Where we are now with Intelligent Tools (AI and Data)

 

This video by John Sumser is his talk on where we are right now with Intelligent Tools (AI and Data) in the HR and Recruiting space in the times of the coronavirus pandemic. Spoilers: it’s probably not what you would expect. John covers the key points companies need to understand about AI and other intelligent tools to successfully implement systems and gain their full benefits.

 
 
You can view the full transcript below the Q&A section below.
 
 

 
 
We’d love to hear your thoughts about this presentation. Post your comments here on HRExaminer.com or John’s YouTube Channel »
 

 


 
Q&A

 
There are several key points here:
 

  1. An intelligent tool is something that uses machine learning, AI, or some other form of advanced statistics to help you make sense out of data. They are very good at understanding the past and not so good at understanding the future. I prefer to call them intelligent tools.
  2.  

  3. Any intelligent tool is only as good as its data. As we’ve learned from the pandemic, sample size matters. So does how you choose and define what you are measuring.
  4.  

  5. All intelligent tools require diligent maintenance both of the tool and the data itself. The underlying models have a tendency to drift and need readjusting from time to time. Sometimes, when the circumstances change radically, the entire model is thrown off kilter and must be rebuilt.
  6.  

  7. Using intelligent tools is a journey, not a destination. The best result possible is that your questions will continuously get better. Adoption and usage of these tools will make your organization more flexible over time.
  8.  

  9. Intelligent tools cannot give you answers. They can offer solid input to your decision making. But the decisions should always be made by people. Intelligent tools are assistants.
  10.  

  11. The pandemic is going to change our ideas about privacy and what can or should be monitored. This will vary between cultures.

 
How will AI reinvent HR functions and transfer the way we work, taking into particular consideration the current state of affairs presented by COVID-19?
 

  1. It’s an extremely interesting time. The pandemic and our responses to it will forever change the way we think about HRTech in general and AI/intelligent tools in particular. Many things that we have taken for granted are upended by our current circumstances. This is partly due to the fact that there has not been a pandemic since the HR Tech and digital data gathering began and we will need to see if our pre-pandemic data is still useful, especially for predictions about our current pandemic conditions and what happens next.
  2.  

  3. Tools like Socrates.ai integrate all of the HRTech systems, employee manuals, and other sources of policy into a single conversational database. It’s less of a chatbot and more of an intelligent knowledge manager. As work becomes more and more distributed, HR’s role as the provider of answers and continuity grows. This is an area where automation really helps.
  4.  

  5. We may not be quite so desperate for scarce talent going forward. Succession management, capacity analysis, and detailed workforce planning will become central issues. It’s more that the pandemic has alerted us to a weakness. While we were looking for help outside of our organizations, we lost track of our internal expertise.
  6.  

  7. In that regard, keep your eyes out for individuals who have exclusive expertise. Sometimes, they are the only one who knows all of the secrets to making some process work. These are the actual essential workers. Find them and develop replacement plans. Figure out who can replace them if they get sick or die.
  8.  

  9. As we head towards the pandemic’s next plateau, it’s time to start thinking about what business you’ll be in then. Make decisions about restructuring and efficiency with something more than cost savings in mind. Many industries will be forever changed.
  10.  

  11. It may be a while before office buildings are fully utilized again (if ever). After all, who exactly are you getting on the elevator with? That means we will be redesigning the physical nature of work for the near to mid-term future.
  12.  

  13. We already receive tons of data from employees through our networks… keystrokes, browser behavior, work documents, communications, video streams, collaboration. Expect intelligent tools to begin to understand both their content and structure. We will redesign work based on data. We will be using intelligent tools to manipulate and understand that data.
  14.  

 
How can HR professionals successfully integrate machine intelligence with their workforce?
 

  1. The most important thing is safety. The pandemic teaches us that we need to be safer in very fundamental ways: wear a mask, distance from others, wash your hands, don’t touch your face, and monitor yourself and others for symptoms. This is stressful at levels we are just beginning to understand. AI will teach us about both physical and emotional safety.
  2.  

