AI adoption is accelerating, but for most organisations it is still ad hoc. Tools are bought reactively, budgets keep growing, and few leaders can say with confidence what is being used, what is working, or what return they are getting. Some companies are already going further, cutting roles on the assumption that machines can simply take over tasks from people. What rarely gets examined is what those decisions do to the people who remain, the culture they work in, and the customers they serve.
So what do organisations need to understand about how work happens before they let AI reshape it?
My guest this week is Sam Naficy, Chairman and CEO of Prodoscore, a data analytics company that studies employee engagement, productivity, and collaboration in large enterprises. In our conversation, Sam shares what his data reveals about AI use and productivity, the cultural consequences of automating roles, and why visibility into work has to come before automation.
In the interview, we discuss:
- Why most organisations are still in the early stages of AI adoption
- The growing gap between AI spending and visibility into what is being used
- What the data shows about heavy AI users and productivity
- Why some employees adopt AI faster than others
- The unforeseen impact of automation on workplace culture
- What happens to the teams that remain when their AI replaces their colleagues
- Stress-testing workforce changes before making them
- The line between visibility and surveillance
- What does the future of work look like?
Follow this podcast on Apple Podcasts.
Follow this podcast on Spotify.
Transcript
Matt Alder 0:00
Many organisations are investing heavily in AI, with some already cutting jobs on the strength of what it could do within their company. Yet very few can actually accurately describe how the work actually happens inside their own businesses. So what are they missing? And what is it costing them? Keep listening to find out.
Advert 0:23
Support for this podcast comes from Cyberr, the professional network for cyber security. If you’re hiring for cybersecurity roles, you’ll know that they are more complicated and time-consuming than more generic requisitions. The briefs are full of acronyms. Filtering through candidate pools is incredibly complex, and outbound response rates are a fraction of what you see elsewhere. Specialist agencies will take away the burden, but at a high cost. And every week a role stays open, your organisation is at heightened risk. Cyberr Recruiter is built to close that gap, saving you time, effort and money in the process. Every Cyberr Member profile carries a trust score. Certifications are validated through Credly and identity is verified at sign up. Through their custom search agent, CIRA, your team brief in a role in plain language and get a qualified shortlist within minutes. You can find out more by going to cyberr.ai/hire. And Cyberr is spelt C-Y-B-E-R-R. That’s cyberr.ai/hire.
Matt Alder 1:50
Hi there, welcome to episode 814 of Recruiting Future with me, Matt Alder. AI adoption is accelerating, but for most organisations it’s still on a very ad hoc basis. And few leaders can say with confidence what’s being used, what’s working, or what ROI they’re getting. Some companies are already going further, cutting roles on the assumption that machines can simply take over tasks from people. What rarely gets examined is what those decisions do to the people who remain, the culture they work in, and the customers they serve. So what do organisations need to understand about how work happens before they let AI reshape it? My guest this week is Sam Naficy, Chairman and CEO of Prodoscore, a data analytics company that studies employee engagement, productivity, and collaboration in large enterprises. In our conversation, Sam shares what his data reveals about AI use and productivity, the cultural consequences of automating roles, and why visibility into work has to come before automation. Hi, Sam, and welcome to the podcast.
Sam Naficy 3:03
Good morning, Matt. Happy to be here. Thanks for having me.
Matt Alder 3:06
A pleasure to have you on the show. Please, could you introduce yourself and tell everyone what you do?
Sam Naficy 3:11
Great. My name is Sam Naficy. I’m the CEO of a company called Prodoscore.
Matt Alder 3:19
Fantastic. And tell us a little bit about Prodoscore.
Sam Naficy 3:19
At a high level, Matt, it’s a data analytics company that looks at employee engagement, retention, collaboration, and kind of productivity as a whole, using AI and algorithms to kind of give assessments and outcomes for large enterprise organisations.
Matt Alder 3:35
Fantastic. And you’ve got a huge amount of really interesting data that we’re going to dive into. But to start with, just a kind of a bit of a general question. In terms of AI, from what you’re seeing, how are organisations actually adopting it? Because there’s a huge amount of hype out there. What are you kind of seeing as the reality?
Sam Naficy 3:55
I think the reality, we’re probably using a baseball analogy in the second or third inning of AI adoption, migration. And we believe with a lot of data that I’ll share momentarily, it’s been more ad hoc. It’s been more by the seat of our pants, reactionary, and let’s adopt AI, let’s get all the tools that we’re fully familiar with, ChatGPT and Claude and Gemini and all the tools without really looking at the impact on the organisation, on outcomes, and on customer relationships. So we think that that’s a big area of focus as we move up from the third inning of the baseball analogy to the sixth and seventh and more maturation of this kind of movement.
