Series guide: hub 281 | Part A overview | Part B high-risk | Part C (this piece)
This continues the “Will AI cause unemployment?” series. This piece looks at the functions being enhanced — where people still hold judgment, relationships, and governance — and a handbook managers can put to work at once.
The enhance / accelerate / eliminate three-axis framework (shared across this series)
- Enhance: people still make the judgments and hold the relationships, but AI amplifies materials, simulation, and insight.
- Accelerate: processes are compressed and cycle times shrink; if you do not rewrite the structure, you will only make mistakes faster.
- Eliminate: tasks that can be standardized, verified, and written back are automated or largely replaced.
I. Enhance: judgment and relationships are amplified
6. Legal & Compliance
AI is changing how legal and compliance work gets done, at a basic level. Companies used to review compliance after the fact — legal would look at a product only once it was built, or step in after a problem had already appeared. AI is moving compliance from “after the fact” to “in the moment”: through risk modeling and prediction, a system can monitor relevant regulatory risk while the business is running, and flag a potential issue before it lands. Contract review is changing in a similar way. Counsel used to read a contract clause by clause; AI can now help generate a first draft, and compare a new contract quickly against a standard template or an existing version, pointing out differences and potential risks at once. In a dense, fast-changing legal environment, looking up newly issued rules by hand can no longer keep up. AI can track regulatory changes across jurisdictions, compare them with the company’s current policies, and help the legal team see, in time, which changes will affect how the company operates.
In that shift, AI becomes a strong assistant for legal and compliance work. First, with natural-language processing and knowledge graphs, AI can parse contract clauses quickly, surface implied legal risks, and even generate a review summary against the company’s compliance standards, prompting counsel to look at the details that matter. Second, version comparison becomes much less laborious — whether the differences are across revisions of one contract or across clauses in different contracts, AI can highlight them visually and retrieve similar past cases or precedents for reference. AI platforms can also watch the global legal and regulatory environment around the clock: when a relevant rule is updated or a new requirement appears, the system can push an alert with a summary of the change, so the company can grasp the scope of the impact more quickly.
As AI takes over a large volume of tedious, standardized work, some traditional legal and compliance roles are being eliminated or reduced. Junior legal assistants used to spend long hours assembling contracts and filling in regulatory filings; much of that routine can now be done by automation. Comparing contracts word by word to find differences is already fading — an AI comparison tool can finish in seconds what several people would need hours to proof. Staff whose job is standardized process review will also have to change, because rule-checking and checklist verification are faster and more reliable when AI does them. In that setting, the value of legal professionals shows up more clearly in the work AI is not good at: interpreting complex provisions, negotiation strategy in unusual situations, and professional judgment in the gray areas AI cannot decide. By handing the tedious work to AI, a legal team can concentrate on higher-level risk control and strategic decisions, and become a genuine guardian and strategic partner to the business.
7. Product Management
In the AI era, the product-management role is changing from a traditional project pusher into an orchestrator of human–AI work. Product managers used to own the roadmap and coordinate resources so that features shipped in sequence. Now the PM is closer to a commander who schedules AI and people to work together. They are no longer only setting a timetable; they are designing a loop of value validation: testing quickly, and continually checking whether the design actually creates user value. PMs once spent a lot of time organizing interviews and gathering requirements. AI tools can now simulate user behavior and forecast how a design might be used, so a PM can see likely reactions before a product is launched. From a coordinator of user research, the PM is turning into a strategist who uses AI for behavior simulation and scenario rehearsal.
In that process, AI becomes a strong assistant. Through intelligent agent systems, a PM can automatically gather user feedback and product-usage data from many channels, watch competitors in something closer to real time, and turn a mass of information into a basis for decisions. Generative AI also shortens the trial cycle — for example, by quickly producing a prototype, interface mock, or copy, so a PM can run a small test, collect feedback, and iterate in a much shorter time. AI can also, from early usage data, forecast challenges after launch, and even produce a suggested list for the next iteration. That means a product manager can adjust strategy from data with more composure, instead of relying on intuition alone.
As AI is worked more deeply into product development, some product managers who lack a data habit, or who only execute, are at risk of being eliminated. A PM who cannot analyze data and decides only from experience will find it hard to keep up with a product rhythm built on AI and data. Likewise, the purely executive PM — the one who only allocates resources and tracks progress — will see much of that work taken by automation, such as a project system that sends reminders and consolidates status on its own. Another kind of PM, used to stacking slides to present a case and light on empirical validation, will struggle in an environment that favors data and speed. The next generation of product talent will need both data analysis and the skill of working with AI if they are to set product direction. They will spend more time on deeper user needs, on setting a product vision, and on working closely with technical tools to realize that vision — the bridge between user value and technical delivery.
