Leaders discuss automation, upskilling and the next era of support.
.png)
GMT20250409-160415_Recording_640x360 (1)
0:04
Good morning.
0:05
Welcome to the experience dialogue.
0:06
In this conversation we have guests with very varied experience and we are just talking like friends, interrupting each other or you know, even disagreeing with each other, which is excellent.
0:20
And today I have two amazing guests here with me.
0:24
Sarah, I spoke to you first, so I'm just introducing you first for no other reason, Sarah from Concentrix.
0:30
What was very interesting for me when I talked to Sarah is she actually has spent over 18 years in the BPO industry and she has come all along and she's one of the few women in service that I've seen who have grown completely in the service side and seen things grow.
0:49
So I would love to get her experience and her background behind what she has seen over multiple decades.
0:57
Welcome, Sarah.
0:58
Thank you.
0:59
Really glad to be here.
1:00
Lovely to talk to you, Kay.
1:03
Rohit.
1:04
You have spent a ton of time in the tech industry, mostly with infrastructure, platforms, very complex products, and you have also spent the time and Technical Support side of the realm and have grown very much into the Technical Support side of the realm.
1:21
You are with data bricks right now, which is kind of right in the middle of how data comes together and how AI is processed.
1:32
So I think, you know, getting your expertise and your changes over the decades and your background from Technical Support would be a very interesting and very excellent mix with Sarah.
1:43
So welcome, Rohit.
1:45
Oh, thank you so much, Gabe.
1:46
Glad to be here and nice to meet you as well, Sarah.
1:51
So we didn't really plan for this particular discussion other than just putting down some topics, right?
1:59
So forget about, you know, and you both don't have to answer every question.
2:04
So if there is something very salient that you want to bring it up, just bring it up.
2:08
OK.
2:08
So in terms of career experience, I know I had highlighted a few things here, but it would be great to understand, Rohit, from you, what you have seen as a change that has happened since you started your career to where it is right now in terms of your career as a support executive.
2:29
That's one we will, you know, let's start that first and then we'll talk about support in itself.
2:36
Yeah, to give you a little bit of context, like before I got into the workforce industry itself, even during my academics, I was always inclined towards AIML kind of stuff.
2:45
But the AIML then was never the AIML what we see today.
2:50
You know, the vision was so much more core to a group of people who are really great experts in in the AIML field.
2:58
Even the projects that I did were very focused on real time systems, embedded systems, machines, robots and stuff, you know, may solving robots that would solve mazes by themselves.
3:10
So ML and AI were very focused in certain aspects of, of, of the industry.
3:17
But as, as we, as I, as I've got into the workforce, I joined as a real time systems engineer.
3:23
I started out as a core developer, you know, coding in assembly, CC plus plus very systems level programming, moving on to more high level languages like C# and Java and other things.
3:34
But over time what I've realized is every every two to three years we were seeing a shift whether it's, you know, in the in terms of improved high level systems and customers themselves adopting a lot of things on their end and forcing vendors to, you know, keep upskilling themselves to provide, you know, better services.
3:57
So all I've always seen that the industry drives all product innovation, customers drive product innovation.
4:03
So that's something that I learned really early on.
4:06
And moving into the support organization eventually allowed me to also see the side of how customers expect support from these vendors.
4:15
So All in all, over time, you know, being part of companies like TCS in India, Tata Consultancy, moving to companies like Subex that's purely focused on telecom application development, then moving to Hortonworks, Hadoop, the big data revolution got the opportunity to even, you know, spend some time authoring a book around it.
4:36
There was, it's been a great right so far.
4:40
And now in data bricks, it's bringing all that experience together, you know, right from the day of coding a small robot to solve a maze to using AI to solve customer problems has been amazing.
4:52
So I, I, I feel that we are in a very, very great phase of customer support timeline.
5:00
Very true, very true.
5:02
Sarah, would you like to take your perspective on the question?
5:06
Yeah, I'm really interested in, in Rohit's comments because we're really both ends of the spectrum, right.
5:11
The, the experience that I have is kind of front of house, you know, with the direct customer engagement through BPO and Rohit, you're really in the, the machine room at the back, right, kind of driving the efficiencies.
