Workforce Capability Intelligence is the future beyond talent intelligence. It uses science, AI and the same methods as intelligence agencies. With this, it will support high-skill roles, and provide the framework for a complete transformation of workforce management.
Talent intelligence vision looks at skills, and how to improve hiring, upskilling, workforce planning and internal mobility. Workforce Capability Intelligence goes further, to build a picture of what people currently do, and what they potentially could do. It uses the full spectrum of skills, experience breadth and depth, performance, behaviour and cultural fit.
Impossibly ambitious? Not now, with AI support and using the techniques developed by intelligence agencies to support state security, policing and environmental management.
The future is Workforce Capability Management. We’ll be able to manage staff with the same scientific efficiency we use for managing finances, product and processes. And we’ll be able to give people richer working lives.
The author is a co-founder of the Metatalent.ai talent intelligence platform. The article below are is personal views about the future of what organisations will ultimately achieve when they move to Workforce Capability Intelligence. The Metatalent.ai platform and services provide a major step towards that. https://Metatalent.ai/
Problems with Talent Intelligence
To understand why Workforce Capability Intelligence is needed, it helps to look first at the much older discipline of talent intelligence.
Talent intelligence is the discipline of collecting data about employee’s skills, and analysing this data. It’s history goes back decades, but it’s now going through a transformation as a result of AI and big data.
For common roles
Talent intelligence is used for common roles where there’s a lot known about the skills for the role.
It provides knowledge about the skills needed for a role. It helps improve management decisions, and improves activities like recruitment, succession planning and reorganisation. And it can help bring the right training to employees to help them progress their careers.
However it’s got challenges, as set out in “Talent Intelligence: Transforming HR decisions with skills-based data” https://www.techwolf.ai/resources/blog/talent-intelligence-transforming-hr-decisions-with-skills-based-data .
- Poor-quality data in conventional HR systems. There’s little of it, and it’s typically based on inconsistent and non-standard phraseology. Companies need to move to new ways of classifying skills, training courses and roles.
- Working practices can evolve rapidly, especially in high-skill areas. Your lists of skills, courses and roles need to keep evolving.
- Organisational resistance. Adopting talent intelligence involves changes of attitude and behaviour, as well as transformation of processes, technology, contractual issues, decision-making, and the management of training suppliers.
At Metatalent.ai, we help clients handle all this, in the platform and related professional services. But that’s just a start.
False assumptions for high-skill roles
The reason for aspiring towards Workforce Capability Intelligence is that talent intelligence only supports common roles. It’s based on 4 false assumptions which cause problems when it’s used for high-skilled jobs.
1. We know all the skills needed by experts.
High-skilled jobs involve a range of soft/power skills, the individuals often fill more than one role, and the skills are evolving because of AI and other changes. To recruit people or develop them into new roles, we need a detailed understanding of everything people are doing for their job, and how they bring experience from different areas. We don’t have this, at the moment.
2. People can be assesses on limited information.
We make judgements about new recruits based on CVs, tests, and interviews … and we regularly get it wrong. It’s from a lack of information, analysis and related structure. We need vastly more information – that’s potentially possible for internal appointments.
3. Problem handling can be ignored.
In high skill roles, once people have gone beyond a basic level, problem handling can be more important than repetition of a safe formula. It relates to deeply understanding skills, and what works and why. However the depth of experience doesn’t show in simple classification of skills and years of experience.
4. Cross-domain experience doesn’t count.
High skill roles attract people with a breadth of experience, taken from loosely comparable skills. But breadth, doesn’t show in a simple classification of skills.
& also …
This is still just looking at skills and related experience. It doesn’t cover performance, behaviour and cultural fit. And there’s also the matter of individual’s motivation.
We must go much further. We need an approach that it neutral and trustworthy.
Workforce Capability Intelligence – FAQs
What is Workforce Capability Intelligence?
Workforce Capability Intelligence addresses what people are capable of achieving, if they have the intent.
It includes skills, experience (breadth and depth), performance, behaviour and cultural fit.
Ultimately, it’s potential is constrained only by legal and ethical concerns.
How does state intelligence fit with Workforce Capability Intelligence?
Workforce Capability Intelligence in industry is about hope and people’s potential. A large part of state intelligence involves the opposite: threats of “bad” events by “rogue actors” – dangerous individuals or groups.
By focussing on hope and potential, we’ve flipped from obstructing bad actors to helping good actors. The methods are very similar.
(N.B. In state intelligence, they talk in terms of risks and threats. Paraphrased: a risk is a specific bad event that may happen, and a threat refers to a series of unknown events from rogue actors. Threats are assessed in terms of capability of the rogue actor, and they’re intent. That’s where the word “capability” comes from, in this context.)