Matt Alder 4:39
Absolutely. And there’s obviously a lot of money being spent on these tools. And I was talking to a CFO the other week, and they said the biggest challenge they have around AI is almost software creep in terms of buying more and more sort of solutions and AI enabled tools that might sort of overlap with each other. So a lot of money being spent. What kind of visibility do companies actually have on what’s being used and whether it’s working?
Sam Naficy 5:05
Great question. One of the many components of our software in Prodoscore is giving that visibility. And I’ll give you a similar example to what you gave with the CFO conversation you had. I was speaking to a CEO of about a 4,500 employee organisation literally last week, and they had adopted licences for Claude and spent hundreds of thousands of dollars, I don’t remember the number, but a significant amount of money on those licences. And realised quickly using our tool, Prodoscore, was that so many of the employees were using non-Claude AI tools, primarily ChatGPT. And the assessment was, well, why am I paying for all the Claude licences if the rank and file are using ChatGPT. Let’s look at this. Is it an adoption issue? Is it a training issue? Are they more comfortable with ChatGPT because it was more of a legacy product? So that kind of nuance and visibility is what we’re trying to provide. There’s a lot of money being spent around AI tools, like you mentioned. What are we getting out of it? What’s the return on that investment?
Matt Alder 6:17
And what else are you sort of seeing from the AI use that you’re tracking?
Sam Naficy 6:17
Okay. There’s no doubt we see productivity increase. We have data that we released a couple of weeks ago that kind of medium users, moderate users of AI see a productivity jump of 15, 20%, but heavy users are in the mid 30% range of increase in productivity versus their peers. So what we do, Matt, is we look at and assess employees within the same peer group. So basically identical roles and looking at their flow of activity during the day. And we see, as I said, for heavy users of AI tools, productivity increase in the mid 30% range.
Matt Alder 6:58
Interesting. And do you know what it is that makes people heavy users? What’s going on with the adoption there? Why are some people adopting it? Other people aren’t.
Sam Naficy 7:07
I think it’s inherent to new tools, right? As human beings, you know, I’m probably the least user of AI within our organisation. For example, I’m also the oldest person here. I’m the guy with no hair and white on the sides. So I think it’s part of that. Part of it is training. Part of it is reluctance. Part of it is getting stuck in our old ways. And I’m giving myself as an example, as I’ve picked up adoption, as my team has kind of got me on board of Claude and using it and improving, it’s fascinating of how it’s made me productive. And it’s kind of organised my life and my different areas of my day, looking at my workflows. One of the things we do, Matt, is we look at workflows and optimise the workflow during the average day of an average employee.
Matt Alder 7:55
Interesting. I suppose on that point, what impact is AI having there, but particularly on things like the culture and how employees communicate with each other when the work is starting to become automated?
Sam Naficy 8:09
That to me has become my big, big mantra in the last two, three months. And the impact of culture was not foreseen by any of the organisations we talked to. And you hear it out there in Meta’s press release and Salesforce and Microsoft. So here’s the analogy. We adopt AI and because of it, we could potentially use a tool like Prodoscore to look at workflow automation, workflow optimisation, and we get rid of X number of employees because of chatbots and agents are able to resolve those issues. What happens to the culture for the remaining employees in the organisation when that RIF, reduction in staff, reduction in force happens. It was a completely unforeseen outcome and byproduct of the AI migration. So we’ve been the big advocates of let’s assess that before you pull the trigger because it leads to an incredible erosion of culture, disappointment, and what’s the outcome look like for the end user clients? They never looked at that issue. Let’s get rid of X number of employees, AI can do the work, resolve the tickets, customer service, whatever the areas were that AI quickly became adopted. Unforeseen was the circumstances of the outcomes of what the customer experience looked like. What did the remaining employees in the organisation that communicated with their colleagues that are no longer there?
Matt Alder 9:39
I think that’s really interesting because there are some kind of fundamental mistakes being made at the moment by almost just assuming that humans and machines are the same, that actually, you know, you can break a workflow down into tasks and a machine can do those tasks. Therefore, a machine should replace the human. But as you say, there’s so much more to it, isn’t there, from the human aspect of things?
Sam Naficy 10:02
Completely. Here, I give this example often, but imagine a sales department and a customer service department that historically were very enmeshed. There was the pre-revenue, the revenue, and then the customer success would take it from cradle to grave in the history of that relationship with that client. They communicated all the time. So imagine for an organisation that may have reduced its customer success team because of the repetitive functions there are now using AI for onboarding, training, QBRs, all the stuff that customer success does. Now, historically, the sales team would be communicating with their colleagues on the CS team. What does the sales team do now when the colleagues are no longer there on CS, but it’s a chatbot? Do I write an email to the chat? Do I Slack it? What does it respond to me? Do I go have coffee with it? It’s a whole nuance that it’s unknown. And that’s where I touch about the culture issue because you have the remaining people in the sales organisation, in my analogy, still with the company, while customer success may have been reduced to a fraction of its former employee count.
