ARTICLE: HOW WRITERS CAN FLOURISH WHEN AGENTIC AI BECOMES THE NORM
“It is not the strongest of the species that survive, nor the most intelligent, but the one most responsive to change.” Charles Darwin
I put together this report with the help of an AI Agent that analyses research.
Generating this kind of report helps me understand complex subjects fast.
Like most of us nowadays, I use a ‘conversational’ structure of prompts to assemble the information.
I’m a research geek.
Luckily, as a writer and copywriter, research is one of the most aspects of my craft.
In this case, I’ve organised the text a bit. But I haven’t performed an in-depth copyedit.
There’s no need to add any artistic flair to it.
It’s a report - so I’ve left it as it is.
Therefore, although I apologise for the ‘raw AI words’ and lack of pazzazz, it’s useful reading for anyone who wants to know:
What’s going on the world of work
Whether creativity matters any more
How writers can flourish alongside Agentic AI
My own thoughts are in brackets and italics.
Summary - THE BIGGEST CHALLENGES ARE NOT TECHNICAL
This report is a strategic analysis.
It’s based on a key text from McKinsey, "The Future of Work is Agentic."
The main finding is that Agentic AI is a major change.
It’s different from generative AI.
Agentic AI can “perceive”, decide, and act on its own (this just means it’s really good as sifting through a lot of data and making connections. It’s not thinking like you or I do. It can’t. It doesn’t have consciousness).
This is creating a "parallel workforce" of digital workers.
For leaders, this means a new approach is needed.
AI should not be seen as just a tool. It must be managed as a new kind of digital employee.
This requires IT and HR to work together.
They must manage a mix of human and AI workers.
The biggest challenges are not technical.
They are about how to manage an organisation and its people. Leaders must prepare for this.
They need to manage change and retrain employees.
The goal is to build trust and adoption.
In the end, successful companies will not just automate tasks.
They will use their human workers for empathy and complex judgment.
This will give them a unique advantage over competitors.
Part I: Agentic AI - A New Paradigm for Automation and Judgment
1.0 Defining the Distinction - From Generative to Agentic
The conversation about AI often focuses on generative AI.
This is a technology that creates content.
It works based on a person's prompt or instruction.
But the research points to a new evolution.
This is the shift to agentic AI.
This new technology is not just an improved version of generative models. It’s a completely different kind of system.
It can’t only generate content - it’s also able to carry out a complete and complex task or a full project.
It does this with some autonomy and ‘judgment’.1
The main difference for agentic AI is its closed-loop system. This system includes what is known as computer perception, decision, and action.
An AI agent analyses what’s happening. It applies an algorithmic type of judgement. Then, it takes a specific action.
That action then gives feedback to the agent.
This helps it learn and improve its future performance.1
This process changes the technology. It goes from a creative tool to a proactive worker.
It can take on entire processes.
The article calls this the "next step" in AI.
It says an agent can decide what to do on its own.
It could even create a "digital replica of the entire workforce of an organisation."1 (I’d like to see how that works out. In many cases it would probably be at best a bit weird. And at worst an utter sh**show. But that’s just my opinion).
This is a huge leap.
It moves from helping with single tasks to automating entire business processes.
This means a much higher potential for productivity. But it also means more complex implementation. (And, I would suspect, more room for utter disaster on a company-wide level if not done with care. I’m thinking of the British Post Office workers who were blamed for the mistakes of a digital system. Some of them lost everything. )
1.1 Use Cases and Early Adopter Insights
Agentic AI is still very new.
Early tests are happening in "deterministic" environments. These are places with clear, rule-based processes.1
They are safe places to test the AI. The McKinsey analysis shows some early examples:
IT and Customer Service: Agents are used for common, high-volume tasks. This includes managing IT help desk tickets. They also handle customer service questions. Agents can solve simple problems by themselves. They can also send complex cases to a human. They learn from each interaction to get better.1 (Most people hate talking to a customer service robot. They’re often useless and frustrating. But maybe they’ll improve.)
Human Resources: The uses in HR are very promising. The article notes agents are used for hiring. They can "clean records, score and rank candidates, and schedule interviews."1 This automates the most time-consuming parts of hiring. It lets human recruiters focus on building relationships with candidates.