What are the benefits of using Workforce Capability Intelligence?
It can be used for high-skill complex roles, as well as common ones.
It offers:
– a higher success rate in candidate-matching,
– improved efficiency in reorganisation,
– easier adaption to using disruptive technologies like AI,
– greater preparation and resilience to major disasters, and
– greater insights into the balance of your teams.
Used wisely, it can help with staff retention and career development.
Can Workforce Capability Intelligence be automated?
Partially. Data collection, most of the processing, some analysis functions, and some reporting can be automated.
But there are key decisions to make at the start, quality assurance and management oversight throughout. Interactions between humans (such as in interviews) are critical, as are human judgements. And humans can recognise outliers and patterns that AI misses.
Key to the human judgement, is the quality of the intelligence presented to them. Trust is the foundation of workforce capability intelligence, as covered below.
(Could there be a single technology providing a full WCI solution, including workforce management? That’s a separate discussion, not covered here.)
Key principles
There are 5 key principles that underly intelligence processes, whether for state security, policing, environmental protection or for workforce capability.
- Establish the objective
- Be scientific – intelligence answers questions
- Employ multiple intelligence sources
- Trust – focus on reliability
- Iteration – if it’s not good enough, go back and keep trying
Intelligence objectives
Look at people and what they bring to their work to achieve results.
Within that there are different operational objectives. As in the examples below, they only partially overlap:
– Identify internal candidates for an upcoming role (succession or new appointment).
– Assess what is needed to develop someone into a new role.
– Understand the skills used by key experts, to support reorganisation or introduction of disruptive technologies (like AI).
– Look at how to improve the efficiency of a team.
– Find ways to reduce staff churn.
– In a merger, understand the capability differences between the acquired people and the existing organisation.
With each objective, different kinds of intelligence, research and analysis are needed.
And for each objective, there needs to be a clear view of the business case for intelligence. That’s both for the immediate challenges, and for the recurring ones.
(Academic note: In conventional intelligence literature, objectives are sometimes referred to as “framing questions”.)
Be scientific – intelligence answers questions
Intelligence is not collected randomly in the off chance it will be useful. That would be too expensive, as well as being unethical.
Intelligence is collected because it potentially helps answer a major question that relates to one or more of the intelligence objectives. So each objective may have a series of questions.
For example, in a challenge to find an internal successor for a role, questions might include these (and more):
– Could the candidate have the necessary technical skills within 4 months?
– What are the main weaknesses in soft skills of the candidate?
– How well do they understand the business/operational context?
– What outlier characteristics does the candidate exhibit? (Outliers can be good.)
– What’s their capability at handling unknown situations?
– How well do they fit culturally with the team?
– How motivated are they?
As a science, these are similar to testable hypotheses. As part of building confidence in the conclusions, it’s important to test the null hypothesis: that the question can’t be answered, or is meaningless. For example, the answer to the question “Could the candidate have the necessary technical skills within 4 months?” might be “Can’t tell”.
And as a science, important questions need to be tested using completely different methods, to see if they give the same answer. That’s one of the reasons for having multiple intelligence sources, as below.
Multiple intelligence sources
Intelligence answers questions. Different sources answer different questions.
Below are some examples of intelligence sources, for Workforce Capability Intelligence:
- Public intelligence, from open sources, curated by the worker
- CVs.
- Blogs.
- Social media, including LinkedIn.
- Interactive sources
- Questionnaires and narratives about specific aspects of an employee’s skills or their career aspirations. (Free text narratives can be interpreted using Generative AI.)
- Psychometric assessments.
- Task-related games.
- Observations monitored during structured training
- In online training, the learner’s strengths and weaknesses provide feedback, and can highlight other characteristics (such as analytic thinking and creativity).
- Observation at work
- Use AI to scan the key documents that employees write at work, and assess competence in specific skills.
- Likewise, formal emails can highlight competencies and weaknesses.
- And transcriptions of dialogue in meetings, can also indicate competencies.
When planning which intelligence sources to use, take care with key questions that need to be tested with independent methods. Each test needs to rely on different sources. A technique for planning this (with the support of an intelligence analyst) is a series of matrixes, one for each objective. The matrix lists the evidence to be collected against the question.
Below, an example of part of an intelligence collection plan, in the context of identifying internal candidates, and where alternative tests are required for each question.
