AI outlook

This article is an outlook of how the world could change in the coming 5 years due to AI, as of now, 2025. This is largely my personal viewpoint. Feel invited to debate :)

Case Studies

I want to start by looking at case studies. I think a great starting point for predicting the future is to have a look at what’s happening right now. The LLM boom was just about 3 years ago at the end of 2022 and it has brought a lot of change into the AI world since. Let’s have a look.

Enhancing Productivity

AI, in particular LLMs, can be used to accelerate business processes. A fantastic example is Lumen Technologies, they integrated Microsoft Copilot into their processes, such as sales and customer service. This includes using LLMs, Microsoft Graph, and custom connectors to automate routine tasks - allowing teams to focus on things like direct customer interactions.

Productivity Gains: Sellers reduce research time from 4 hours to 15 minutes per outreach, potentially unlocking $50 million in annual revenue without workforce expansion. Copilot automates note-taking, financial reports, client health scoring, and presentation creation, streamlining workflows in Outlook, Salesforce, and meetings.

Employee Empowerment: Leaders emphasize a culture of empathy, teamwork, and work-life balance. Copilot frees employees for “more human” tasks, fostering joy and resilience. Champions and prompt libraries accelerated adoption, with consulting partner Valorem Reply supporting workshops.

Business Transformation: Supports Lumen’s “North Star” strategy for innovation and customer obsession. Executives like CEO Kate Johnson, CRO Ashley Haynes-Gaspar, CMO Ryan Asdourian, and others highlight how it enables faster scaling, better preparation for meetings, and industry disruption in networking, cloud, and security services.

So the key takeaway here is, they used AI to heavily accelerate or fully automate repetitive business processes (routine operational tasks, administrative workflows etc).

See the case study here.

Microsoft Customer Service

Customer Service is a huge topic as well. Call centers for example are immensely expensive and a smart combination of AI and tools gives really good results that compare to a human taking a call (think of speech-to-text, LLM, RAG, Agent, text-to-speech for example).

Microsoft reportedly saved half a billion by replacing call centers with AI.

See the report here.

Self Driving Cars

Waymo’s deployment in Phoenix shows how autonomous vehicles are becoming reality. They’ve been operating fully autonomous taxi services since 2020, with no human safety driver - just passengers and AI.

Real-World Impact: Waymo’s cars have driven millions of autonomous miles, handling complex scenarios like construction zones, emergency vehicles, and unpredictable human behavior. Their safety record is actually better than human drivers in many categories.

The Tech Stack: It’s a fascinating combination of LiDAR, cameras, radar, and AI that processes this sensor fusion in real-time. The AI doesn’t just see the road - it predicts what other drivers, pedestrians, and cyclists might do next.

Business Model Shift: This isn’t just about replacing drivers - it’s reshaping transportation entirely. Uber and Lyft are already investing heavily because autonomous fleets could reduce ride costs by 60-80%.

The broader implication? We’re not just automating driving - we’re reimagining mobility as a service.

The Startup Signal

Here’s something telling: if you look at Y Combinator’s recent batches, roughly 40% of startups are AI-focused. And within those AI startups, a huge chunk (probably half) are tackling customer service automation specifically.

This isn’t coincidence - it’s the market signaling where the low-hanging fruit is. Customer service is expensive, repetitive, and has clear ROI metrics. Perfect for AI disruption.

When the world’s top startup accelerator sees this pattern, it tells us something important about where the AI transformation is heading first.

IoT Integration

Volkswagen Group shows how AI transforms manufacturing through IoT integration. They’re using AWS to process massive amounts of sensor data from their production lines in real-time.

Smart Manufacturing: AI algorithms analyze data from thousands of IoT sensors across their factories, predicting equipment failures before they happen and optimizing production schedules. This reduces downtime by up to 30% and cuts maintenance costs significantly.

Digital Twin Technology: VW creates digital replicas of their entire production process, using AI to simulate different scenarios and optimize workflows before implementing changes in the real world. Pretty neat stuff.

Quality Control: Computer vision AI systems inspect vehicles on the assembly line, catching defects that human inspectors might miss. The system learns from every inspection, continuously improving accuracy.

The key insight here is that IoT + AI creates a feedback loop where physical processes inform digital models, which then optimize physical operations. It’s not just about collecting data - it’s about acting on it intelligently.

See the case study here.

The Future Landscape

Based on these patterns, here’s how I see the next 5 years unfolding. The following categories aren’t mutually exclusive, but they help structure the thinking:

Automation & Acceleration

AI accelerates workflows and automates repetitive cognitive tasks.

Content generation (text, images, video, code)
Customer service (chatbots, voice agents)
Back-office automation (document handling, compliance checks)
Productivity augmentation (copilots inside Office tools, CRMs, ERPs)

Predictive

AI becomes the forecasting and risk-spotting engine.

