling deepfakes, impersonations and automated attacks at an industrialized scale, and that it has removed old warning signs such as spelling errors in scam messages.
Does PNC use AI in lending decisions?
PNC has not said so publicly. Its credit statements concern AI as a risk: the head of corporate and institutional banking said in March 2026 that every loan credit action is assessed for AI impacts, with about $5.3 billion of software exposure watched closely.
What AI risks does PNC disclose?
Its 10-Ks say AI may make attempts to breach its security more effective, that attacks increasingly use generative AI and deepfakes, and that AI and machine-learning models are often less interpretable than traditional models. They also note that humans may sometimes spot anomalous data better than automated tools.
- US 2026/0037979 A1: Systems and methods for training and applying machine learning systems in fraud detection (PNC; published 2026-02-05)
- US 2025/0348880 A1: Computer systems and methods for monitoring customer financial transactions for potential financial crimes (PNC; published 2025-11-13)
- US 2025/0190992 A1: Scoring payments based on likelihood of reversal (PNC; published 2025-06-12)
- US 2025/0124444 A1: Continuous event monitoring, identification of risk signals, and acceleration of fraud risk analysis (PNC; published 2025-04-17)
- Why PNC is building its own 'AI factory'
- PNC Financial Reaffirms Outlook as Loans, Fees and AI Investments Gain Momentum
- GenAI and Multi-Layered Security: How Banks Can Fight the Digital Transformation of Fraud (Leaders video, with transcript)
- PNC Financial at RBC Conference: Strategic Expansion and Growth (transcript)
- PNC Financial Services Group Annual Report on Form 10-K for 2025
- US 2025/0104080 A1: Systems and methods for training and applying machine learning systems in fraud detection (PNC; published 2025-03-27)
- PNC at Barclays conference: growth, AI and branch push (transcript)
- The PNC Financial Services Group, Inc. 2026 Proxy Statement (DEF 14A)
- Behavioral Analytics Inject 'Smart Friction' Into Battle Against Scams and Payments Fraud
- PNC Financial Services Group Annual Report on Form 10-K for 2024
- PNC Financial at Bank of America Conference: Strategic Growth in 2026 (transcript)
- PNC (PNC) Q2 2026 Earnings Call Transcript
- US 12,321,323 B1: Data migration factory (assignee PNC Financial Services Group; filed 2024-02-20)
- BofA and PNC respond to rising AI fears
- PNC's New Mobile App Delivers an Intuitive, Personalized Experience to Help Clients Seamlessly Manage Their Day-to-Day Financial Lives
- PNC Bank rolls out new mobile banking app
- PNC Financial Services Group Shareholders OK 4 Proposals as CEO Touts 'Strongest' Year
When PNC moves on AI, you'll read it here first.
when one of the banks we cover moves on AI, it is in the next morning's brief · six sourced stories · 7 am PT · free
plus every regulation, bank and agent page update, the morning after · leave any morning