Welcome to AI Pulse! This month’s edition is less about scale and more about smarts. From domain-specific models reshaping industries to agentic AI reinventing financial crime prevention, the focus is shifting to outcomes, trust, and precision. Gartner predicts that by 2027, task-specific AI will outpace general-purpose LLMs, driven by the need for accuracy and cost efficiency. Meanwhile, agentic AI is proving that true transformation comes from designing AI to act autonomously within a fully re-engineered business process, not just assist in the current (broken) one.
But with great power comes great responsibility. As businesses double down on AI governance, risk mitigation strategies like human-in-the-loop oversight and data access restrictions are becoming non-negotiable.
Last but not least, don’t miss the “ABBYY Ascend 2025.2 – Clarity in Motion” recap! It’s packed with insights on how AI is maturing to meet real-world business challenges. Read on, Sir or Madam, and discover what you need to know about the latest in AI.
Beyond the AI Bubble: How AI Grows Up and Gets to Work
The next leap in AI isn’t about building bigger models — it’s about building smarter ones. Domain-specific and process-aware AI must truly understand how business runs: documents, data, decisions. Instead of chasing scale, outcomes count. It needs models that learn the language of industries and the logic of processes.
Gartner now forecasts that by 2027, organizations will use task-specific models three times more than general-purpose LLMs, driven by the need for accuracy, cost efficiency, and domain alignment. Because in business, trust and precision aren’t optional — they are the foundation of impact.
Agentic AI: The Real Transformation in Fighting Financial Crime
While banks have seen modest gains from analytics and GenAI (think 15–20% productivity boosts) they’ve mostly applied AI to support, not transform, human workflows. Agentic AI changes that. When banks design from the process up and involve frontline analysts, agentic AI can deliver 20x gains by taking on whole tasks autonomously—not just assisting investigators, but acting for them.
Success hinges on three factors: a clear, modular view of the KYC or AML process, an AI-first data remediation strategy, and architecture that spans both compliance and business operations. This isn’t automation. It’s reinvention.
Risk Mitigation Budgets Swell as Enterprise AI Adoption Grows
For all the value AI offers, companies can’t afford to ignore the risks. Nearly all (98%) of enterprises plan to increase governance budgets in the next financial year, with the average business anticipating a 24% jump. And more than 4 in 5 businesses cite AI risks as a driving force behind modernizing governance practices.
Businesses are already leaning on several mechanisms to improve risk mitigation, such as human-in-the-loop oversight, data access restrictions and AI access via trusted technology providers. Getting governance right will remain an iterative process as AI technologies mature and businesses analyze potential dangers.
Back in the '80s, the Caribbean island of Anguilla was randomly assigned the country code .ai—long before “artificial intelligence” meant anything to most of us. Fast forward to today: it's a digital goldmine. Revenue hit $40M in 2023, with over 850,000 .ai domains now live (up from just 50,000 in 2020). As AI takes over the world, Anguilla’s domain windfall is funding roads, schools, and a front-row seat to the tech boom—by sheer lucky coincidence.
🤖41% of employees report having received AI-generated workslop in the last month, and they estimate that an average of 15% of the content they receive at work qualifies as "content that appears polished but lacks real substance, offloading cognitive labor onto coworkers."
💰It carries a real cost for companies. Employees reported spending an average of almost 2 hours dealing with each instance of workslop, which translates into over $9 million per year in lost productivity for an enterprise-sized company, according to Stanford Social Media Lab and BetterUp.
Myths About Agentic AI that CIOs Should Ignore As you move forward with agentic AI, ignore the rumor mill. For CIOs, that starts by debunking four AI myths that are holding businesses back. Read the full article in Information Week by Maxime Vermeir.
Generative AI Confessions survey highlights key obstacles businesses face when implementing generative AI. Nearly a third of UK leaders cited lack of staff skills, 30% found training models harder than expected, and 25% struggled with integration. Misuse of tools was also reported by 20%.
FinTrU Streamlines Regulatory Workflows with ABBYY Document AI
Financial Services firm FinTrU developed a document intelligence solution using ABBYY Document AI to automate data extraction from large, complex regulatory documents—and achieved a 99% first-time pass rate in compliance review.
Transforming Know Your Customer (KYC) Workflows with AI Agents
IBM and ABBYY enable financial institutions and insurance companies to process large volumes of documents quickly and reliably, while ensuring complete traceability, integrated regulatory compliance, and optimal audit readiness.
The Innovation Hub is more than just a gallery of projects; it is a collaborative ecosystem built for our developer community. Read the blog to see what you can expect and how you can contribute your ideas.
Guest post: Process-First Automation with Document AI and GenAI
Without context, extracted data is just numbers and text. With process understanding, it becomes actionable intelligence that drives real outcomes. Explore process-first automation with Document AI and Gen AI, using customer inquiry management as an example.
IDC Spotlight - Enterprise Automation Needs More than Just LLMs to Unlock Business Value
As LLMs transform how we handle documents and unstructured data, companies must think strategically about how to use this new wave of intelligent automation. This report helps guide organizations on how to do just that—starting with a foundation of accurate data.
Nominations are now open for the 2025 ABBYY Partner Awards!
If you are an ABBYY Partner, here’s something exciting to recognize your extraordinary achievements in innovation, customer impact, and collaborative expertise. Nominations are considered across 9 categories – entries are accepted until December 31, 2025.
Join ABBYY and iSecure on November 18 in Zurich for the AI Summit Financial Services – an exclusive half-day roundtable for banking executives. Together we’ll explore how AI and intelligent automation are transforming compliance, KYC, fraud detection, and data residency in Switzerland.
Did you miss Ascend 2025.2 or want to replay any portion of it? Check out the replay here. See how we’re turning complexity into clarity and clarity into action in the world of GenAI and agentic automation.
Webinar: LLMs in Document Processing: Where They Fit, Where They Don’t
November 13 - This session gives you a clear decision framework to choose between LLMs and purpose-built IDP, and to combine them when it pays off. You will see patterns and anti-patterns, ROI breakpoints, and real examples where LLMs add value
Missed ABBYY’s AI Summit Singapore? You can now watch it on demand. Hear from ABBYY leaders and industry experts as they explore how organisations are using Process and Document AI to drive smarter and faster business decisions.
World Financial Innovation Series (WFIS) Indonesia – November 25–26
Join ABBYY at the WFIS Indonesia in Jakarta. We’ll be engaging with leaders across banking, financial services, and fintech to showcase how ABBYY Process and Document AI are transforming customer experience, strengthening compliance, and driving efficiency across the financial services ecosystem.
You have received this e-mail because you are an ABBYY customer or partner, have requested further information on our products, solutions or services, registered for an ABBYY trial version, and/or agreed to receive news from ABBYY.