11. Corporate Strategy
AI is also changing how corporate strategy is made, and at what cadence. Strategy used to be an annual planning discussion at the top, with the management team applying classic frameworks (SWOT, Porter’s five forces, and the like) to draw a future map. That “annual plan” model is being replaced by more immediate, dynamic adjustment. Managers are leaning more on an experimental model: small pilots and continuous data feedback to test a strategic hypothesis, and an early change of direction when needed. AI lets a company simulate situations in a virtual environment — a sudden shift in demand, a competitor’s new move, a change in the policy environment — and see earlier what different options might lead to, reducing the blind spots in a decision. Industry observation is also moving from manual work to a multidimensional forecasting system: AI can analyze economic data, consumer trends, and technology-development signals at the same time, and identify potential trends and hidden correlations faster than a person, so that planning sits on a more evidence-based forecast.
AI gives strategy teams a richer set of tools and strengthens senior decisions. An AI system can, for example, draw a competitive landscape automatically, show how the main players are positioned and where they are moving, and help a company find a gap or an opening. It can also, from a large volume of industry reports and news, generate a path of how a trend might evolve, so a decision-maker can see more directly where a technology or a business model might go in the next few years. AI can even produce a first draft of a strategy analysis — a SWOT, a suggested market-entry mode, a product-portfolio adjustment — replacing a time-consuming manual process. Further still, for each option, AI can quantify risk and return in advance: through Monte Carlo simulation or other models, it can estimate financial performance under different scenarios, and help the management team compare risk and reward before deciding.
As these capabilities appear, some traditional strategy-analysis work is being eliminated or weakened. Consultant-analysts who used to collect material by hand, cut and paste reports, and make slides to support a decision now face direct competition from machines that can search public information in moments and assemble a draft. A strategy deck that only applies a framework and lacks a distinctive insight also loses value, because AI templates already contain the standard analysis; the real difference comes from an original view. Static industry maps, showing the landscape at a single point in time, are becoming outdated; companies need a view that can be updated, which AI can provide continuously. In future, strategy professionals will put their energy into what AI cannot replace: creative strategic hypotheses, political and cultural factors that are hard to quantify, and — most important — using experience and judgment to test whether an AI recommendation is feasible, and to make the call when the choice is a hard one.
12. Corporate Communications
In a digital era, the role of corporate communications is also changing as AI comes in. Traditional communications focused on raising brand visibility, through press releases and press conferences. The emphasis now is on shaping the brand’s semantic field — its narrative space — in something closer to real time. That means a company is not only sending a message; it has to keep building a favorable context for the brand, across online and offline channels, taking part in discussion and influencing the direction of public conversation. Communications people used to write a release on a schedule. Now they pay more attention to distributing content precisely to different audiences and building trust over time — not a one-way “publish,” but a semantic interaction with stakeholders (customers, media, investors, and others) so that the company’s values and messages are understood and accepted as intended. Crisis work is also moving from after-the-fact repair to rehearsal in advance: where a team once drafted a statement by hand when a crisis hit, AI can now simulate crisis scenarios and generate response scripts, so the company can face a sudden event with more composure.
AI tools give communications teams several kinds of help, and raise both speed and precision. First, generative AI can quickly produce drafts of many kinds of materials — a press release, a speech for a senior executive, a statement for a KOL collaboration — for the team to refine, saving a large amount of writing time. Second, an AI-driven media-monitoring system can scan social media and news around the clock, and identify, in something closer to real time, a discussion or a potential crisis that will affect the brand. When negative volume rises, the system can simulate how the conversation might spread and suggest a communication strategy and wording, helping the team set a line quickly. Third, AI supports multilingual, cross-cultural communication — automatically translating a message and adjusting the wording for a local market, so the brand’s voice stays consistent globally and still close to local culture.
Some traditional communications roles are then eliminated or weakened. Demand for copy editors whose only job is to write drafts falls, because most standard releases or statements can have a first draft generated quickly by AI. Manual clipping and social-listening roles are also gradually replaced — where people once had to read coverage every day and collect competitor news, an intelligent system can now assemble and analyze that information, more immediately and more completely. Support staff who only made slides and compiled communications materials also face pressure, because AI can visualize key information and generate a report at once. In future, communications professionals will put more of their energy into strategy: inventing a communication approach, personally managing media and community relationships, and making sure every public conversation fits the company’s long-term image. In short, AI takes on a large volume of tedious execution; people concentrate on using judgment and ideas to protect reputation and build a lasting bond of trust.