5:23
I think in, you know, I've, I've been with Concentrix for nearly 20 years.
5:27
And I think that the biggest change that I have seen is that when I started my career, everything was about having enough people to manage the customer demand, right?
5:36
And it was a very easy calculation to say number of calls times time taken equals the number of people that you need, right?
5:43
And it meant that for our people, some of our jobs were really quite dull, right?
5:49
It was very repetitive.
5:51
It was very hard to keep people engaged.
5:53
It wasn't an aspirational career and for people, so as leaders that was quite tough.
5:59
But over time, you know, with the development of technology and with the input of automation, the real acceleration of that now, you know, with with the growth of bots and with the opportunity and AI, we have seen that our jobs have become so much more interesting.
6:15
A lot of the repetitive tasks and you know, a lot of those process driven steps that we would have dealt with with a human can now be dealt with much more efficiently.
6:25
Buy a machine.
6:26
And what that means for our people is that their jobs are actually a whole bunch more interesting because they're dealing with really complex customer demand.
6:36
They're they're dealing with people who have exhausted all possible opportunities for finding answers to their question or finding the right product for them.
6:43
And, and so the skill level of the people that work for us now is just a world away from the skill level of the people who worked with us 20 years ago.
6:53
And I think that's great for the industry.
6:55
You know, it really helps us from a, a motivational point of view because it really gives people that interest in their job when they come in every day and, you know, something, something that makes the brain tick and keeps people going.
7:09
So, you know, for me, the the BPO industry just gets more exciting and, you know, with every new development that comes along, Yeah, You know, it doesn't matter whether I'm talking to a BPO leader, Technical Support leader, field service leader, customer support, customer service, all sorts of flavours, right.
7:31
I'm hearing the same thing that the people's jobs are changing considerably.
7:36
One of the interesting perspectives for both of you is we do AI agent master class where people from all levels, mostly a agent master class so far has been executives so far.
7:49
And Carrier, we do a career roundtable where we have people from all levels coming in, right.
7:55
But the interesting thing we noticed is everybody, including the executive level are now saying that the jobs are more around coding and understanding the code, understanding more about software, understanding, you know, bringing in new skill sets and stuff into the mix.
8:15
And Rohit, you're not the only engine, you know, executive that I've seen who have started as an engineer, software engineer and who's, you know, leading support, right?
8:25
So there's a lot of people that I'm seeing that are coming in that way.
8:29
And do you see that trend also?
8:31
That's question number one.
8:32
That's the easiest part.
8:33
The second thing is, how are you preparing yourself for that change?
8:38
Yeah, I mean, this this whole revolution of AI agents and agency in general is, is going in a direction where everyone wants to optimize.
8:49
They they all want, you know, how can we save, save people save, you know, improve efficiency, improve productivity.
8:58
So save in any, any factor like whether time, whether it's money, it's always about how we can optimize.
9:04
So we are seeing that shift for sure in every, every step of the way.
9:08
I mean from a very simple pull request into, into a code, you want to automate the generation of the documentation of the pull request.
9:17
It's, it's as simple as, you know, a small change is going to create a big impact over time.
9:22
So at every possible opportunity that we we try to optimize, we are trying to embed things like AI or systems or tools that actually facilitate in that.
9:35
So to build out all this, there is a lot of development and coding that happens, but we are also leveraging AI to do that development and coding.
9:42
So we're seeing a lot of influence like systems like Cursor, you know, GitHub Copilot being like used day in day out to actually build out these tools to facilitate every step of the way.
9:55
And that's that's definitely happening.
9:58
I mean, every, every leader in every org is expecting, you know, that shift to happen.
10:03
That's correct, Rohit, it's very true.
10:05
You know, we are not going to escape AI.
10:07
It's going to happen especially, you know, I think we are all in that realm.
10:12
But I'm trying to think from your perspective of decades of experience, like what is it you are doing to prepare yourself, right?
10:19
So for example, we have customers with whom they want to, they don't, you know, just going and logging, bringing out the logs takes like 10-15 minutes.
10:29
We're automating that, right?
10:30
So that's why we are a full service ecosystem.