Objective) Identify internal candidates for an upcoming role | |||
Questions | Necessary tech skills? | Necessary soft skills? | Business understanding? |
Intel source | |||
CV review | AI assessment. Human cross-check. | AI assessment. Human cross-check. | |
Questionnaire | Psychometric assessment. | ||
Trial | Test in simulation. | Test in simulation. | |
Structured interview | Questions from tech expert. | Questions from business expert. | |
etc |
(Academic note: This is a structured augmentation. In police intelligence, the columns would be hypotheses and the rows would be evidence.)
Trust – focus on reliability
The decision makers need to know what they can trust.
Getting a major decision wrong can be disastrous. In state security intelligence, lives can be lost and damage can be considerable. In police intelligence, major crimes are not avoided. In environmental intelligence, there is widespread damage that effects people and the economy.
In workforce capability intelligence, failure can mean:
– Expense, from inappropriate promotions and reassignments
– Personal loss for the worker, facing failure from unreasonable expectations.
– Wasted effort spent on retaining staff.
– A reorganisation results in disrupted operations, because the impact wasn’t understood.
It’s important to get alternative opinions for key questions.
In addition:
– For each intelligence source, assess the reliability of the raw data, and how much reliability deteriorates in the maths or AI that derives conclusions from the data. This can be partially automatic, based on precedent, however reliability assessments can be very wrong and humans should review them.
– Human researchers and analysts should always look for outliers – cases where the intelligence could mean something significantly different to the obvious.
– Ask whether the question itself is valid. (The null hypothesis.)
– As humans, understand the strengths and weakness of the different methods, in part to identify the null hypothesis.
– As humans, be prepared to Say no! If the collective intelligence is not sufficient, refuse to decide, or demand more intelligence gathering and analysis, and if necessary do it again and again – as in the intelligence cycle, below.
Iteration – the intelligence cycle
The process for intelligence follows a cycle of 6 separate activities, that follow a sequence.
The initial Planning and Direction can be extensive, but after that the sequence is very much faster. Tools may be used to help, but it’s essentially a human activity, involving different people.
This is followed by Data Collection, which can involve delay, such as for answering questionnaires and arranging interviews.
Processing the results of the data is potentially suitable for automation, including with AI. But it still needs overview by a researcher. And where automation doesn’t yet exist, older manual methods are still needed.
Analysis and Production, brings together the processed results and produces options and/or conclusions for the decision maker. This includes assessment on its trustworthiness.
Dissemination involves getting the right information to the right people. It should be automated. However it’s also highly sensitive, because it can contain personal information.
Decision-making is for the next step. Choose between options, choose not to decide, or demand more intelligence. The process should be entirely manual. (If it was automated, there would be a variety of legal implications.)
Conclusion: 5 lessons and a takeaway
1. Talent intelligence looks only at skills, but to solve the problems it addresses, we need to look at much more skills.
2. Adapt methods from the state intelligence community.
3. Trust is fundamental – we need to aim for very high levels of transparency and reliability.
4. Workforce Capability Intelligence can be highly automated, but the human element is critical.
5. For high-skilled roles we need extensive understanding about what people are actually doing.
The takeaway:
– With Capability Intelligence, we can help people achieve more.
– And we can build organisations that are more resilient to the crises that are sadly common in our world.
References
Original intelligence cycle. See “Intelligence Power in Peace and War”, by Michael Herman. Published by Cambridge University Press, 1996. https://www.cambridge.org/core/books/intelligence-power-in-peace-and-war/39B13810C2D49FD2894827D9BA373CCB
Action-on intelligence process. See chapter 5 of “Securing the State”, by David Omand. Published by Hurst, 2010, ISBN 978-1-84904-188-1. https://www.hurstpublishers.com/book/securing-the-state/
Analysis techniques. See “Intelligence Analysis: A Target-Centric Approach, 8th edition”, by Robert M. Clark. Published, by Sage Press. https://uk.sagepub.com/en-gb/eur/intelligence-analysis/book287809 . (Note that some elements of analytic thinking, apply in research work.)
Unknown unknowns. See manage-unknown-unknowns and improving-management on this website.
Historical context. The earliest remaining thesis on espionage dates to the 5th century BC: Sun Tzu’s “The Art of War”, Chapter 13, “The attack by fire”. See https://suntzuartofwar.org/
Caveat by the author, Adrian Cowderoy
The opinions in this article are mine, and are influenced by my work in documenting aspects of the intelligence profession (here on this website) which came from my fiction-writing. The comments do not necessarily represent those of Metatalent.ai Ltd – a company I co-founded. (For more, see https://Metatalent.ai/ )
Metatalent.ai provides an AI-enabled talent intelligence platform, supporting talent acquisition, talent development and talent management. The platform provides collection, processing and analysis functions that supports Capability Intelligence, in the area of skills. However more is needed by organisations. The professional services from Metatalent.ai can help you move to this new future.