I have been advising the team at Salesforce for over a decade now, and Dreamforce has always been a blast, but this year it felt more like a turning point, specifically with respect to how AI will get deployed and adopted within the enterprise.? Here are my takeaways:
- The narrative around AI to date has been technology-centric, AI in search of its destiny if you will, the stories swirling around OpenAI and its peers, as well as the hyperscalers at Microsoft, Amazon, and Google.? Customers are encouraged to adopt AI and bring it into their IT stacks on a DIY basis.? This will work for early adopters, I expect, but not for the majority of customers on the other side of the chasm.
- The challenge that is keeping the CIOs I spoke with at Dreamforce up at night is all the data wrangling that must take place to make any material deployment of AI work.? History has taught them once data has been ETL’d (Extracted, Transformed, and Loaded), it is like a decaying swamp—it begins to rot even before you have finished the first loading in and will continue to do so for the rest of time.? We saw this with data warehouses and data marts, and it is hard to see why data lakes would be immune.
- That said, Salesforce, under the leadership of Steve Fisher, introduced a game-changing innovation in data wrangling at DreamForce last year, DataCloud.? It is based on a Zero-Copy principle of data wrangling whereby you federate the schemas to the data and then extract it on an as-desired basis.? No warehouses, no rotting.? Add this to Mulesoft, their API connector to third-party systems of record, and this offers a strong foundation for a viable sustainable data access strategy.
- This year’s big story was the emergence of AI agents as a much-needed extension to the workforce.? This will be a godsend to regulated industries that trap a lot of human capital in low-value workflows that are nonetheless mandatory, be that in banking, insurance, health care, real estate, social services, audit and assurance, or the like.? In parallel, there is an aging out of the workforce challenge facing the energy, manufacturing, and field services sectors, where they are losing the human expertise necessary to keep the lights on.? Even IT is subject to this problem, as anyone who has had to maintain a COBOL program written in the 1980s well knows.? AI is well suited to address both sets of problems either through autonomous agents or co-pilots.
- With respect to co-piloting, an AI-first approach struggles with the challenge of hallucinations, which is a show-stopper for pragmatists wanting to deploy at scale.? But systems of engagement and systems of record are chock full of workflows that have proven reliability.? AI agents integrated into those proven workflows, and surrounded by appropriate guardrails, should be deployable today.
- Finally, in the roundtables I facilitated, one recurrent theme was that top management and the board of directors are anxious to have an AI strategy but unsure of what it should entail.? This makes life very difficult for CIOs, particularly where a material investment of time, talent, and budget is required.? One way to create clarity here is to assess the probability of your enterprise’s operating model being disrupted by AI, and if so, in what timeframe?? The closer in this is, the more risk you need to take, which will likely entail reengineering market-facing processes in your Performance Zone to upgrade your effectiveness.? The further out the risk, on the other hand, the more likely you will focus on reengineering internal processes in the Productivity Zone to improve your efficiency.? The latter feels safer, of course, but only if you are not under imminent threat of disruption.?
That’s what I think.? What do you think?
Educator, Researcher, Entrepreneur, and Consultant Focused on Technology, Innovation & Entrepreneurship.
1 个月Brilliant, thanks for sharing Geoffrey Moore. Your insights should impact the development and implementation of AI driven strategies; it will certainly help my efforts.
Creating Value with Strategy | Strategy Consultant @ Visualise | Lead Coach @ Strategyzer, Leanstack | BSI Balanced Scorecard Professional (BSP) & Senior Associate | Blue Ocean Strategy Certified | Six Sigma Black Belt??
2 个月Accepting ethical data practices could boost trust in AI.?This change will help many companies use AI better.
Thank you for this summary Geoff, how does an enterprise know if their product suite is going to be in the path of the AI product road map? Seems like we have an infinite number of products being built but that most will end up being standard offerings by the big players so the new new thing today becomes a flash in the pan. Any general rule of thumb to follow here? Many thanks again for your insights and wisdom.
eMazzanti Technologies - 4x Microsoft Partner of the Year, CISSP
2 个月It's compelling to see AI evolve towards practical enterprise solutions. Sustainable data strategies are indeed crucial for successful implementation and adoption. Do you think other sectors will follow suit? Geoffrey Moore
Thanks for sharing your takeaways, Geoffrey Moore. It's an exciting time as enterprises prioritize sustainable, efficient data strategies to make AI truly transformative.