Matt Alder 11:17
And has your sort of data told you anything else about how work’s changing, how people communicate? Are there some sort of surprising findings in there?
Sam Naficy 11:27
Yeah, look, we’re all about outcomes and both outcomes internally for the organisation, and then ultimately externally in the customer relationship of that company with its clients. And you’ve seen this public information. You know, people have let go of thousands of employees only quickly to rehire back some of them, right. Our thesis has been, before you have that reduction of force and migration to AI, let’s stress test it. Let’s look at the data around it and ensure that your employees are kind of sound and healthy that are remaining. And most importantly, the customer outcomes and the customer success remains as it was or if not improved. Because then it’s going to be really, really disadvantageous to the company to migrate to AI when you have potential churn or attrition of your customers.
Matt Alder 12:22
I think an interesting point here as well, because there is a great deal of fear in the workforce around things that could be seen as surveillance. We’ve seen a very well-publicised case of Meta using people to train AI to do their jobs and then laying them off. With software like yours, when you’re sort of collecting that detailed data on how people work, where’s the line between sort of visibility and overstepping into surveillance for you?
Sam Naficy 12:53
Great question. And that’s been the number one question asked of us in the five years of our history. A lot less now, but certainly in the beginning, the pronounced question was surveillance, big brother, you’re micromanaging people. We don’t do any of that, Matt. And the distinguishing difference between us and some of the tools that do that is we don’t look at your mouse clicks and URLs and keystrokes at all. But more importantly, is the employees themselves have their own dashboard within Prodoscore. So nothing is hidden from them. So when I come in as an employee, if I’m a Prodoscore customer, I come in, I have my dashboard. I see what Sam is doing. And I see what my peers in the identical role are doing. So we give full visibility to the employee, to middle management, and to the C-suite.
Matt Alder 13:48
I think that’s interesting. Because that must be interesting from an onboarding perspective as well. When people join the organisations, they can actually see the shape in which people work and how they use the software.
Sam Naficy 13:57
Completely. Here’s a perfect example that we had a company whose onboarding would take up to 90 days. They deployed Prodoscore. They looked at what a good onboarding and a good employee onboarding outcome looked like. Post-deploying Prodoscore, they reduced their onboarding time to 45 days. So we showed what a good employee was and how to replicate it. And any employee that’s coming in and it’s good and wants to thrive and succeed loves a tool like ours. By the way, I give this analogy a lot. 20 years ago, salespeople detested Salesforce. Like, oh my God, you’re looking at my pipeline. You’re looking at my call volume. Leave me alone. I don’t want to get involved in Salesforce. Obviously, now it’s ubiquitous and salespeople can’t live without it. So we feel a tool like ours, if it’s properly deployed and adopted, it allows the flexibility for the employee to potentially be remote, be a hybrid employee of three, two, four, one days in the office versus remote, and give the employer the engagement visibility they want to provide the flexibility to the employee. So it could be a win-win if used properly.
Matt Alder 15:14
As a final question, if we sort of look out into the future, where is this going? What does the future of work look like, particularly as AI gets even more embedded into organisations?
Sam Naficy 15:27
That’s a great question. I think of that 24/7, Matt. And we believe, partially selfishly, is that for any of this AI strategy to work, you need activity-level visibility. For AI to work properly, you need granular activity level work and workflows to be able to optimise those. And tools like ours provide that visibility to management. Exactly what does Sam do in his eight to five, eight to seven, whatever my duration of my workday, what do I do exactly during the sequence of my day? And by seeing the activity level at that granular level, you’re able to show it, replicate for other employees within that group and cohort. And then if you want to have AI come in, that’s the intelligent way to deploy AI when we fully assess the workflow and then we optimise it, right? The issue and hiccups have been people have tried to optimise and jump to AI without having visibility to workflows. The last thing I’ll say is this. I think this idea, this morbid world that we’re all going to be replaced with AI is BS. I truly don’t believe that extreme is going to happen. I think we’re going to push back and there’ll be a modicum of that middle of the road. We’re going to need people. People want to talk to people, engage with humans at some level. I’m not saying AI doesn’t have a big part in the future of work. It absolutely does. We’re seeing it. It’s done great for different roles in different organisations. But this idea of humans being obsolete and the obsolescence of human beings is way overly exaggerated in our belief.
Matt Alder 17:10
Sam, thank you very much for talking to me.
Sam Naficy 17:12
Absolutely, Matt. My pleasure.
Matt Alder 17:14
My thanks to Sam. You can follow this podcast on Apple Podcasts, on Spotify, or wherever you listen to your podcasts. You can search all the past episodes at recruitingfuture.com. On that site you can also subscribe to our weekly newsletter, Recruiting Future Feast, and get the inside track on everything that’s coming up on the show. Thanks very much for listening. I’ll be back next time and I hope you’ll join me.
