Innovating at Speed, With Generative AI

Generative AI has the potential to dramatically increase the speed at which companies realize their business objectives to optimize the end-to-end speed and quality of value creation.

Generative AI has the potential to dramatically increase the speed at which companies realize their business objectives. But if a company hasn’t mastered Business Agility – optimizing the end-to-end speed and quality of value creation – they will struggle to get the benefits of generative AI. If they haven’t mastered rapid product discovery, with tight feedback cycles, their AI work will suffer. This blog will explore key elements of Innovating at Speed with generative AI and what operating models and ways of working support success in this time of almost unbelievable change. This blog will cover:

  • Fundamental assumptions:
    • Why are companies interested in generative AI?
    • What are the baseline capabilities needed for innovating at speed?
  • What makes generative AI different?
  • What are the key elements for innovating at speed with AI?
  • What is Eliassen’s internal experience?

For this article, “generative AI” will occasionally be called “AI” for brevity.


Fundamental Assumptions

Let’s establish working assumptions before going into the next level of detail.

  • Companies will primarily use AI to support their existing strategies. It is a means to an end; using AI isn’t a goal.
  • We’re focusing on applications of generative AI based on existing models provided by third parties (e.g., Open AI, Google). That doesn’t mean companies won’t use other forms of AI, but that’s not the focus of this article.
  • Strong foundations of Business Agility are required for innovating at speed with AI. This includes:

    • Flexible funding models, enabled through Lean Portfolio Management.
    • A product vs. project mindset, having intact teams focus on an evolving roadmap vs. creating project teams from scratch for every new idea.
    • Focus on outcomes and goals over features.
    • Rapid identification and testing of assumptions to lower risk.
    • A preference for rapid learning, deploying small bits of functionality, vs large releases.
    • High-performance teams are able to transform intent rapidly and predictably into a working product with limited dependencies on other teams.
    • Excellent technical practices that enable as many deployments per day as needed to achieve business objectives.

If these are missing or weak, speed to value will be impeded.


What Makes Generative AI Different?

AI makes some things faster, possible, or economically feasible. However, there is a lot that makes generative AI challenging. Until you have tried using a Large Language Model to do a particular job, you can’t know if it will do a good job, an average job, or a poor job compared to what you might have wanted. You may have to try things multiple ways. And you may end up abandoning an idea because nothing you try produces a usable result. What this means practically is that there is a lot of uncertainty and experimentation required to find out if your product or service idea for AI is a good one.

How does this differ from programming? The Scrum framework exists because unpredictability is a natural aspect of software development. Scrum addresses this through transparency, inspection, and adaptation, with a general expectation that things will progress as planned within each Sprint. When facing considerable uncertainty, Scrum teams might undertake a 'spike' to gather information and diminish unknowns. However, AI involves a higher degree of uncertainty, resembling a continuous sequence of spikes. When engaging with AI, expect the unexpected and prepare for frequent directional shifts. Due to this increased level of unpredictability, a Kanban approach, with its focus on continuous planning, might better manage the fluid nature of AI projects than Scrum. Now, let’s move on to the key elements for innovating at speed with AI.


The Key Elements for Innovating at Speed with Generative AI

Here are things that will enable your AI strategy to be most effective. Some of these echo themes from business agility but with more detail about why they matter for AI.

  • Adopting an adaptable strategic framework, like Lean Portfolio Management, is crucial due to the swiftly evolving field of AI. Organizations must be poised to quickly evaluate how new AI advancements align with or impact their strategic objectives.
  • Concentrating on goals rather than specific features during planning. For instance, a goal-oriented objective like "reducing the number of support calls to our call center by 10%" empowers product managers and product owners to explore a variety of methods to reach that goal. This is particularly important in the context of AI, where experimentation, quick adaptation, and learning are crucial. In contrast, a feature-focused approach, such as "implement a chatbot for customer queries," prescribes a solution without room for flexibility and may not guarantee the best path to achieving the overarching business goal. It's the adaptability in goal-focused planning that aligns with the dynamic nature of AI, allowing for pivots and adjustments as AI capabilities and applications evolve.
  • Favoring Kanban over Scrum, as mentioned in the prior section to support rapid adjustment to plans as you experiment and learn.
  • Streamlining work with legal, risk, compliance, cybersecurity, and data governance, as well as anyone else involved in reviewing and approving new uses of AI. Organizations with long feedback loops can’t quickly deploy new innovations.
    • Align on what you can do without additional approval so not everything requires review.
    • Develop a process for rapid review when anything comes up that departs from standard guidelines.
  • Build the technical capabilities to enable rapid deployment since the best learning comes from users and customers touching what you built, even if it is in limited release.

Eliassen’s Internal Experience

Eliassen employs an internal team that specializes in applying generative AI. This team efficiently addresses organizational requirements, frequently generating new value within short timeframes. To ensure success, we have implemented several structures facilitating rapid value delivery.

  • The leader of our AI initiatives is also involved in refining work processes, including onboarding, training, and development. This dual role enables swift identification and implementation of innovative ideas.
  • We utilize the Kanban method for agile project management, enabling quick directional adjustments.
  • We have established guidelines outlining the appropriate use of generative AI.
  • We are in the process of integrating data collection mechanisms to refine our approaches based on user feedback.
  • The executive leadership team actively participates and provides support for our AI initiatives.


With these structures in place, we can innovate at speed.


AI offers unique opportunities for companies to achieve their strategic objectives and move into whole new product areas. However, AI also needs some special handling to ensure IT organizations get timely results from their work. By looking at your end-to-end process, and upgrading it where needed, you increase the likelihood of success. And, above all, business agility competencies are essential in this time of dynamic, rapid change.