How much of AI are we going to need to call it a productivity plateau?

I continue my previous article on “Optimal-mix-ai-gen-ai-co-pilots-agentic-human” mix. Here comes another transformative piece. I have been tracking it closely around AI PC’s.

This came into perspective towards end of Q3-2024. The push towards Windows 11 migration propelled this change. MS 365 vs. Win 365 strategies are getting debated. The aim is to maximize the value of a Cloud-based, AI/Gen-AI model compatible future state.

Why AI PC’s :

Built on edge or micro-edge computing architecture. It runs on specialized NPUs combining powerful CPUs and GPUs. This setup enhances data ingestion, processing, and inference to make workplace computing smarter. It also caters to protecting data sensitivity, privacy, security, compliance around cloud based PC fleet management model, serving low latency. So what AI workload is it processing to boost your productivity ?

Existing AI family (AI, Gen-AI, Agentic AI, Co-pilot)

The use, adoption and productivities are manifested in 3 distinct areas

• Workplace productivity leveraging product suites like Microsoft Copilots, Github Copilots, Google Workspace-Gemini enabled, Now Assist, Salesforce Einstein, SAP Joule etc. OR custom built LLM based-Chatbot-Virtual assistants
• Business / IT Operations productivity leveraging automating BPO / ITO processes – Agentic AP, Agentic Claims, Agentic Policy Issuance, Agentic Server decongestion, AIOPS, End point self-heal and other Infra heal use cases where we (SI’s and MSP’s) are putting curated custom-built Gen-AI/Agentic-AI based designs to maximize automated operations and right shift human in the loop
• There are many industrialized use cases. These include applied AI in molecular research and drug accuracy determination in clinical trials. Health care BOTs summarize your symptoms and diagnosis with or without medical practitioner intervention. Other examples are insurance premium pricing models, device manufacturing, and IT / OT technologies.

AI PCs the Fourth cross-sectional mark:

AI PC’s are revolutionizing to add productivity the moment you enter your business day. They integrate your Workplace tools and applications, Business operations applications, or Industrialized models. These models are present in your device. They integrate, aggregate, prioritize, and summarize to help you stride your day effectively.

For example, based on your role, it prompts you on the most important business task. This task have been laying around in emails or your notes. It can optimize your calendar and mails based on what has transpired while you were away. It flags-off or auto-heal any device, hardware, or software-related issue. It takes action before you even interacted with a Chat BOT. It tracks what you are working on. It guides you to take notes or communicate. It prompts some knowledge articles to assist. It assists you with prompts and searches on your browser or on your business applications. It can prompt collaboration with a group of users most suitable to help your current work. It can integrating with your end point and security monitoring tools to make the devices more resilient and proactively managed. It can integrate with ITSM. This integration allows it to raise a service request for an issue that it anticipates. It does this without your or chatbot intervention.

Then is there still a problem we need to solve with strategy:

Yes, there are few problems I foresee for which we as CIO partners needs to solve for the CIO’s to drive adoption.

• First, focus on human comfort. Alleviate the fear of extra interference in my day to day work. Keep the minimalistic intervention that “really” improves my own accuracy and mental map.
• Second, measure the end to end ROI of AI adoption. Create a compelling case before the CFO to fund that “extra premium” for AI PC adoption.
• Third, as MSPs and System Integrators; (a) How do we calculate the net ROI and NPV? These calculations involve the multi-year future dated contracts we are signing with our clients. The situation is now more complex with the AI family being topped up with AI PCs. (b) How do we measure productivity? How do we account for committed savings to the CIO and CFO? (c) Which layer of AI-led innovation takes the credit for that productivity? We need to push investment in that layer. We should ignore the rest. After all we have to pick our battles with limited resources that we have at our disposal !

In summary, the advent of technology will complicate adoption and use case definition. This will continue until we are convinced we have reached the productivity plateau. Till then, we need to keep identifying personas of importance. We must also measure interactions and time elapsed for journeys. Additionally, we should leverage models to determine the exact value for AI/tech replacement. This approach should make some economic sense.

Note: Each topic and line item above is worthy of debate and discussions, curated to your context. This is necessary to maximize the value of strategy. Also, I maintain no financial or investment positions in any or all the tools mentioned above. These are just point of view and might differ across specialists, analysts, and expert networks