  3. The next important thing is numeracy, the ability to be comfortable with math and statistics. The pandemic teaches us about sample size, measurement, and multivariate forecasting. These are baseline competencies for a workforce that uses intelligent tools well.
  4.  

  5. As it has been with the pandemic, we will be learning as we go with intelligent tools. What we thought was right last week (or even yesterday) usually turns out to only be a part of the whole answer. Being comfortable that we are continuously learning is a skill that can be taught.
  6.  

  7. Certainty, as in ‘we have to do it this way because the machine says so,’ is a problem to avoid. All decisions are made at a point in time with a given set of data. Nothing is certain. But, by keeping close tabs on your expectations and assumptions, you can make good decisions and keep learning.
  8.  

  9. People will trust the output of a machine before the machine has earned that trust, simply because it’s a machine. They will also lose confidence when the machine seems to make a mistake. Neither point of view is useful. Machines are like novice interns. They do what they’re told, only stop when you tell them, and can’t innovate.
  10.  

  11. Basically, introduce intelligent tools to the workforce with the reminder that the machines are not the human. It is the human who is responsible and accountable.
  12.  

 
You will be sharing details about an ‘AI Safety’ statement at HR Tech Festival 2020. Can you elaborate on this and how it impacts both employers and employees?
 

  1. The essence of AI Safety is the anticipation and monitoring of unintended consequences. The key here is to make sure that there is a team of people who are thinking out worst case scenarios and monitoring outliers. This might have been called an ‘ethics committee’ in earlier times.
  2.  

  3. The most important first step is to have a current inventory of all of the intelligent tools in your systems.
  4.  

  5. Most data models age as their underlying assumptions become invalidated by the data they process. Clear monitoring of performance degradation should be an essential part of every single instance of intelligence in your system.
  6.  

  7. The health of the underlying data is important to monitor. Imagine, for example, that one of your intelligent tools estimates the risk that a person will leave their job. The models are built on historical data from an extraordinarily good economic climate with serious talent shortages. That data won’t be inherently useful in a downturn with high unemployment.
  8.  

  9. Take the machine’s opinion with the same grain of salt you’d take any other opinion. Verify that the opinion makes sense in the broader context of your experience, strategy, and goals. Explore new issues that arise.
  10.  

  11. Intelligent tools have the net effect of moving decision making closer to the places that produce value. Safely managing the workforce’s newfound responsibilities involves patience and understanding.
  12.  

  13. Understand that sometimes the entire data set has to be trashed. Once the model is completely out of synch with reality, you have to start over. One replaces models in intelligent tools the way one replaces tires on a car. Get new ones. Don’t patch the old ones.
  14.  

  15. The most important thing is monitoring and governance. Every organization should have a team who monitor, assess, and question intelligent tools deployments. Because we learn so much, this has to be an ongoing function with rotating membership. Rotating membership helps keep the group’s perspective fresh.

 
What are the other key takeaways delegates can expect from your presentations at HR Tech Festival 2020?
 

  1. The basics of preparing for AI in your workforce.
  2.  

  3. How organizations are using intelligent tools in a pandemic time.
  4.  

  5. The fundamental elements of an HR AI Ethics program.
  6.  

  7. A renewed understanding of the range of possibilities offered by intelligent tools
  8.  

  9. An opportunity to rethink the way that HR is executed.

 
Can you also share your thoughts with us on the concept of remote working, and do you see this shaping the future of work? Within that sphere, what is the role you see AI and other intelligent tools playing?
 

  1. The idea that work has to be centralized in order to be effective is built on a factory model. When we were building things, we needed to all be in the same place. If you work in a company that builds, sells, distributes, or stores things, you probably still need a centralized workforce. For everyone else (maybe 30%), a distributed workforce is a better idea.
  2.  

  3. The Work from Home model (WFH) makes some pretty bold assumptions about employees. For instance, the idea that the company needn’t pay rent, utilities, and so on is suspect. Fairly quickly, we’ll see this simple idea evolve into a more complex and business-like relationship.
  4.  

