An AI Tech Stack
The artificial intelligence technology landscape is complex. The volume of tools and applications can quickly make it feel overwhelming.
Business and IT leaders struggle to align limited budgets with what looks like a sea of unlimited technologies. And unless they demonstrate clear payback from their investments, they put those budgets at risk.
What's needed to escape this complexity is an approach that directly aligns technology with measurable business outcomes and creates a simplified roadmap for technology planning, procurement and payback.
This can be done with a strategically organized portfolio commonly referred to as an AI Tech Stack.
And as our AI consultant staff show on a repeated basis, it can be done in a 3-step process.
An AI Tech Stack is a managed portfolio of business software applications and cloud services designed to create operational synergies, reduce software complexity and future-proof technology investments.
Align Technology with Business Advantage
Begin with an AI strategy that defines the applications to achieve your business goals in the least time, cost and risk.
That's different that the more common but ill-advised approach of acquiring systems and tools in a technology vacuum.
Artificial intelligence is transformational. So, you can align AI apps with business priorities. And there's a technology strategy for this. It's called PACE.
PACE aligns the methods business leaders use to create competitive advantages with technologies that empower and accelerate those advantages.
AI is not a one size fits all solution. So, you need to look at applications in the context of the business value they provide and their unique characteristics. Some of those characteristics are illustrated below.
Each layer in the pyramid achieves different objectives. And is managed differently.
Systems of Record support standard, foundational capabilities. They are important but by themselves generally provide little to no business differentiation. They incur fewer changes and longer life cycles. They receive periodic updates. But releases are iterative and incremental, not wholesale changes or replacement to the underlying application.
Systems of differentiation support the company's different ideas with differentiating technologies. For most companies, different ideas permit the company to improve operations, outperform competitors, or take customers and market share away from competitors. The business benefits are worthwhile but often temporary. In many cases, the benefits are a zero-sum game.
Systems of innovation support new, novel or even experimental ideas. That includes new products or services, new markets and new revenue streams. Where systems of differentiation provide incremental advances, systems of innovation deliver order of magnitude gains. These types of capabilities outflank competitors and realize new competitive advantages. To achieve strategic or transformational change, these tools and capabilities should be allocated more of the limited technology investment.
Different companies may classify the same technology differently. Also, the same tool or capability may change layers due to maturity or obsolescence. Also recognize that Systems of Record are required to support more nimble Systems of Differentiation and Innovation.
Each of the layers have different objectives, maturity, payback, risk and pace of change.
Recognizing these differences simplifies IT management and future proofs technology spend.
Another benefit of the PACE strategy is that it allocates more investment to the capabilities that drive the most business value.
Many people think Gartner created PACE. They didn't. But they made it very popular.
Align the Types of AI
Different types of AI serve different objectives.
The three primary types are shown below.
Predictive AI:
- Makes predictions, recommendations and decisions using statistical models and machine learning (ML) techniques
- Uses historical data and statistical algorithms to forecast (predict) future events or behaviors
- Assigns weighted probability and likelihood to the results
- Applies structured and time series data
Generative AI:
- Uses company content and large language models (LLM) to deliver answers to questions
- Generates new content in the forms of text, image, audio, video and code
- Summarizes multiple pieces of content or complex content
- Translates data into different formats
- Works with unstructured data
Autonomous AI:
- Agents apply natural language processing (NLP) and LLM to engage humans and perform tasks
- They can operate independently, making decisions and performing tasks without human intervention or consent
- They quickly adapt to changing conditions and continuously improve performance without human intervention
The technology becomes especially transformational when these different types are used together to perform tasks that previously could only be performed manually.
Assemble as Desired
Now you can organize the most effective software portfolio.
Categorizing artificial intelligence technologies by their business contribution and type allows you to assemble them for specific business results.
You also have the flexibility to swap components as your business strategy, IT budget or other conditions change.
An illustrative example is shown below.
Technology is Not a Business Result
This approach is unique in that it firmly defines technology that delivers business outcomes. Recognize that most software portfolios define desired technology based on speeds, feeds or technical merits.
But also recognize technology is not a business outcome. It's a tool that when combined with methods or processes achieves results. Without that lineage, you risk adopting software that may be best in class but does not empower your company strategy or business priorities.
Procuring AI software based on technical attributes isn't going to improve business outcomes any more than putting a new engine in your car will make you a better driver.
Final Considerations
An effective AI Tech Stack should do several things.
- It should define a holistic technology portfolio that replaces a mixture of ad hoc and piecemeal applications and tools.
- It should define the technologies to best achieve the most important business outcomes. Things like improving staff productivity, better engaging customers, reducing operating costs and accelerating revenue growth.
- It should lower overall IT cost and future proof technology decisions.
- And it should define the fewest technologies to achieve the most business objectives in the least time.
Without this approach, most companies will acquire departmental and piecemeal systems to solve urgent but isolated problems. That may help with the problem of the day, but quite often contributes to software sprawl, data siloes, lack of integration and temporary results.
In contrast, a strategic portfolio procures technology in a sequence and pursuant to a plan. Rather than buy tools because they solve the problem of the day, they are adopted in a planned order where each app builds upon the prior. It's a strategy that will result in less maintenance, lower total cost of ownership (TCO), higher utilization and much higher ROI.