AI#-6 - How AI Will Transform The Workplace
How will AI transform HR departments? How will employees get trained and re-skilled in an AI world? Which functions will be disrupted the most? What will an AI-enabled employee look like?
We constantly talk in the news today about the innovation AI can bring to our daily jobs and dream of days when Tim Ferry’s 4-hour week is a reality. I find it equally fascinating to explore the implications of AI in the workplace. How will AI transform HR departments? How will employees get trained and re-skilled in an AI world? Which functions will be disrupted the most? What will an AI-enabled employee look like? Is AI in the workforce good or bad for graduates?
Those are the questions I will briefly cover today. I still do not have a crystal ball, but Christmas is coming. Take my analysis and predictions with a pinch of salt. No one can predict the future of AI.
Let’s get started!
Towards An AI-Powered HR System?
In 2020, Eightfold.ai surveyed 250 HR professionals on their use of AI for various HR functions, including employee records, payroll processing, recruitment, and onboarding. An overwhelming 92% planned to expand AI use in HR within the next 12 to 18 months, especially in performance management and payroll processing. According to IDC, by 2024, 80% of global 2000 organizations will employ AI/ML "managers" for employee-related decisions.
An ex-colleague of mine interviewed with a company using AI tools in interviews. The AI recorded the discussion and generated a report for the recruiter. It is unclear whether the AI performed body language analysis and voice pattern recognition. It is just a matter of time, anyway.
ChatGPT or Bard can already help draft interview questions. Soon, AI may become an active participant and suggest questions in real-time. Otter.ai already supports that use case for work meetings; the extension to interviews is a logical one. AI tools for the HR department are already widely available. You can read more on this topic here, here, and here. I could not find any data that would substantiate the claims AI tools are making.
Screening interviews could easily be automated. An interviewer may work with his AI to prepare the interview questions and let the AI run the interview. The AI could analyze all interviews, rank the applicants, and make a hiring recommendation. My niece had to submit a recorded video for a job application. It is a no-brainer to run a fully automated live interview with an AI bot or avatar that is skilled at discovery calls.
What about avatars? It becomes harder and harder to tell them apart from real humans. Are you ready to interview with an avatar? A company like Amazon could save time and money on entry-level jobs and high-volume hiring to prepare for the shopping season.
From this angle, one could think that HR will have all the leverage. You would be wrong. There are numerous AI tools to help job seekers. Zain and Awais Khan, authors of the Superhuman newsletter, shared this AI + ChatGPT job search cheat sheet and a list of AI tools for job hunting. A simple search on how to use ChatGPT for a job search will return many results. This Linkedin post helps you write prompts for your resume and mock interviews. There is a huge market for AI apps targeting job seekers. I would not be surprised if we saw a few unicorns.
Employee Performance Management (EPM)
Every company is struggling to get new hires up to speed quickly. In the future, once hired, the employee will be paired with an AI assistant that will guide him through the onboarding process, tell him what to read, the meetings to attend, the people to meet, and more. Every company is struggling to get new hires up to speed quickly. The market is there for the right AI tool. Interestingly, it may be easier for startups to penetrate the Enterprise in dire need of such technology. It is going to be harder to build an AI onboarding app for small and midsize companies due to the lack of data and limited training opportunities for the AI model.
Several apps exist on the market today that focus on EPM: Lattice, BambooHR, 15Five. They promise to drive employee performance and engagement, help manage career growth, and increase overall retention.
EPM includes processes such as regular 1:1 meetings with your report, a quarterly performance review, a yearly career review, an individual development plan, a salary raise, and a bonus.
This field is ripe for disruption, IMHO. Managing employee performance is hard. If you are a manager, chances are that you have received training and been asked to follow your company's performance management framework. Now, think of an AI assistant that coaches you. An AI assistant that helps you organize your employee performance tasks for each individual report. That would be handy, don’t you think?
The AI could help you ask good questions in 1:1s, take notes on your behalf, create summaries for your quarterly check-in, and make factual recommendations on salary raises, bonuses, promotions, or PIP. Each manager would run its own instance of the company’s AI. You could imagine the HR system being built on a centralized machine learning model that helps ensure fair and equitable performance management across all departments. This would be disruptive!
It does not stop here. Imagine all employees having a company AI assistant for their daily work. The AI would know everything about you. It would analyze every interaction, what you say in a meeting, what you write, and more. With all those insights, it is not far-fetched to think that your AI and your manager’s AI would be able to exchange information about you and make a recommendation about your promotion.