  5. There are all sorts of things to be sorted out about the WFH arrangement
    a. When should employees meet in person? Should the supervisor visit the home worker at home?
    b. How do you replace water-cooler socialization?
    c. Is it reasonable to think that a culture can be extended without physical proximity? That’s a big bet.
    d. What is a reasonable level of employee monitoring. It’s reasonable to suspect that managers will want to compensate for the time they aren’t spending in the presence of their team.
  6.  

  7. AI will be a part of all sorts of aspects of distributed work from performance monitoring and appraisal to automated time keeping to work/project management. In order to manage a distributed team, managers will need the force multiplier of intelligent tools.

 


 

Transcript of talk

 
John Sumser:

[00:00:00] My name’s John Sumser and I’m here today to talk about intelligent tools and where we are right now. It’s been a disturbing time. And you might guess that things ground to a halt in the world of AI and data and people analytics, and you couldn’t be further than the truth. So I’m here to tell you about that.
 
 
[00:00:20] When I use the term intelligent tools I’m referring to all of the things that people have called AI and predictive analytics and machine learning and natural language processing. I don’t really think there’s much intelligence involved just yet. And we’ll see some of that over the course of this presentation. So when I say intelligent tools, I’m referring to that basket of stuff that is machine intelligence being applied to the work of HR.
 
 
[00:00:50] Again, intelligent tools are everything that has gone under the rubric of artificial intelligence, machine language, machine learning, natural language processing, deep learning, predictive analytics.
 
 
[00:01:06] Okay. So I run a company called HRExaminer.com. It’s a publishing house that does industry analysis as its fundamental business.
 
 
[00:01:17] There’s a website, there’s a weekly newsletter. We have about 40,000 subscribers. Most of them in senior management around the world. And we look at the edges of HR rather than current conventional practice. There’s there are other people who do a great job of looking at what the benchmarks are today. Our job at the HR Examiner is to look over the edge to see what’s next.
 
 
[00:01:42] So weekly newsletters, we do an annual report about, the state of intelligent tools. This year will be our fourth year of doing that annual report. And we have podcasts on Thursdays and Fridays. There’s Thursday show is a wrap up of the week’s news in HR technology. The Friday show is interviews with key executives.
 
 
[00:02:09] We’re in a credibly uncertain time add in times of uncertainty there’s only one real way to get ahead, which is embrace the fact that you don’t have the answers. For my money, I get very, very worried when I run into people, who have the answers right now, because we don’t know really what we’re dealing with.
 
 
[00:02:32] And so I say, embrace the uncertainty, figure out what the very next thing to do is, and do it, and then repeat that process. And then as you’ll see in the presentation, I also think you need to look out a little bit beyond the next thing to do. But don’t expect to have a big picture or a grand idea of how things are going to be before you get started.
 
 
[00:02:55] Things are going to be very ambiguous and very uncertain, for a very long time. Whenever I do a presentation, I like to set the expectations early on. Here are the things you’re going to take away from this talk. One, machines now have opinions. It used to be the case. The data that you put in was data that you got out, but today machines take that data and they convert them into ideas about what’s going to happen.
 
 
[00:03:24] What’s going to happen next, what should happen? And so the proper way to treat this output of the machines is not to this factual data, but that it’s one opinion that fits into your overall decision making. People are using the crisis. This crisis that we’re in to increase the quality of their work by incorporating intelligent tools.
 
 
[00:03:49] I’ll give you a number of examples and some companies to pay attention to. One thing is very clear. Our organizations are going to have to be very fluid and very resilient for the foreseeable future. That’s all decent is not being great at firefighting. If you want to put your company out of business, let everybody go fight fires as they come up.
 
 
[00:04:13] Okay. Keep don’t pay attention to where you’re going. The foundation has to be great data, and now is a time to get your data in order. This is the time to really dig in and solve the data problems in your shop. For me, that’s acts as the art of continuously asking why and what will happen if that’s X is concerned with the best outcomes for people, employees, and other stakeholders, the company itself, and the surrounding community, the crisis gets us in the practice of slowing down our decisions until we really tried to figure it out.
 
