Xingchu Liu, Ph.D.’s Post

View profile for Xingchu Liu, Ph.D.

Chief Analytics & AI Officer at Pfizer | Advisory and Board Member

Super bullish view of AI agents by Satya Nadella recently at a BG2 podcast with Bill Gurley and Brad Gerstner, that SaaS business applications will be soon gone away and replaced by AI-agent-powered solutions. Quite interesting position given Microsoft's giant share of the SaaS market. I do like Satya's simplified and abstract view of SaaS applications as a combination of databases and business logic. At the same time, I also see 3 major challenges (some of them were called out as well) 1. Engineering - common connectors to allow AI agents to access and take actions without human intervention. Today a user can open and operate across multiple SaaS applications by different vendors once licenses are purchased. However, it's a totally different story for AI agents, which will take an industry-wide effort for enablement; 2. AI - although the recent GenAI advancement is super exciting, my conservative view of business AI intelligence is still quite limitedly impacted by these advancement, especially with domain expertise and accuracy. Considerable work is still needed to make AI agent smart enough. 3. People - the business adoption and integration is always non-trivial. Even till today, there's still a lot work to be done on the SaaS side. AI-agent integration will require the next level. Also, automation comes with higher risks and lower transparency. Both need robust process/solution in place and culture shift. I am actually quite excited about AI agents, and do believe this is coming strong. I also think we need to solve all 3 challenges together to be truly successful, that'll take dedication and commitment to see fruitful results. What do you think? (This is the original interview of a 90mins video. There're shortened versions that can also be found on Google.) https://lnkd.in/gx_tr8A6 #ai #SaaS #AIagents #GenAI #analytics #machinelearning #data #automation #businessapplication #CAO #CDO #CDAO #transformation #Microsoft

Satya Nadella | BG2 w/ Bill Gurley & Brad Gerstner

https://www.youtube.com/

Liran Eisenberg

VP Engineering | Head of Data | Transformational Leadership | Mentorship | Technology Roadmaps | Cloud Computing | VP analytics

2mo

When you consider Microsoft's SASS business model, these AI agents would increase their market share. Imagine a world where you can tell your analytics solution which metrics to display in your dashboard. This would eliminate the major IT headaches associated with analytics and empower the end user. I can also envision a world where the various cloud providers enable these AI agents to interact, similar to how they collaborated on data sharing of cloud buckets. In the realm of SASS applications, if SAP and Salesforce ever teamed up, that would be a game changer. However, history tells a different story :) Interesting interview, thanks for sharing!

Buddhika Madduma

CEO @ LayerNext | Strategic Intelligence for CFO office

1mo

Xingchu Liu, Ph.D.I would add another point to your observation: data. Most enterprise data systems are not ready to work with AI agents or agentic workflows. What do I mean by data systems not being AI-ready? A lack of metadata, improper data models in warehouses, disconnected data silos, and lineage issues. A lot of work needs to be done to address these data layer challenges. However, I believe we will see AI agents specifically built to handle these issues.

See more comments

To view or add a comment, sign in

Explore topics