For this to happen, the company would have to move to an AI-based HR system. Once adopted, what is the true role of the manager? Who is truly in charge of managing you? Those are interesting ethical questions for the reader to ponder.
White Collar Productivity Gains
There is no doubt AI apps and systems will bring tremendous productivity gains to white collar jobs. It is more a question of when than if. Let’s walk through a few practical examples.
Finance: this department spreads across multiple areas and teams: general ledger accountants, account payables and receivables, payroll, tax, and financial analysts. Each area may comprise large teams. Expert AI financial tools will automate manual tasks and assist experts in doing in-depth financial analysis. The outcome? A flattening of the organization with just a few experts equipped with the best AI tool to do their job.
Legal: the same paradigm could repeat. Companies need to abide by strict legal frameworks, and AI, once trained properly in a legal field, should surpass a lawyer in understanding the intricacy of legal systems and use this knowledge to draft legal documents compliant with local laws: sales contracts, employment contracts, HR employment and labor laws, litigation, and IP. A friend of mine at a tech startup used AI to draft a complex employment contract for an employee working abroad. His lawyer simply reviewed the draft and approved it with no edits! There was no need to involve his legal assistants in doing research and drafting the proposal. My friend saved hours in legal fees.
Sales: companies like Hubspot and Salesforce are already bringing intelligent tools to streamline business processes. The entire deal flow, from prospecting to negotiating and closing the deal, will soon be entirely AI-enabled. OneShot.ai, founded by an ex-Salesforce engineer, is building a generative AI platform to automate the SDR calls. Because buyers will use their own AI tools to source and select vendors, the role of sales rep may be reinvented. Intelligent marketplaces may play a big role in disrupting traditional sales by forcing vendors to list on marketplaces, at least in High-Tech and Manufacturing. Will we see a world of AI-to-AI sales with no physical sales rep or buyer?
Marketing: This field is poised for disruption with a plethora of AI tools. AI apps such as Jasper.ai, Rapidly, and Fraise.io promise to help you generate marketing content such as emails, landing pages, social content, and SEO optimized websites. CRM companies like Huspot and Salesforce are beefing up their marketing tools with AI for campaign management and orchestration. The future marketing department may be much smaller, with specialists covering broad areas thanks to AI assistants.
AI-assisted work paradigm
In a recent article, AI-04 How AI Will Transform The Software Industry, we mention Ajay Agarwal’s book Power and Prediction and three levels of innovation: AI point solutions, AI app solutions, and AI system solutions. The focus so far seems to be on AI-enhanced workflows and automation through point solutions and apps. We have yet to see system disruptions (HR is a candidate).
Those changes are happening right in front of our eyes. Microsoft relaunched the Office 365 apps suite with Microsoft 365 Copilot, which combines the power of LLMs and your data from the Microsoft Graph. AI is embedded into the app as its own operating system. Google is following Microsoft with their Duet AI for Google Workspaces, with the goal of boasting the productivity of Google Apps users. We also see tremendous innovation within single-purpose apps that help you with meetings, spreadsheets, presentations, videos, and so on. It seems that new AI apps launch every day. I stopped counting.
How does it impact the workplace? For one, it makes employees with the right AI tools more productive. Less time is spent on low-value tasks (email management, scheduling meetings, reviewing long papers). It is possible that the trend in the workplace is towards smaller teams with experts covering more ground.
By pairing SMEs with a specialized AI assistant, we may see 1x to 3x productivity improvements, especially in those functions with many supporting roles. The AI will know your preferences, organize your week, help you think about challenges, and work towards a resolution.
Besides the reduction of supporting jobs across functions, there are longer-term considerations. The reduction of mid-level white-collar jobs creates a skill gap. To be hireable, you may suddenly have to become an expert in your field and work well with AI tools. In such a market, employers will have the upper hand, with more demand than offers. The natural career evolution from entry-level to mid-career and executive may be disrupted.
Pushing the reasoning one step further, the only constant in this scheme is the AI. My concern is a world where AI becomes the de facto source of truth, whether it is accurate or not. Functional experts could become irrelevant. Lower-wage employees, trained by expert AIs, would do their job.
I am not saying this is what is going to happen. I am just analyzing a possible outcome if AI is truly successful at helping us in our jobs or making us believe so. The counterargument consists of saying that AI is not capable of doing all those great things and that it will not replace human expertise any time soon. We already see an entire pan of jobs at risk, as discussed in AI-05 - How AI Will Transform Business. We’d rather be prepared than sorry!