 
[00:04:46] All of the unintended contract, right. And then finally, I hope you’ll take away a better understanding of how to make decisions that don’t fire on you as you move down the road.
 
 
[00:05:01] One of my favorite ways of sinking about leadership has always been that it is like being at the head of Alliance cars, driving across bridge on a very foggy, like. You can’t see the sides of the bridge. You can’t see the end of the bridge. You can see a little bit pastor headlights in front of you. It looks like that picture of the top left.
 
 
[00:05:25] It’s very challenging to drive in the fog. And that’s often what leadership is like today. That problem of driving in the fog. Flows all the way down the organization. And so, so as managers in organizations, following an inspired executive, we’re used to having tail lights, still follow, and there are not so many taillights to follow.
 
 
[00:05:51] Now we’re all driving in the fog and great leadership is understanding that we’re driving in the fog, driving in an appropriate speed because we’re driving in the fog. And trying to see a little bit further out in front of where your headlights will take you so that you can make sure that the next step gets you lined up for the right outcome comes down the road, get used to having no big picture for awhile.
 
 
[00:06:19] It’s not coming back anytime. Soon, as I mentioned, I’d rather not call these emerging tools that are starting to be everywhere. Artificial intelligence. They’re mostly extraordinary and useful applications of math to big piles of data it’s useful, but it’s not very intelligent in a couple of minutes.
 
 
[00:06:42] I’ll talk to you about what’s happened to our business piles of data and why? Yeah. Thinking of this as an intelligent process is kind of dangerous. Nonetheless, these tools are being baked into human capital management systems everywhere you look. Workday for example has 25 discrete machine learning functions in its talent management system alone.
 
 
[00:07:06] Ultimate software deployed its entire workflow around intelligent tools and coaching what they did ultimate software it’s like they were going to put a new foundation under the house. So they lifted the whole house off and they installed. Intelligent tools as the foundation of their overall software package.
 
 
[00:07:26] This is an amazing job machine learning is a technique for examining data to discover repeated patterns. It’s really great at chess. It’s really great that the game of go, it’s not so good in driverless cars in cities. I’m sure you’ve heard some of those stories where. Where a driverless car that does just fine out on the highway, goes into the city.
 
 
[00:07:55] And there are a thousand things that you can’t predict every moment when you heard the shit and you can’t predict it, the ball that comes bouncing out of nowhere, the cat that runs across the street or the way that people are double parked in the middle of a skinny little alley. And so machine learning is not solving the problem for.
 
 
[00:08:15] Car manufacturers in the way that they thought and are estimates of when we’re going to see fully autonomous vehicles, it’s moving further and further down the road. So it’s a technique for discovering repeated patterns. It works in things with fixed rules and finite settings. Yeah, it doesn’t work so well.
 
 
[00:08:35] It’s complicated places where you have a merchant phenomena and emergent phenomena. Are those things that happen every day in organization where something that you completely didn’t expect, changes, everything sound familiar, natural language processing is a way of kerning texts into math wages and then doing mathy kinds of stuff with it.
 
 
[00:08:57] It’s great for building search engines. It’s great for seeing patterns across resumes. It’s how you distill skills out of large quantities of resumes and job descriptions. Very very useful. The whole idea is that if you find a repeated pattern, you can use that repeated pattern to make it prediction. All of that is based on stable, repeated history.
 
 
[00:09:24] More about that in a moment, the most important thing you could take away from this talk is that machines can offer a remarkable opinion about what’s going on, but there will always be just that and opinion if offers a point of view. Okay. I don’t know if you’ve noticed, but where I live, everybody has left the office building and moved into their house just to do work.
 
 
[00:09:50] And when that happened, all sorts of historical predictability has vanished things that looks like falling off a log to make quarterly numbers are not happening. People who used to work very carefully together or not able to do it. Yeah. Some of the things that, that are broken, our understanding of attrition no longer really works because there’s this variable of people getting really sick and dying from something that wasn’t part of our mapping of nutrition in the past, because of the various health concerns running around the system, our benefits
demands are changing.
 
