AI Pulse Podcast: Tackling the Big Questions in Enterprise AI
by Max Vermeir, Senior Director of AI Strategy
These days, a new AI breakthrough gets unveiled every month or two. The headlines move fast, and the demos faster. But when those same ideas hit real enterprise workflows, more often than not, the momentum comes to a stop.
That tension is why we created the AI Pulse Podcast. We wanted to explore not just what’s new in AI, but what actually works in the real world. Our latest AI Pulse episodes take a look at this from two different perspectives: the journalist watching the enterprise AI industry evolve, and the practitioners building scalable, purpose-built AI solutions.
The journalist’s view: From open source to agentic AI
To understand why enterprise AI feels like it’s moving at warp speed, we turned to tech journalist Adrian Bridgwater, who has spent 25 years tracking how trends rise and collapse. In his words, the last 18 months have been “ridiculous”: RAG one quarter, small language models the next; agentic AI on the horizon.
Adrian’s take is that the AI making headlines isn’t the AI that survives inside real organizations. Rather, enterprises need systems that respect rules and work with existing workflows while keeping costs from ballooning out of control. That means AI has to come with guardrails, integrations, and controls already built in, so teams don’t have to engineer everything from scratch.
“If I had to predict what the industry will be saying or doing by Christmas, it’ll be that there will be some attempt at additional standardization, and for sure there’ll be some productization to move some of the functions toward being more self-service.” — Adrian Bridgewater, Journalist
Watch the full AI Pulse episode:
Learning from the pros: The future is combining LLMs with IDP
If Adrian shows us the bird’s-eye view of where AI is headed, the experts on our latest episode show us what it actually takes to make it work. Collins Sarmento, Director of Product, IDC, and Process Automation at Initium Softworks, and ABBYY’s own Deepak Goyal break down intelligent document processing (IDP) as it actually runs inside enterprises.
Even today, invoices, leases, claims, and mortgage packets still demand hours of manual data entry at many companies. IDP is what replaces that work. It classifies documents, extracts the right fields, validates them, and hands off clean, structured data to downstream systems. That transforms a 24-hour turnaround into a few minutes—with compliance and audit trails built in, to boot.
Why not just use LLMs to do that work? Collins and Deepak point out that LLMs actually need IDP to give them accurate, reliable facts they can work with. In fact, generative AI is accelerating demand for IDP as organizations realize that large-scale automation only works if the data feeding the model is clean.
“We learned pretty quickly that the LM is not the silver bullet.... Where the real magic happens is when Gen AI becomes an intelligent layer on top of a solid IDP foundation.” — Collins Sarmento, Director of Product, IDC, and Process Automation at Initium Softworks
Watch the full AI Pulse episode:
Follow the pulse of real-world AI
At AI Pulse, we don’t just chase trends. Instead, we look at what actually works when AI hits real enterprise workflows. Together with my co-host Dr. Marlene Wolfgruber, ABBYY’s Product Marketing Lead for AI, we dissect how businesses are automating and scaling with AI.
Our recent episodes take on questions many are wrestling with right now:
- How to deploy generative AI safely and securely in production
- What it really takes to bring AI agents into enterprise environments
- Where AI is headed next, from regulation to ethics to sustainability
If you’re serious about real-world AI, you’ll want to join us. Tune in and subscribe to the AI Pulse Podcast.