Intent-Based Skilling

The concept of skill-based organization is gaining traction. While it’s hailed as a superior way to measure, manage, and forecast organizational workforce needs and capabilities, current models often feel broad and ambiguous. Many lack a clear roadmap to transition to this skill-centric future, both in terms of strategic planning and execution. That is, I’m finding a lack of pathway supporting the switch to that ideal skill-based future, including the « WHAT » for associated strategy and planning; and the « HOW » on execution mechanisms for smooth roll out. At present, the predominant solution seems to be establishing an organization-wide skill taxonomy, but this often complicates the situation rather than resolving it.

 

In my view, transitioning to a skill-based organization necessitates forming new, adaptive connections and interactions between the talent data worlds (data about talent, whether captured through internal processes, or external data about the talent market), and the physical world of talents (what is happening within the organization and its business ecosystem). The challenge now is not only about the type and immediacy of these connections but also their quality and reliability. Yet such a new model is complex. Given the numerous parameters to evaluate and adjust skills, managing a skill-based organization would be untenable without an intelligent system offering prescriptive capabilities.

 

My vision for a skill-based organization is one where Talent (with a capital T) is a fully adaptable and flexible resource. This would mean overcoming all business scalability challenges, from basic talent access to precise visibility over existing Talent. It would also mean constantly anticipating and resolving low adaptability of talent resources due -for example- to talent-induced costs or low transformation velocity. In a mature skill-based setup, talent management would primarily revolve around skill management, with many talent-related decisions being autonomously computed and executed.

  

Aspiring skills-based organizations should embrace the concept of “Intent-Based Skilling” (IBS)

 

In networking, as users, devices, and applications have exploded, the environment has become highly complex. The concept of “Intent-based networking” (IBN) is a solution to keep modern networks running smoothly. In essence, IBN is an evolution of SDN (Software-Defined Network) and it transforms a manual network into a controller-led network. The goal is to continuously monitor and adjust network performance to obtain desired business outcomes. Put very simply, business intent is imputed into the IBN, then the system outputs network policies that can be applied consistently.

 

I find the principle of IBN very much aligned to today’s skill-based implementation challenges, and advocate for what I am calling “Intent-Based Skilling”, or IBS. To simply detail the concept and its benefits, but also to avoid tech jargon and better bridge the concept to skills-based organizations, I’ll use “Intent-Based Casting” as an analogy.

 

 

What is intent-based casting (in a theater)?

 

Imagine a large theater company planning multiple productions, each requiring a mix of actors, singers, dancers, and other talents. As the number of productions and performers grows, optimizing how talents and roles matches becomes challenging. “Intent-based casting” goal would be to disrupt the historical audition process. Instead of a long series of casting with prospective artists, the director defines the “intent” (e.g., “I need a soprano who can also juggle with machetes”). That’s when an intelligent system captures that intent and ideally finds the perfect match of artist(s) based on available talent. The aim is to ensure that every performance is what has been ideally envisioned by the director.

 

Historically, the casting process was manual and based on the subjective judgement of the director or equivalent casting director and staff. They’d rely heavily on people’s memory, previous experiences, and physical portfolios when considering artists for roles. The casting process was naturally experience-dependent, time-consuming, inconsistent because of human error/bias/lack of knowledge, and of course a slow-reactive mechanism hardly capable of coping with sudden performers’ drop out which could compromise the quality of the overall performance.

 

In contrast, Intent-Based Casting is proactive, data-driven, and consistent:

 

• Data-Driven - A comprehensive database provides real-time insights into every performer’s strength, weaknesses, and unique skills, ensuring that no talent goes unnoticed.
• Efficiency - Since the system already knows the abilities of each performer, it can suggest optimal matches without repeated auditions. It’s a quicker, more streamlined process.
• Consistency - By automating the matching process, the system eliminates human biases and errors, ensuring a consistent quality in casting.
• Proactivity - If changes are needed, the system can immediately suggest alternatives based on the intent and required skills, ensuring the show goes on without a hitch.

 

 

The Three Phases of “Intent-Based Casting” (and Intent-Based Skilling)

 

The IBC evolves from traditional casting calls. It uses a central database, a “casting controller”, if you will, which has a detailed profile of every artist within the company, and similarly detailed information about external artists. This database knows the strengths, weaknesses, past performances, and unique talents of every individual. Whether the production is a Greek tragedy, a broadway-like musical, or a modern dance performance, the system ensures the best artists is placed where they shine the brightest.