 
[00:10:30] The requirements for paid time off are different. If you’re working out of your house, Who makes a good part of a team and how many people you need and what backup looks like for a team to work and solve a specific problem. That’s all true. Contingency plans change because the contingencies are there even things as simple as bereavement leave no longer mean the same thing.
 
 
[00:10:54] Workflow timing is different when you have to schedule every interaction in the zoom meeting assumptions about privacy teams. Communication flows, change message. Control changes. Call center volume is okay. Particularly of HR job descriptions don’t mean the same thing. Any longer management competencies are different.
 
 
[00:11:18] When you manage a distributed organization, operational analytics don’t have any validity over time. Right now, the volume of job applications is skyrocketing. Hiring pace is slowing. The definition of who’s a social and who’s not essential is under fundamentally examination. And what are the engagement score means is up for grabs.
 
 
[00:11:42] And so these various central components of the HR workload, Oh, we’re all subject to question right now. And the historical data has really completely changed. That means that. Many of the embedded machine learning tools that you find in your software already are going to produce more questionable results than they did in the past.
 
 
[00:12:07] And you’re gonna have put more energy and do arguing with the machines. A conclusion nothing’s predictable right now. Uh, nothing, nothing at all right now, except . Yeah, tomorrow morning and the sun will come out. And it is a remarkable opportunity to them build a clean slate in the world of your data and its structure in the HR department.
 
 
[00:12:36] It’s a remarkable opportunity to embed the architecture necessary to build for the future, a fully data driven decision making process. The first step is to get your data in order. And I’ll talk about that in a second. You do this whole process one step at a time, but I want to underline something that I’ve been saying, which is that you can’t let solving the prices, get in the way of coherent decision making for the long haul.
 
 
[00:13:07] Even if you can’t see where the long haul takes you and you do that one step at a time, you are, you get very wary of short term fixes and bandaids. You designed with ethics in mind. So now perfectly wonderful time to start thinking about the ethical consequences of your decisions and you learn how to argue with machines.
 
 
[00:13:31] So let’s look at a couple of those things. First of all, getting your data in order means having complete profiles for all employees, for the most part. HRS systems have about 25% completion rates for the personnel profiles inside of their systems. But if you’re going to have yeah. Agile fluid, resilient organization, that’s able to shift based on the circumstances that come its way.
 
 
[00:14:05] You need to know as much as you can about everybody who works for you. So complete profiles for all employees is so important. That you might consider withholding paychecks until people yes. But get there profiles filled out and then you need to establish goals for prediction and tracking of work and work.
 
 
[00:14:25] Wow. There’s going to be an extraordinary set of temptations to track everything because that’s a way of coping with uncertainty and it’s more important to track a few things. Well, then everything all at once. The other thing that’s really worth understanding is that you should be building employee like profiles for the people who are in your recruiting pipeline.
 
 
[00:14:53] We’re in uncharted territory, where there is reason to, I believe that 20% of your workforce is going to be okay. Four, nine to 12 weeks with the disease and recovering. Another three to 5% of the workforce is going to die. People are going to need all sorts of extra bereavement leave because they are, uh, taking care of loved ones who have recently passed.
 
 
[00:15:22] And so you’re going to draw on your recruiting pipeline in ways that you have in the past as part of succession parenting for frontline jobs. And then finally. You want to have a company data model so that you understand how all of the data goes together. And this is a nontrivial thing to do. Building a company data model.
 
 
[00:15:44] It takes a team and it takes some time.
 
 
[00:15:50] No, we just talked about are there’s a hiccup in the quality of the output of machines. That’s the way that it was up til the 1st of February is not the way that it is now. But even before then there were problems with the machine output and they come from the timeliness of the data or latency, the sufficiency of the data, the completeness of the data, sometimes machines are just playing wrong because they can’t see certain things.
 
 
[00:16:20] They don’t have a conscience. They don’t have common sense. They don’t have intuition. All data and data quality given their way all the time. And you have to be wary of these things. When the machine offers you a recommendation, how do you do this? How do you argue with the machine? Well, when do you get a recommendation from a machine, regardless of the area of the recommendation, you want to sit back and look at it and you want to try to understand.
 