Training and Development: Agents are also used for training. They can simulate customer calls for employees. They offer "live, detailed scoring and coaching."1 This gives supervisors specific feedback on skills. (Sounds a bit odd to me. I’m not sure if any form of training beyond the most basic kind of thing will work with bots. I’ve been teaching for decades and it makes no sense. But there you go. Once again, maybe they’ll improve.)
The report also mentions a "coordinating agent."
This AI is designed to "interact with these different, underlying agents."1
This is a new model for company structure.
Instead of a human management team, we might see AI "managers" overseeing AI "teams."
This structure can be scaled instantly. It is not limited by human managers. The idea of entire departments with "zero-FTEs" or zero full-time employees is a complete new way to think about corporate structure.1 (I wonder if this will mean more freelance work for copywriters, content writers and so on. If so, let’s hope this goes with good pay and reasonable deadlines. Not sure if robots can set humanly possible deadlines, but we’ll see. Quality takes time. And creativity doesn’t grow on trees.)
Part II: The Strategic and Economic Imperatives of an Agentic Workforce
2.0 The Decoupling of Capacity and Monetisation
Agentic AI brings a new idea to business strategy.
It "decouples the creation of capacity" from the "monetisation of capacity."1
In the past, a company's ability to do work was linked to its number of employees.
As a company grew, so did its workforce.
Agentic AI changes this.
It creates a digital workforce that can be scaled. It is not limited by human labour.
The immediate economic effect is a dual choice for leaders.
The McKinsey report says about a third of executives might use AI to cut jobs in the next 12-18 months.1
But a larger group, almost 50%, is considering using AI as "digital labour."
They will keep their staff and boost productivity instead.1
This second view sees the true value of AI. It goes beyond just cutting costs.
The analysis shows that AI can create new services.
These services were once "cost prohibitive" with a human workforce.1
The example of a personalised travel concierge shows this. A human version would be too expensive. An agentic AI can offer it on a large scale.
This turns a luxury service into something many people can access.
The real value is not just in making things cheaper.
It's in using this new capacity to create new services and revenue. (And that takes imagination, and creativity. Oh good.)
2.1 Building the Parallel Workforce: A Class of Digital Labor
Companies must start thinking of AI agents as a "parallel workforce."
This is a new class of digital workers.1
This is not just a new term.
It is a vital step for strategic planning.
It means companies need a formal strategy for their AI workers.
This is just like the strategy they have for their human employees.
The analysis suggests that some companies already do this. They are exploring "zero-FTE departments" run only by agents.1
The idea of "digital labour" changes how we think about "employee" and "work."
This is more than just adopting a new tool. It is about adding a new kind of worker to the company.
This brings up new questions.
How is an AI agent's performance measured?
Who is responsible for the agent's mistakes?
How do we account for "AI headcount" in company reports?
Companies and regulators will need to find answers to these questions. (And no doubt train a bunch of human workers to watch over these little digital employees so they don’t get all glitchy.)
2.2 The Evolution of Enterprise Functions: IT as the New HR
Adding an AI workforce will change the roles of IT and HR.
The article quotes Nvidia CEO Jensen Huang.
He says, "IT becoming the HR of AI agents."1
This suggests a new reality.
The IT department's job will grow.
It will go from providing technology to managing a digital workforce.
This includes its development, training, and performance.
Creating and managing an agent is a team effort. It requires new skills.
The article lists three steps:
Defining the Business Need: First, define why the AI is needed. Set the rules for its function.1
Development and Procurement: Next, build or buy the AI. This might be a mix of both.1
Tuning and Training: The "tuning" process trains the agent. This needs a team. A subject matter expert guides the agent. A content specialist provides data. A prompt engineer refines its performance.1
The analysis shows that IT and HR will depend on each other.
IT will handle the technology.
HR will be "critical in driving change management and helping employees adapt" to the new workforce.1
This requires new levels of collaboration.
The old way of keeping departments separate will not work.
Companies must break down these silos.
They need a united strategy for AI.
This might lead to new roles, like a Chief AI Officer (CAIO). This person would connect IT and HR. (And what the Chief Creativity Officer - who connects the AI with tools and content that reflect the company’s branding?)
