This month’s AI Pulse decodes the latest in artificial intelligence faster than you can say, “Beam me up, Scotty!” 🚀
Marlene explores Google Colab’s new Data Science Agent, highlighting how natural language is becoming the ultimate interface for data workflows. Meanwhile, Slavena tackles the GenAI hype, channeling her inner Seinfeld to remind us that AI isn’t a magic wand but a tool that needs the right strategy to truly shine. And Jon explores the rise of small language models (SLMs), proving that sometimes, smaller really is better—especially when it comes to cost, privacy, and precision.
Don’t miss our Fun Fact on the OG AI system, SHRDLU, or the jaw-dropping stats in ‘AI by the Numbers’.
But most of all, check out our AI Summits near you, and come meet the experts in real-life!
The Future of Data Science: Talking to AI Instead of Coding?
This month, I came across an exciting update from Google: Colab’s Data Science Agent now generates complete, ready-to-run notebooks from simple natural language prompts. While AI-generated code isn’t new, what’s different is how it seamlessly transforms human intent into structured workflows—handling everything from data imports to analysis steps automatically.
As a computational linguist, I find this shift fascinating. Natural language is becoming the ultimate interface—we’re no longer writing code, but simply describing what we want, and AI handles the rest. This could be a game-changer for researchers and analysts, allowing them to focus on insights instead of tedious setup.
GenAI "promised heaven and earth" for business process automation. Two years in, with expected 30-40% productivity gains still elusive, enterprises face disillusionment and adoption slowdowns. A shame, given GenAI's immense potential to augment humans and accelerate operations.
As Jerry Seinfeld put it: "We're smart enough to invent AI, dumb enough to need it, and still so stupid we're not sure if we did the right thing." In many cases, we didn’t—hype blinded enterprises into over-reliance on GenAI.
The lesson? GenAI isn't a silver bullet. Purpose-built solutions, tailored to specific tasks, deliver a reliable basis for GenAI implementation. Success lies in choosing the right tool for the right job, setting clear goals, and defining ROI from the outset. Enterprises must also address critical concerns like data privacy, model explainability, and safety. Without strong governance and realistic expectations, GenAI risks becoming just another overhyped technology rather than a true driver of transformation.
Faced with rising computing costs and model hallucinations, CIOs are increasingly turning to small language models (SLMs) as a practical alternative for enterprise AI applications. Unlike large language models (LLMs), SLMs require fewer resources, making them a cost-effective solution for businesses aiming to tackle domain-specific challenges. Forrester predicts SLM adoption will increase by over 60% in 2025 as enterprises seek tailored AI solutions with focused expertise.
Notably, 50% of enterprises have explored SLMs within the 1 to 10 billion parameter range in the past year, according to Gartner. Industries like healthcare are already using these models to handle specialized terminology while benefiting from reduced costs and stronger data privacy. Is your enterprise ready to leverage more purpose-built AI solutions? Read on to discover how small language models are streamlining operations while mitigating challenges.
SHRDLU: The AI That First Understood Us (1968-1970)
Between 1968 and 1970, AI researchers at MIT developed SHRDLU, an early system that could understand natural language commands and execute them in a simple virtual world. Users could type instructions like “Pick up the red block and put it on the blue one”, and the system would respond intelligently—an early glimpse into the dream of computers that truly understand us.
Fast forward to today, and we’re living in that reality. From AI agents that generate complex code to systems that process and act on natural language in real-time, the way we interact with technology is fundamentally changing, making AI more intuitive, accessible, and powerful than ever.
In an AI-first world, end-to-end workflow orchestration and automation is critical to delivering target business outcomes and gaining competitive advantage. However, today’s digital infrastructures and processes are both complex and interdependent, making it a significant challenge, especially when AI technology is incorporated into the environment.
🚩 85% of organizations face challenges in scaling and operationalizing AI across their business 🚔 84% say a lack of transparency in applying AI applications within business processes is causing regulatory compliance issues 📈 93% believe AI must be fully integrated into orchestrated processes to maximize the return on investment and business value
Customer story: Erste Digital Improves Document Processing Efficiency by 40%
As part of the Erste Group, a financial services leader in Central Europe, Erste Digital is committed to leveraging AI to automate the bank’s processes, with an emphasis on ease of use for citizen development. ABBYY’s suite of intelligent automation solutions delivers the capabilities they need to achieve their modernization objectives, resulting in improved experiences for both customers and employees.
Blog: Process Mining Vs. Process Discovery: The Wrong Question
In this post, Jon clarifies what process mining and process discovery are, explains why they shouldn’t be seen as separate tools, and reveals the comprehensive solution organizations need to truly optimize their processes.
AI in Practice: Real-World Applications & Case Studies
Case Study: Konica Minolta Wins Its Clients More Time, Money, and Accuracy
The Konica Minolta Business Solutions Spain team recognized that manual document processes were causing productivity losses and a heightened risk of error for their clients. The time required to process invoices, contracts, and other documents was also diverting staff away from more engaging, value-add tasks, impacting employee satisfaction.
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.