 
[00:16:51] What’s the date of the machine used to make this conclusion and what you draw the conclusion, the same way, this business of arguing with the machine, it’s going to get harder and more important as time goes on. And we depend the and more on machines. Yes. So what’s working, what’s working out there. I’ll tell you the volume of queries into the HR department.
 
 
[00:17:18] It’s exploding to the ceiling, the volume of applications being sent because of, you know, the employment rates is exploding. Meanwhile, search is in terrible shape because it’s HR and search is better than terrible shape. So the things that work, reduce question volume and hurry up the response, they improved the quality of self service results.
 
 
[00:17:42] They expand the effectiveness of search. They helped me deliver better quality experiences, applicants who are in your pipeline. They allow you to see further ahead. Yeah. Improve communications, quality and consistency repair. And from employee trust increased internal mobility effectiveness. And this is a key one.
 
 
[00:18:08] You’re going to be moving people around the organization. Because of the impact of the disease on your staffing. And so you want to be able to understand who would be good, where, and be able to move them quickly. And it needed to be able to account for the fact that you could lose as many as 20% of the people at any given time.
 
 
[00:18:28] And then the last thing is tools work is that re-imagined succession as business continuity planning. So it’s not just succession at the top of the organization. What is success in the places where you have individuals who are the only people who know how to do this specific? The first area is there are a lot of really interesting things going on in conversation.
 
 
[00:18:56] Socrates, who are one of my favorite little companies does systems integration with conversations. And so their idea is that any employee ought to be able to ask. Uh, the machine, that coherent question about something white, how many days off would I get to have my new baby? And currently that takes an HR professional looking things up on six or seven different systems to be able to formulate an answer.
 
 
[00:19:24] And you can answer questions that complex with Socrates. They currently. Help you get your policies in shape and get the automation in shape so that the machine to answer about 9,000 questions, professional depth is another thing. If you’re dealing in an environment with rigorous regulation that the HR folks are following capacity helps build tools that dig deep into the regulations to find out answers.
 
 
[00:19:53] Candidate experience company is like. Textio or Saddam people offer really interesting ways to ensure the quality of the cow’s experience over time. And they are typically moving into internal mobility, uh, as an extension of the candidate experience for domino novella is an interesting company that does project performance.
 
 
[00:20:17] And that means. They’ve got a project management framework, critical paths thinking in large complicated projects with hundreds of tasks, but performance management and task communication happens inside of that framework with a constant eyes to the consequences of decisions on the final outcome. And then Textio does amazing stuff with communications effectiveness, their tool.
 
 
[00:20:44] Takes the language that you put in your documents and analyzes it to see if it’s going to have the intended impact. The audio search is a super bugaboo. Yeah. The human capital management systems, while everybody was getting Googled in the consumer world, the. HCM world has not been quite so fortunate with its search tool.
 
 
[00:21:11] So it’s often the case that the data is in the system, but you can’t find it. There are companies like hiring solved and Socrates who are working on different aspects of search, hiring solved, solves the search problem and the overall talent acquisition function. So all of the databases under one roof with one search engine using hiring solved.
 
 
[00:21:36] And then Socrates solves search on a policy and a employee data respect. Yep.
 
 
[00:21:46] And this is one of my favorite areas, sediment, attention monitoring, you know, the, um, the engagement score it’s kind of broken right now in America. 84% of employees are looking for. Uh, uh, looking for new work, they’re scared of having their job taken away from them. And so the idea that if you ask them a question about their engagement in the organization, they’re going to give you a straight for the answer.
 
 
[00:22:13] It is a little fanciful. What you need to do with engagement scores is balanced them with some sort of validation. And there are two basic approaches to doing that. One is sentiment analysis, which means. Taking free text responses to surveys and jelling a S a single thread of sentiment out of that stuff.
 
