.svg)
Twelve months on from my article speculating on the future of software in the 2020s and beyond, I’ve decided to take another look. That piece was written pre-launch of my new consulting company, Noremo. So what’s changed? And what endures?
Well very clearly in the software and the wider business world, news media and politicians from Presidents and Prime Ministers downwards, the hype surrounding AI has gone stratospheric. Generative AI, LLMs and Agents have caught the excitement of every entrepreneur, speculator and software developer from Abidjan to Zibo.
Marc Benioff, founder and CEO of Salesforce has said:
“The No. 1 thing I think about every single day is how to explain to my customers the huge value that they can get from Agentforce. That's why I felt it was so important that I'm eating my own dogfood. I've moved my whole Support infrastructure to the Agentic platform; I will move my Sales platform to the Agentic platform; I will move my whole company to the Agentic platform for no other reason but to show my customers what is possible today compared to just 6 months ago; that customers don't understand this and what it can mean for them. That's all I'm focused on. My No. 1 focus is Agentforce.” https://youtu.be/Xgsxi7IGMEU?feature=shared&t=1100
Here’s what Satya Nadella, CEO of Microsoft has said:
“A SaaS application is a business app… The notion that business applications exist - that's probably something that will all collapse in the Agent era. If you think about it, Biz Apps are essentially CRUD databases with a bunch of business logic. The business logic is all going to these Agents. These agents are going to be multi-repo CRUD; they are not going to discriminate what the back-end is; they are going to update multiple databases and all the logic will be in the AI tier. And once the AI tier is where all the logic is then people will start replacing the back-ends.” https://youtu.be/9NtsnzRFJ_o?feature=shared&t=2745
What’s my view? There is huge opportunity, but beware the hype!!! And watch out for the legacy.
Are you selling software? Or investing in companies that are creating software products to sell?
Or are you buying software? And looking for ways to drive operational efficiencies in your business?
Your answer may direct you to take two quite different routes.
If you’re selling software, then the AI hype is good and you want to be in front of the next wave of innovation. You’re probably seeking new customers or looking for ways to upgrade the features of your software products / platform to retain your existing customers’ business. The list of AI leaders all trying to sell you stuff include:
The Pre-Internet red dwarfs (or crocodiles!)
+ IBM, Oracle, SAP, Microsoft
Unlike the dinosaurs, they survived the mass-extinction event at the end of the Cretaceous 66 million years ago - AKA the dot-com boom and bust! They have hundreds of thousands of Tech and non-Tech Enterprise customers, strong cashflows, first class R&D teams, and understand what it takes to leap forward to the next innovation wave, despite their legacy base!
The SaaS main sequence
+ Salesforce, Workday, Adobe, Infor
Regular high margin SaaS-only companies with SaaS business models and competitive moats to protect. They have industry focus, strong cashflows and a throng of captive customers
The Cloud hyperscalers and Social Media blue giants
+ AWS, Microsoft, Google, Meta, X
The leading public cloud providers and social media companies, with strong business models and competitive moats to protect, but also have immense capital reserves to deploy and their own internal challenges
The Generative AI jedis
+ Anthropic, Open AI, Gemini, Grok, Mistral, Perplexity, Others!!!
Young, bright and hungry the Generative AI superstars that are building their own LLMs, consumer brand recognition and drawing down the huge capital resources of their sovereign wealth fund or big tech backers to train their models
The Single-Industry or Single-Domain focused AI protostars
+ No names offered
This space is still evolving too fast. Travel planners, Service Self-Discovery, AI coding assistants and more. TechCrunch remains your goto hype source for news
As a software product platform investor / builder / seller then the AI hyper drive view (that AI is going to transform every business and every job on Planet Earth in the next decade) helps them justify their astronomical levels of capital investment in GPU hardware and data centres to train new generations of models. Or that was until DeepSeek r1 came along and showed that there was a lot more innovation to come that’s going to moderate the slope of the cost curve.
If you’re buying software, then watch out! You’ll be better off in the long-term focusing first on your Tech Stack and Data Modernisation efforts. You can do this before your teams get sucked into that forever cycling AI proof-of-concept vortex and while you wait for the GenAI winners to emerge.
If you are a buyer for a medium or large size enterprise that’s been in existence for 10 years or more, (Noremo’s target consulting clients!) then the likelihood is that you’ll have a whole spaghetti mess of different legacy, on-premise and cloud-first systems. You’ve probably got many duplicate processes, manual workflows, messy inconsistent data and lack a single view of a customer / product / order etc..
Messy data is one of the reasons why LLMs start to hallucinate and in a typical enterprise there is an awful lot of inconsistent, incomplete, duplicate, defective and dirty data to be aware of. So don’t be surprised when you struggle to take your GenAI experiments from flashy prototype to single client pilot to mass production at scale. Getting to production at scale takes a whole lot longer and is a whole lot more complicated than simply creating a handful of cool-looking proof of concepts and single-scenario live pilots; especially if you are working in a regulated industry and have a duty of care to your customers.
By all means it is going to be very valuable to experiment with new GenAI tools around the edges of your business; eg. to help your sales / marketing / service / finance / product development teams to be more efficient in their current ways of working. So do continue with that.
But GenAI is not going to do the really hard work for you and help you re-define, streamline and automate your core processes - where duplicate processes, absent or inconsistent data and manual workflow steps that require multiple humans to deal with the uncertainties of information and the fudging of tasks exist. You’re going to need to do the hard graft yourselves - and perhaps with the support of a small team of external experts who’ve worked on similar challenges before.
The greatest near-term value you are going to create is by simplifying and streamlining your processes, automating those manual workflow steps and interactions, mastering your data structures, cleansing your old data, creating new sources of data, retiring your legacy systems and reducing your technical debt, so that in 2-5 years time; as the GenAI winners emerge you, your processes and your data will be in a way better place to take maximum advantage!
If you can do the hard and dirty work first, then you’ll be in a way better place to drive process automation and accelerate how your business performs work at high velocity and at scale. And with the same number of employees, grow your revenues way faster than your costs.
Historically, technology change happens more slowly than the initial hype-stars suggested, but then the long-term impact is often way bigger than anyone could have imagined. So do stick with it. Your business’s economic survival depends on it.
Does that change what I was thinking 12 months ago? Not really. The Acceleration is here to stay.
For those of you that are interested and want an excuse to visit the pulsating city of Berlin Germany, in June, I’ll be speaking at the SuperReturn Technology Value Creation event on Tuesday, 3rd June 2025. See below link for the full agenda.
https://informaconnect.com/superreturn-international/technology-value-creation-summit/
My original article
https://www.noremo.uk/news/what-makes-software-fit-for-the-2020s-and-beyond
Photo credit: NASA Hubble Space Telescope
https://unsplash.com/@hubblespacetelescope