13. Research & Development (R&D)
AI is also changing R&D in a material way. Traditional research often followed a linear path: a researcher formed a hypothesis, ran experiments in sequence, changed one parameter at a time, and then decided the next step. With AI, research is moving toward parallel rehearsal of hypotheses: through computer simulation and machine-learning models, a team can explore several scenarios at once, and have AI evaluate which paths are most promising for further experiment. Parameter tuning that used to require repeated manual trials can now be handed to AI, which generates combinations of variables and searches for a better solution — whether a formula ratio in a chemistry experiment or a model’s hyperparameters, an algorithm can often find a strong setting more quickly. And where scientific knowledge once depended on what a researcher had accumulated and looked up by hand, with inevitable gaps, a team can now compare the latest papers and patents in time, so that an experimental design sits on prior work and does not repeat a dead end someone else has already walked.
AI tools give researchers a strong boost and speed up the path to something new. In new-materials or drug development, for example, AI can suggest candidate formulations or molecular structures from a target property, and shrink the range of trial and error. It can also suggest optimizations in engineering design: from the shape of a part to the efficiency of an algorithm, machine calculation can find a near-optimal option. Experiment data management and analysis also benefit: steps and results can be logged automatically, archived in a structured way, and turned into charts so a researcher can see a trend at once. During an experiment, computer vision can help track an observation (for example, a subtle change in cells, or in a materials stress test) and surface something the eye would not catch in time. On the knowledge-search side, a researcher no longer needs to spend long hours hunting literature: AI can gather relevant papers and patents, and quickly produce abstracts and key findings, so the team can keep up with the global pulse of the field.
As these capabilities come online, a good deal of traditional research-support work is being eliminated or reduced. The technician role dedicated to repeatedly adjusting experimental parameters is in less demand once automated lab platforms and optimization algorithms are in place — a machine can try combinations around the clock, and find a better condition faster than a person. The assistant who looked up literature also becomes less necessary; AI can finish a literature review and extract the main points in a short time. Hand-drawn charts and manually compiled lab records have already been replaced by visualization tools and electronic lab notebooks. Future researchers will concentrate more on what AI cannot replace: posing a genuinely original research question, applying reasoned skepticism to an algorithmic conclusion, and turning a technical result into a concrete application. AI will be a strong assistant to the research team; people will still set the direction and give the work its meaning.
II. Organization-capability functions: curating human–AI collaboration
10. Human Resources
AI is also changing what human resources work consists of. HR used to be centered on hiring, performance, retention, and development — a “manager of talent.” In the AI era, HR is gradually becoming a curator of human–AI collaboration systems. As a large volume of repetitive work is taken by AI, HR’s job is no longer only about people; it also includes designing how people and AI work together, bringing in the right intelligent tools, and making sure employees have the skills to work with AI. Performance management is also moving from a formal annual KPI review toward a more dynamic analysis of behavior and potential: AI can follow an employee’s behavioral data across work platforms over time, and help judge actual contribution, collaboration networks, and growth potential, so HR can more accurately find a star, or someone who needs support. In recruiting, the manual screening of résumés and the repetitive scheduling of first interviews are being replaced by smarter matching — AI can parse a role’s needs and a candidate’s materials, produce a shortlist, and even run a first screen through a chatbot, raising hiring efficiency substantially.
AI tools also give HR several kinds of enhancement, and put talent development and organizational care on a more evidence-based footing. In hiring, for example, an AI interview assistant can score a candidate from facial expression, voice, and answers, and estimate fit and future potential as an input to a human decision. After joining, each employee can have a growth path built by AI: from career goals and current skills, the system can recommend courses, on-the-job learning, or a mentoring plan. Further still, through sentiment analysis of internal email and work chat, AI can monitor mood and team morale closer to real time; if it detects a signal of high stress or rising attrition risk, it can prompt HR to step in earlier. These techniques move HR from reactive to proactive, catching a problem earlier and offering support closer to a person’s actual needs.
With these developments, some traditional HR functions are being eliminated or weakened. Recruiters who screen résumés by hand will be needed in much smaller numbers, because AI can read a large volume of CVs in a short time and pick those that meet the criteria. In training administration, the practice of tracking course progress one by one and teaching from a standard slide deck also fades as intelligent learning platforms arrive — AI can not only remind people to finish training, but also suggest the next step from what they have learned. Roles that exist mainly to update personnel data in a spreadsheet or calculate attendance bonuses also shrink as an HRMS automates them. Facing this change, HR practitioners will focus more on the new problems of people and machines working together: shaping culture, setting a future talent strategy, attending to mental health and career development, and acting as a bridge between senior management and AI systems, so that the technology fits the organization’s values and actually enables people.