Predictive maintenance (factories, infrastructure, aircraft, power grids)
Demand and supply forecasting (retail, logistics, agriculture, energy)
Financial predictions (credit scoring, fraud detection, trading support)
Healthcare predictions (disease risk, treatment outcomes, early diagnostics)

Industry 4.0

Though overlapping with automation and predictive, it’s worth calling out separately.

Smart factories (robots, digital twins, quality control)
Autonomous logistics (warehouses, shipping, supply chains)
Integration of IoT + AI for real-time optimization

Self Driving

Autonomy applied to vehicles and mobility systems.

Road: cars, trucks, delivery robots
Air: drones, air-taxis, autonomous inspection systems
Sea: autonomous cargo ships, port logistics
Infrastructure: traffic optimization, fleet management, multimodal systems

Co-thinker and Planning

AI as a strategic assistant rather than just a tool.

Brainstorming partner, ideation support
Business planning: scenario simulations, strategy modeling, portfolio optimization
Scientific discovery: proposing hypotheses, designing experiments
Complex decision support: policy design, risk trade-off evaluation

Points to think about

Energy Consumption: Training large AI models requires massive computational power. GPT-4 training reportedly cost OpenAI over $100 million in compute alone. As AI scales, so does energy demand - this could become a real constraint.

Security & Privacy: More AI means more data processing, more attack surfaces, and new types of vulnerabilities. Adversarial attacks on AI models are becoming sophisticated, and privacy-preserving AI techniques like federated learning are still early-stage.

The Chip Wars: GPUs (graphics processing units) vs specialized AI chips is getting interesting. GPUs are general-purpose parallel processors originally designed for graphics but perfect for AI training. Companies like NVIDIA dominate here. But specialized AI chips (TPUs, neuromorphic processors) are emerging for specific tasks. The chip shortage taught us how critical this infrastructure is.

Ethical Concerns: AI bias, job displacement, deepfakes, and algorithmic transparency are real issues. The EU AI Act is just the beginning of regulatory frameworks we’ll see emerge.

Market Opportunities: This transformation creates a massive market for digital transformation services. Companies need help not just with technology, but with change management, process redesign, and workforce reskilling.

Skills Gap: The demand for AI talent far exceeds supply. This creates opportunities but also risks - companies rushing to implement AI without proper expertise can make costly mistakes.

The Consulting Angle

All of this creates a perfect storm for AI consulting and digital transformation services. Companies know they need to embrace AI but don’t know where to start. They need help with:

Strategy: Which AI initiatives will actually move the needle for their business?
Implementation: How do you integrate AI into existing workflows without disrupting operations?
Change Management: How do you get teams to adopt new AI-powered tools?
Risk Management: How do you implement AI responsibly while managing security and compliance risks?

The sweet spot is helping companies bridge the gap between AI hype and practical business value. That’s where the real opportunities lie in the next 5 years.

Text Form Summary

Looking at the AI landscape ahead, the transformation (largely) isn’t about replacement - it’s about augmentation and liberation. AI excels at automating the repetitive, data-heavy cognitive tasks that bog us down: processing documents, analyzing patterns, generating first drafts, handling routine customer inquiries.

In financial services, for example, there are great AI models that extract info from financial reports and summarize it for bankers to analyze. This uses the power of generative AI to process a massive amount of text and detect weak points, which is do a degree the core business of banks. Note that Mistral AI offers specifically trained models for this.

This frees humans to focus on what we do best: building meaningful connections, strategic thinking, creative problem-solving, and leveraging emotional intelligence. The magic happens when AI handles the grunt work, allowing people to spend time on relationship-building, innovation, and complex decision-making that requires human judgment.

But here’s the thing - even for strategic work, AI shouldn’t be sidelined. I think leaders should embrace AI as a “thought companion.” Need to explore different scenarios for a business decision? AI can simulate outcomes. Stuck on a complex problem? AI can help you think through it from angles you might miss. It’s like having a brilliant research assistant and brainstorming partner available 24/7.

Customer service is probably the clearest example of this revolution. Instead of humans spending their days answering the same questions over and over, AI handles the routine stuff while humans tackle the complex, emotionally nuanced interactions that actually build customer loyalty.

But it applies to all, from IT, operations, R&D (research & development), finance, to logistics and more. There is also a noticable shift for many positions from hands-on work to more and more reviewing what the AI did - one example is the one of a software engineer.

One can say: AI makes business software intelligent, mainly via process automation and predictive analysis - these two make us use our resources better and make smarter decisions.

The next 5 years will separate the companies that use AI to amplify human potential from those that see it as just another tech tool. The winners will be the ones who figure out this human-AI collaboration dance.