The intelligence management concept

Colin Blackwell-Chairman of HR Committee Vietnam Business Forum

AI’s limitation has always been a lack of human-like intelligence. Considerable work was done for years to advance it to mimic human intelligence, by building neural networks modelled on humans and teaching them in the way people learn. There were glimpses of this working for the last few years, but now it has been successfully released publicly for us all to see, in the form of ChatGPT.

The overall combination of human and AI workers is what companies now have as their intelligence resource, which requires a re-think of how companies are organised. The entire field of human resources is no longer adequate by itself to address the complexity of a business context containing the new AI. A new concept of intelligence resources is being organised by the new field of intelligence management.

The new intelligence management function has to deploy the maximum intelligence in a company’s activities in the fastest, best and cheapest way possible. Each task the company has to perform has to be analysed for the best fit of either a person or AI. Within every current human job description, every individual task will either be replaced or enhanced by AI.

The task lists need reviewing by an HR professional who understands the human aspects, sitting alongside an IT professional who knows AI capabilities. Between them, they will have to reassign tasks for every single job role, creating multiple new instances in the process.

This is similar to an HR headcount optimisation process, which companies will be familiar with. Activities are bundled, expertise is centralised and headcount is reduced to save cost and increase performance. But this time around it is much more complex because it is not just reassigning between people, but reassigning to very new AI capabilities.

The interdependencies and re-design across thousands of people will be a monumental task for HR and IT people working together. Paradoxically, it is AI itself that has the solution. There is nothing more capable of handling this complexity.

Thousands of job descriptions could be loaded into an AI dataset, which is then prompted to identify an ideal balance of what it can do and how the human roles can be best organised.

This can then be reviewed by the HR and IT team and run through multiple iterations before the best mix and transition plan is arrived at.

As powerful as AI instances are, they are by no means perfect. They still need managing in the way an employee does. They need all the same things – instructing, training, and reviewing their results. This will need high-level human management for the foreseeable future, so it is mostly the less senior human roles that are replaced. This also creates entirely new versions of human skill sets required.

Instructing an AI is often called prompt engineering, but this makes it sound more difficult than it actually is. Think of the search terms needed when finding something on the internet – it is just that, but kicked up a few gears. It forces a human manager to say things precisely and with full descriptions, which is not at all dissimilar to how one manages people either. In both cases, one has to set SMART objectives, meaning those that are specific, measurable, achievable, relevant, and time-bound.

Training an AI involves loading relevant data sets, explaining broader objectives and fine-tuning performance. This will be the task that replaces what most software programmers do now. Reviewing AI output will be what every manager in a company does. A core human role is to understand, sense, check, and act on the answers provided by it.

With people and AI working with such close interdependency, the new intelligence resources department will have to merge almost all the old HR and IT functions into new sub-functions. Training AI and people has to be done in tandem, creating a new intelligence training function. If a company is expanding, human recruitment and AI instance creation also have to happen in tandem, creating an intelligence expansion function, which combines IT equipment upgrade, AI instance creation, and the old HR recruitment team.

The effect on individual careers will be equally profound. There is plenty of both risk and opportunity, happing at an exponential rate of change. Those employees who struggled with the old technology will likely have a worse outcome now. People who could not do an internet search will be even more unable to prompt an AI instance.

Employees who were confused by last generation AI answers, and did not know what to do with the information, will eventually have even less value. If they were not managed out of businesses, they will present a fatal bottleneck to the overall intelligence capability of businesses.

This requires the intelligence management function to take on an organisational design role, looking at intelligence flows and interactions. An intelligence task needs reviewing and acting upon by person or AI capable of doing so, just as a human manager has to be more capable than a human subordinate to be able to lead them. All aspects of business operations need to be redesigned in terms of optimal intelligence flows, with each stage being given to the most capable resource in the most efficient manner. Hence, specialists in intelligence organisation replace what used to be the separate specialisations of HR organisational design and IT infrastructure.

Whilst it is natural to worry about are risks such as job losses, but the upside should not be underestimated either. Startups will be able to develop products and services lightning fast and with a tiny fraction of the investment needed previously (with interesting implications for the venture capital world). This will lower the barriers to starting many businesses and will create new employment.

Many new roles within existing businesses will be created. New services will become available, increasing the quality of life for consumers. Vital areas like assistive technology for people with disabilities will improve immensely.

There are guaranteed to be startups out there now, thinking how they can become giant killers, with relatively little human effort or cost, replacing old-fashioned lumbering companies. The best of luck to them, but it is also incumbent on every current corporate CEO to get in gear to set up an intelligence management function without delay.