 

Like Intent-Based Network, the system runs on three phases:

1. Translation - Understand requirements and convert them into a specific set of skills with appropriate proficiency and context applicability. This is the generic principle; I’ll add to it later in this post.
2. Activation - Match the required skills combinations with available artists (internal, and ideally also external) to make sure roles will be filled by those best suited to them.
3. Assurance – This is a closed-loop system. It monitors rehearsals, live performances, and any other relevant acting moments to collect performance feedback. An ideal system would also make possible some form of simulation. It’ll serve as a feedback loop to measure how the actual acting meets the director’s vision. If adjustments are needed, such as swapping roles or getting specialized training for a performer, the system facilitates those changes.

 

In this analogy, execution flaws may become very apparent despite a sound design. Specifically, they relate to how the initial intent is captured, and the integrity of any feedback loop. Because these hold similarities with what the Intent-Based Skilling model is, it is important to address them within the analogy. First, capturing the Casting Director’s (and Business) vision as the leading intent. This is likely where most of the upfront challenges for early adopters of the concept would surface. The process for capturing the director (or any business leader’s) intent must be intentional and rigorous. I recommend opting for a granular programmatic approach with four key elements of that phase.

 

• Detailed Input - The system would allow for comprehensive input, where the casting director can describe not just hard skills (like singing or dancing abilities), but also softer aspects like emotional range, chemistry with other actors, or specific looks. This can be done through a combination of dropdown menus, sliders, and free-text inputs.
• Feedback Loop - After each production, the casting director provides feedback on how well the performer fit the role. This helps the system refine its future recommendations. For example, if a performer was rated highly for comedic timing in a play, the system will give them higher priority for comedic roles in the future.
• Dynamic Updates - The system would be continuously updated. As performers gain new skills or have standout performances, their profiles are updated, ensuring the system’s recommendations are always based on the latest information.
• Human-AI Collaboration - While the system can suggest potential matches, the final decision always rests with the casting director. They can review the system’s recommendations, look at performer profiles, and even override the suggestions if they have a particular vision or gut feeling.

 

Second, capturing the elements properly to inject in the feedback loop. The point is generally about making sure of a healthy feedback review mechanism to support the integrity of the system. For the intent-based casting, examples could be to use Audience and Peer Feedback. Indeed, solely relying on the director’s feedback could introduce many challenges such as reintroduction of bias, impossibility to scale (having multiple concurrent shows running), and more. That’s why aggregating feedback from multiple sources is paramount, especially as the show grows in complexity and requires an intricate mix of artists and unique skills. Another mechanism could be Automated Collection of Performance Metrics. That’s another vehicle assuring the system relevance and precision over time is to integrate modules capable of automated data collection. For example, by incorporating technologies that perform automated sampled feedback collection (i.e., collect only random sample over multiple versions of multiple shows), or automated feedback weighting to further avoid bias reintroduction.

 

 

Key Benefits, Features, and as Many Drivers for Considerations

 

Bottom line, driving reasons for the adoption of the IBS concept and the application of its model lay in the key features for that new Intent-Based archetype. They logically resemble the core capabilities of Intent-Based Networking, but with a distinct talent twist. I would resume those in four key features:

 

• Scalability - Whether it’s a small play with a handful of characters or a massive musical production, the system scales to handle any number of roles and performers.
• Customizability - A casting director may adjust the weightage of certain criteria. For example, if for a particular play, they believe that chemistry between actors is more important than individual skill, they can adjust the parameters accordingly.
• Adaptive Learning - As more productions are done and more feedback is provided, the system learns and gets better at predicting which performers will best fit certain roles, aligning more closely with the casting director’s vision over time.
• Time Agnostic (Fast forward/backward) — In addition to adaptive learning capabilities, an ideal system would feed and update in real-time with a variety of parameters precisely characterizing artists and considering their skills plus their proficiency based on applicability in context. The point is to maximize output quality in a highly sensitive and dynamic environment.

 

 

 

 

Disclaimer: A LLM has been used by the author to correct minor grammar and syntax errors, but also to help with readability. However, all and any ideas discussed and developed are the author’s own brain juice.