This is such a big topic that it needs a separate category to itself. Yes, I believe that AI is as key a technology as the Internet, and in retrospect we’ll see it as being as transformative as electricity and steam. What will 2025 bring for AI?
- SDLC: We’ll see AI redefining how software is built. Across the range of the SDLC (Software Development Lifecycle), AI will change requirements, coding, testing, release management, and support. Any developer worth his or her salt already uses AI extensively. Google announced this year that 25% of all code released over a quarter was written by AI. Expect this to grow significantly for most companies. Consequently there will be a mini-crisis for software developers. Some will elevate to AI experts, some will be AI competent, some will find ancillary roles and some will leave the profession. I expect to see a much lower intake across the board in almost every tech firm. So the real panic might be in the crop of students currently planning on careers in software development. In fact, I would extend this to much of IT - in how we manage databases, infrastructure, legacy systems, and more.
- Customer Experience: The second big area of AI impact will be on customer experience - this will be across the spectrum - from insights, to loyalty, retention, contact centres (especially contact centres), self serve on websites, omnichannel models. I believe that early mover businesses will start to explore structural changes. For example contact centres and omnichannel teams often work independently from each other but if both are using AI, it’s got to be the same AI, using the same data, and so why different org structures? I believe CCOs (Chief Customer Officers) will need to have an AI office in 2025.
- Knowledge & Strategy: We don’t talk about Knowledge Management anymore. There have been moments over the past couple of decades where knowledge management has been the flavour of the month, usually tied into specific products, collaboration tools, databases, expert systems, and the like. But it should be clear that with AI, the ability to harness organisational knowledge will get a massive boost. Imagine an AI system in a pharma company that can look across it’s history of biochemical research, existing and past drugs, clinical trials, regulations, skills and competences of people, and external data about diseases, and make recommendations on areas for the company to focus on strategically, or to build new competences. I expect AI to become an important strategy and knowledge tool, and I wouldn’t be surprised if you get some AI products aimed specifically at this space.
- Agents hype vs value gap: Expect to see agentic everything. The rebranding and rewiring of everything SAAS as agentic, or AaaS. Some of it will be real. A lot of it will be hype. In 6 months we’ll see a lot of articles about the limitations of agents, where NOT to use agents etc. Agentic AI is an excellent idea by the way. But it will have its own hype cycle. Many agentic programs will see scope creep and/or fail to deliver the ROI for their extensive investment and inflated expectations. At the more useful end of things, we will see examples of AI executing autonomously - including financial transactions with limits, but with real decision capability. Some people have even spoken of agent marketplaces.
- GeoPolitics: The flavour of the month in January of 2025 is clearly DeepSeek. The Chinese Open Source model that has set the cat amongst the pigeons. It seems to perform better than OpenAI models, cost less, is truly open - i.e. publishes all it’s parameters, and most ironically comes from China. My friend
Shefaly Yogendra Ph.D
talks about Trust and Transformative Tech. This is going to be a key battleground between the so called leaders of the free world who are undergoing their own identity reframing. Taking sides will be the challenge of 2025. National AI strategies and policies will be at the forefront.
- Weather: we’ll get better weather prediction and this will have knock on effects on transportation, retail, and other infrastructure, for the smarter providers.
- PhD AI: the capability of AI systems will reach PhD level ability. It still won’t be AGI but in it’s specific areas of training, AI will be able to deliver PhD levels of intelligence and output.
Predictions Part 2 tomorrow.
#innovation #future #prediction
Insightful read, Ved Sen! Your exploration of AI's evolving role by 2025—particularly around ethical frameworks and human-AI collaboration—aligns closely with Ascendo AI's vision. We’re passionate about building tools that empower industries to harness AI responsibly while fostering innovation. Excited to see Part 2!
Helping Enterprises unlock the power of private, secure and trusted GenAI without having to compromise
1 个月Hi Ved Sen very interesting and a not unreasonable set of predictions. Funnily enough we are just deploying an AI solution into the UK MoD that is used for institutional knowledge capture, management and utilisation. We are engaged with other industries in the UK around re-pointing it to their organisations as there is an increasing concern about knowledge loss due to restructuring and "brain drain" as well as unlocking the knowledge that already exists within documents, processes, Comms and people's heads. I like the focus on use case lad predictions. So much of the AI narrative at the moment is on the technology itself and obsessing about one model Vs another
Non-Executive Director | Chair | Keynote Speaker | Senior Board & CEO Adviser | Digital Transformation Specialist | Guiding Leaders to Success in the Digital Age | FTSE 100 Women to Watch
1 个月Thanks for the mention, Ved. I have views, as you can imagine, on "PhD level". Need to define the term a bit sharpish I feel ??
Thanks Ved - picking up on the point about specific AI products for specific use cases, this would follow the "Big Data" trajectory we saw - first generalised "platforms", then localised and targeted solutions (eg retail insoghts). This time around we are targeting arras which have previously seemed to require human intelligence at degree level. I'm particularly interested therefore in the area of AI applications to Facility and Built Asset design - essentially training on the Civils Design literature to be able to generate fully designed and tested 3d models as good as a Grad would have done. Lots of implications for new grads in 3-4 years time plus insurance liability, client readiness etc. I'm interested in picking up this theme with anyone interested. More thoughts on applications in Asset Management as well!
Sales Director at EJADA
1 个月Plain speak. Looking forward to your next set of predictions. Thanks Ved