AI = Workplace Disruption?
The more I look at AI, the more I believe it will further accelerate business automation across industries and functions that had plateaued after the IT and digital revolutions. Because automation sells and represents one of the most obvious ways to cut costs and improve margins, executives will invest first in automation use cases with potential disruptive effects on the workforce.
In my opinion, the following workforces are most at risk in terms of total job loss:
Support Agents: Tiers-1 agents are the most at risk of being displaced; estimates put it at 70% of the workforce (tens of millions of jobs); eventually, AI disruption might move up to Tiers-2 and Tiers-3 agents and organizations lean towards the AI-assisted worker paradigm we discussed above
Warehouse Workers: Companies like Boston Dynamics are revolutionizing this space with low-cost robots for warehouses; the writing is on the wall. Sadly, this will remove well-paid blue-collar jobs that support middle-class families.
Waiters and “assembly cooks”: Covid showed us you do not need waiters; AppleBee's, for instance, lets you scan the menu and place the order from your phone. With waiter robots, you do not need waiters anymore. The same is true in the kitchen with assembly cooks (low-skilled, more of a prep cook), who can be replaced with robots.
Stores And Supermarket Cashiers: robots and IoT automation might make cashiers irrelevant. Amazon has been experimenting with cashierless stores. This low-paying job is at risk of slowly disappearing, leaving millions of low-skill workers out of work.
I am not making predictions, but if this holds true, this is both an economic and societal tragedy in the making. AI may further accelerate the divide between the haves and the have-nots. I am for progress and see many benefits to AI, but it cannot come at the cost of millions of individuals and families.
Graduates And Young Professionals
Are today’s graduates at an advantage or a disadvantage when it comes to the AI revolution? Will schools offer them the proper training to be employable in an AI-first world?
I touched on the impact of AI on schools and universities here. The public school system is very slow to change and has been on a downward performance trend. It is best shown in the continuously declining literacy rate in the US Army admission tests over the last fifty years. Public schools are not better, but teachers do not fail kids anymore and have become easy graders—my 9th grader here in California can re-take a test she failed as many times as she wants now. Universities ain’t any better, unless you can afford to pay $100K+ a year to go to Harvard, MIT, or Stanford... and still.
Where does it leave most of the students? On their own. AI clearly puts them at a disadvantage. The skill gap is going to further increase, and those entry level jobs will be easily replaced by AI.
Let’s assume that AI truly disrupts the workforce. Three scenarios could play out:
Disrupted By AI: Students will select fields without realizing they are impacted by AI, leading to uncertain job prospects after graduation.
At Risk Of AI Disruption: Tech may be the best example. Imagine pursuing an engineering degree only to discover in four years that many engineering jobs are disappearing due to AI.
Resistant To AI: Savvy students will opt for careers that AI cannot replace. Any plumbers or electricians out there?
If AI truly enables employers to reduce their workforce in critical functions like sales, service, finance, and marketing, it may well mean that they are incentivized to keep a few highly skilled workers and cut low-entry jobs. This would make it much harder for graduates to get their first job and gain precious experience. There are plenty of resilient job opportunities out there. It will just require a different mindset, especially in wealthy communities where university diplomas are considered the holy grail for success.
Workforce Outsourcing
Since the nineties, millions of jobs have been outsourced to low-wage countries: manufacturing, support centers, and tech hubs.
How will AI impact this workforce? It is hard to say. A company has little to no incentive to replace a low-cost workforce. However, some companies may be tempted to train AI-enabled workers in low-cost destinations. That would be a direct threat to highly paid white collar jobs in developed countries.
I have no doubt that AI will impact the outsourcing industry, especially in those industries and functions primed for automation. It is still murky to me how this could play out.
Conclusion
AI will impact the workplace in the next few years and possibly redefine how we work, collaborate, and deliver value to our employer. Companies that invest in mature AI tools will gain productivity and reduce costs. If and when this revolution happens, white-collar workers may be impacted across functions. New graduates with no functional experience and a lack of proven AI expertise may find it hard to get a job that allows them to repay their staggering student loan. On a positive note, change takes longer than expected to happen, giving us plenty of time to prepare. Are we overoptimistic about the true potential of AI?
It’s worth looking at/listening to what Tobi from Shopify has done using AI at Shopify. It’s a great direct use case of the impact of the technology, in a way that’s not just ‘oh look we now use AI wow’. He recently did an interview at the All-in Summit talking about the co-pilot he built. Really great use case and real-life example of what you touch on in your article.