 
[00:22:38] And that’s done very well by ultimate software and then tension analysis, which is done by keen Corp. Um, is the idea of that when something’s up in the department or division the language in that department or division changes. And you can measure the change of the overall language in all of the communications by monitoring the communications and score it, catching it.
 
 
[00:23:04] This is great for spotting trouble spots. It’s also great for you finding spots where, uh, things that are really positive are happening. I call the next class people movers and they are, um, talent acquisition in orientation. Arena.io is an AI matching system hiring solve to does scoring and matching of databases across the system for non people has incredible personalization from the point of applications through internal mobility, add work logic.
 
 
[00:23:45] Does AI driven. RPO sorts of work, all worth paying attention to all being used to great effect. Um, in the current environment,
 
 
[00:23:59] ethics involves examining decisions and policies for the direct, indirect and unintended consequences on employees, stakeholders, the entire ecosystem of the organization as a whole. It’s particularly important to think about ethics in a crisis time when people are making decisions in a real hurry, ethics doesn’t mean a set of rules.
 
 
[00:24:22] This is not, don’t do this. And don’t do that. This is more a slowing down of the decision making process to make sure that we see where the likely fallout it’s going to be for a solution that we need to make to something right away. The basic areas of ethics, um, our bias, privacy, total cost of ownership.
 
 
[00:24:49] And that might surprise you as a, as an ethics problem. Bateman’s reproducibility liability. So let me walk through those really quickly. It’s not possible to eliminate bias from decision making and it probably isn’t even very desirable. But it is really critical to be aware of bias in decision making.
 
 
[00:25:11] And so ethics highlights, bias and allows you to see whether or not bias is leading in the right direction. Privacy is both a legal and a trust question. Um, employees are in a new and precarious position and the regulations that protect their privacy. Are under some pressure to make the workplace more accessible and more healthy.
 
 
[00:25:39] And so you’re going to see a lot of tradeoffs between safety and privacy and it’s uncharted territory and marriage consideration. One of the things that I often recommend with intelligent tools is that you don’t unleash an algorithm in your organization without a legal review. You want to understand what’s going on?
 
 
[00:26:05] When we make decisions particularly about the application of intelligence tools, the total cost of the decision that we’re making needs to be an upfront piece of it. So it’s, it’s, it’s common in times of crisis to make decisions that look cheap in the short run and are very expensive in the long run.
 
 
[00:26:26] And what we want to do going forward is make sure that our organizations are. Not encumbered by debt because we’ve yeah. Some decision making down the road, same thing is true with maintenance, quick decisions, quick installations, quick applications of intelligent tools, create maintenance type errors.
 
 
[00:26:48] And so maintenance is sort of a subset of the total cost argument. And what you want to do is keep an eye on maintenance because maintenance is what allows you. To have crisp output from your systems and get good decision makers, any intelligent tool needs to be able to deliver reproducible results of this is a little bit of a challenge for these tools.
 
 
[00:27:14] And finally you need a heavy, heavy, heavy layer of consideration about what the liability is and the decisions that we’re making. Right? No. How would I describe all of this? These are actually the necessary elements, problem solving in our contemporary environment. And so for real HR should be using ethics as it problem solving tool rather than a barrier.
 
 
[00:27:46] And so, I’ve set these expectations. I hope you feel like I met them. I wanted to be sure that you walked away with the idea that machines have opinions. That there are tools that can help reduce the volume of calls, increase the consistency of answers, manage the flow of applicants. And you can install those things now while you’re solving problems.
 
 
[00:28:11] But I also want to underline that this is a really important time to get your data in order. We just stepped through the idea that ethics is a way of solving problems. And in total, this view that I’ve been giving you should give you a framework for making decisions that allow you to land well with better data and more intelligent tools, reducing the overall manpower of your HR department.
 
 
[00:28:40] Thanks. Again, my name is John Sumser and I’m the principal analyst at HR Examiner. You can get in touch with me at John at HRExaminer.com or follow me on Twitter at @johnsumser.
 
 
[00:28:53] Thank you very much. And, just to leave you with a little something, here are the company’s we discussed.

 

Watch more of John’s talks on his YouTube Channel »