Part III: Navigating the Human Element: Challenges and Opportunities
3.0 The Human-in-the-Loop: Trust, Adoption, and Resistance
Technology is important, but human acceptance is key.
The McKinsey analysis points to a big challenge in adoption.
It notes that "newer employees tend to embrace AI faster."
However, "more tenured employees may resist it."1
This age difference can cause internal problems. It can slow down progress. (I don’t think it’s about age. This is a typical and very boring attitude. I’m not the ‘right age’ to be into AI, but I know more about it than my adult children. Stop with the ageism. It’s so 20th Century. Companies probably just threaten the older employees with forced redundancy without offering them the change to retrain.)
A deeper challenge is a lack of trust.
Employees may not trust a "black box" AI.
This can lead them to "duplicate work by double-checking the AI's results."1
This defeats the purpose of automation.
It creates wasted effort.
The report says these are not just technical problems.
They are human and organisational problems.
Leaders must take charge.
They need to provide a "clear mandate from the top." Leaders should "role model the use of AI."1
Successful AI implementation is a change management problem. Leaders must create a safe environment.
Employees must feel secure. They need to trust the new technology.
Their roles are not being removed, but changed.
If companies do not address the human side, they will face friction and wasted money.
3.1 Reskilling at Scale: The New Differentiators
As AI agents take over simple tasks, new human roles will appear quickly.
The analysis from the article notes a demand for roles like "prompt engineer" and "content specialist."
The supply for these jobs is low.1 This means a huge effort is needed to "reskill at scale."
The workforce must be prepared for the agentic future.
But the biggest change is in what skills are valuable.
Simple, task-based skills will matter less.
Uniquely human skills will matter more.
The article provides a key insight. The ultimate advantage for a company will be a human workforce that is "more empathetic."
It provides a "superior service" and a "human touch."1 This is supported by survey results.
They show that younger generations (as do older generations - duh) still prefer human help for unique customer issues.1
AI will automate predictable and repetitive work.
This makes empathy, emotional intelligence, creativity, and problem-solving much more valuable.
This is a complete re-evaluation of human skills.
Company training must change. It needs to move from teaching specific tasks to building these higher-level human abilities.
Part IV: Synthesis and Strategic Recommendations
4.0 A Framework for Adoption
Organisations should use a structured approach to adopt AI. T
his plan must address both the technology and the people.
Phase I: Strategic Rationale and Definition: The first step is to define the business need for the AI.1 This is a strategic step, not a technical one. It focuses on finding the right business problems to solve. It identifies which processes are best for AI automation.
Phase II: Development and Tuning: The next phase is to build or buy the AI. This may be a mix of both.1 Then comes the "tuning" process. This is a team effort. It needs input from experts to guide the AI. Content specialists provide data. Prompt engineers refine its performance.1
Phase III: Integration and Change Management: The final phase is about the human side. A flexible plan is needed because the technology changes so fast.1 Success depends on a clear direction from leadership. Leaders must be a model for using AI. The performance of both human and AI workers must be evaluated together.1 This integration must build employee trust. It must ensure people feel secure.
4.1 The Human-Centric Advantage
The final analysis shows that AI agents will automate many tasks.
But the ultimate source of a company's advantage will not be technology.
It will be the human workforce.
They will be empowered to provide a "superior service" and a "human touch."1
This is something AI cannot do.
The article is optimistic about this. It sees the potential for personalised services on a massive scale.
It believes humans will be free to focus on non-repetitive tasks.1
The future of work is one where human empathy, creativity, and judgment are the skills that create success.
Conclusion: The Agentic Future and the Empowered Human
The future of work is not about machines replacing people.
It is about a new digital workforce working with the human workforce.
The most successful companies will manage both workforces well.
They will use human empathy and judgment as their main advantage.
Agentic AI will change how work is done. But its success will depend on a company's ability to manage its people.
It will not depend on its ability to manage technology.
By using AI as a new type of digital worker, companies can unlock new levels of productivity. They can create a future where technology and humanity work together.
References
1 Amar, J., Weddle, B., Hancock, B., & Rahilly, L. (2025, June 3). The future of work is agentic [Podcast transcript].
McKinsey Talks Talent. Retrieved from https://www.mckinsey.com/capabilities/people-and-organizational-performance/our-insights/the-future-of-work-is-agentic