14. Learning & Development
Pushed by AI, corporate learning and development is also changing form. Training used to run mainly through a fixed course platform — employees attended according to a shared schedule, and an instructor taught from prepared materials. The work is now moving toward more flexible learning-scenario design and a personalized cadence. Learning is no longer limited to a classroom or a single online course; it is being worked into the daily job, so that people can get knowledge when they need it. Companies are relying less on large, centralized training and more on a real-time recommendation system: AI, from each person’s current role and skill gap, can recommend a resource or a micro-course at any moment, so that people keep growing in the work itself. Training effect used to be measured after the fact, with a survey or a test. Now, with live learning data, HR can see progress and results as they happen, and adjust in the moment.
AI tools give L&D teams a strong assist. First, AI can generate a learning map for each employee: from personal goals, skill level, and career path, it can plan a suitable combination of courses and a sequence — a learning plan built for that person. It can also filter, from a large internal and external knowledge base, content that fits, and push a relevant article, video, or exercise. Second, in delivery, generative AI can automatically produce practice questions and interactive case drills — in sales training, for example, AI can simulate different kinds of customer conversation, let a learner practice, and give feedback at once. The learning path is logged throughout: which exercises were finished, which items were missed repeatedly; AI can adjust what comes next, and actually teach to the person. Finally, training administration becomes more efficient and more transparent: AI can assemble each person’s results, produce a report on knowledge gained and on whether behavior has changed (for example, a lift in performance), and let management see the return on the training investment more clearly.
As AI enters corporate training at scale, some traditional L&D work is being eliminated or transformed. The internal trainer who only reads from a standard deck will be needed less often — repeating slides no longer holds attention, and is being replaced by more immersive, AI-driven learning. Course operations that used to be done by hand (scheduling, notifying learners, collecting feedback) are also automated; an LMS can handle notices and progress tracking. Paper exams and manual marking are already history; online tests and AI scoring not only reduce administration, they can produce an analysis at once. In future, L&D professionals will put their weight on growing a culture of lifelong learning, finding the new skill areas the organization will need, and designing learning programs with a real impact. They will work more with AI, and use data to improve the learning strategy, so that the company’s talent keeps evolving for what comes next.
Closing: take up the change, and work with AI
Taken together, what AI brings is not simply a crisis of unemployment, but a reorganization of the workplace. Every function is being redefined under this pressure: roles that know how to work with AI and create new value will be enhanced; the parts of a process that can be automated will be accelerated; and the links that lack a distinctive human value will, in the end, be eliminated. Future competition at work will no longer turn on seniority, or on fluency in a fixed process. It will turn on whether you can become a new kind of role that can handle AI with some flexibility — able to mobilize intelligent tools, connect human insight with technical capability, and produce results the old model could not.
For an individual, the practical advice is to take up this change rather than wait it out. Keep learning about AI and its tools, raise your own digital literacy, and treat AI as an extension of daily work rather than a threat. Concentrate on building the capabilities a machine cannot replace — original thinking, strategic judgment, empathetic communication, and the ability to integrate across domains — so that you can do more when you work with AI. For a company, the work is to set a talent strategy for an AI era: invest in employees’ AI skills, spread good practice in human–AI collaboration inside the organization, and, at the same time, re-examine current processes and role design, and be willing to adjust so that AI’s strengths are used and people are placed on higher-value tasks. Leaders need a culture that allows a team to try a new, AI-enabled way of working, and a clear ethics and governance frame so that the use of AI fits the company’s values.
We are at the beginning of a historical turn. To watch from the side, or to resist, is to be carried under by the current of change. To choose to adapt and to improve is to have a chance to use this force, and to open a new stretch of a career or a business. The future belongs to those willing to work with AI: to take up the technology, keep a human starting point, and lead in a movement played by people and machines together.
The judgment, restated
My judgment is that the functions being enhanced will win on judgment, governance, and the orchestration of people and machines. A manager’s job is to rewrite the role, not merely to buy a tool.
Three things this week
- Pick one enhanced role and rewrite it in three columns: what a person does / what AI does / what you accept together.
- Set one escalation path for human–AI exceptions (when a person must sign).
- Hold a 30-minute alignment this week: which KPIs should be retired, and which should be rewritten as intent metrics.
Return to the series hub for the full directory of 15 functions, or to Part B to check high-risk tasks; the overview is in Part A. Subscribe to Future Insights Lab, or reply: which cell does your function sit in?