10:33
So everything from entitlement to bringing logs, to processing logs to even bringing down manuals from the various manufacturers, all of that is getting automated.
10:46
But how can I as a, as support executor, make myself ready for this change?
10:53
I think being bringing awareness, you know, you should have some kind of evangelism happening within your team.
10:59
You need at least like I always give this example of having one engineer dedicated to be an evangelist, focusing on trying to bring the latest and greatest break things like, you know, do lots of PO CS everything is not going to be a success.
11:12
So being ahead of the curve is key.
11:15
Things are changing day in day out, doing a lot of experiments internally, Many of those experiments are not going to be successful, but there's going to be a lot of learning in each of these experiments.
11:26
So something that we do internally is we, we try to, you know, keep ourselves aware of what's going to happen next.
11:34
And you know, you know, if, if you, for example, MCP servers are something that's, you know, picking up as, as, as something in the industry, it's, it's, if you really breakdown, it's not rocket science.
11:44
It's very, it's, it's a protocol.
11:45
It can be understood, but being ahead of the curve and understanding what in our ecosystem can really fit into that paradigm and how can we leverage that.
11:55
So being that evangelist, kind of having that evangelist kind of outlook is key.
12:02
Experimenting is key.
12:04
Letting things fail is absolutely OK.
12:07
You're going to learn a lot.
12:08
So that's something that we typically do and we are looking at it, you know, from the customer's perspective and from the support engineer's perspective, not just, you know, 1 angle.
12:18
What can we do to even avoid the origination of a support ticket?
12:22
What can we do from that point, right?
12:25
I always say no ticket, right?
12:26
So the ticket itself is a flawed concept.
12:29
If there was a customer sitting with us and they hear, they are hearing that it's ticket 1234 and they're like, what is it?
12:35
Does this all they refer to us?
12:37
So anyway, agreed Sarah, your perspective on it, please.
12:42
Yeah.
12:42
So I think two sides, 2 answers to that question.
12:45
So the first one externally when we're positioning ourselves, you know, to clients and talking to clients about it, Rohit, you will love this message.
12:53
But our our first conversation is data, right?
12:55
So if you do not have structured architecture around your data, if you do not have clarity on, on what you're looking for, you will not get a good result.
13:03
So, you know, we're still in a world of rubbish in rubbish out Gen.
13:07
AI will not fix that necessarily for you.
13:09
So you know that that is, but that is our clear message.
13:12
Any Gen.
13:13
AI transition strategy needs to start from from your data architecture internally in terms of how I prepare myself and how I'm helping my team prepare myself.
13:23
Is is exactly as Rohit says.
13:25
Just just use it.
13:26
Start practicing, develop use cases, you know, understand how can I use this in my day-to-day.
13:32
That then helps me understand where it's beneficial for others.
13:35
So you know, things as simple as using, we have an internal AI platform, but using that to write an e-mail for me saves me so much time in my day.
13:44
Using it to write reports or generate insight for me saves so much time in my day.
13:48
And getting my team to think about what tasks do you do that are repetitive or are repeating and then can we get AI to replace that?
13:58
So quality checks, quality audits, anything that is assessment based that we can use our AI platforms for.
14:04
And we have forums, evangelical forums like you talked about Rohit, where we share, you know, as a team and then as a function.
14:13
Where have you used IX Hello, this week to benefit?
14:17
What did it save you?
14:18
How much time did it save you or what was the quality increase you got from using this platform?
14:24
And it just takes, you know, somebody to say, well, I used it to assess all of our coaching output and it gives me this result for somebody else to then think, oh, well, that would be relevant in my business if I adapted it in this way.
14:36
And then ideas grow and develop and some work, some don't.
14:39
But like you say, you always learn.
14:42
So I think enabling the conversation internally, having people constantly try and then sharing the use cases is really how we're driving the conversation internally, which helps us then drive the conversation externally.
14:58
Yeah.
14:59
One thing that Rohit mentioned is having somebody as an evangelist and within the team is, you know, I was recently talking to another service leader whose team was thinking they had change management issues on AI.
15:16
And Sarah, from your perspective, do you have other tools in your belt like having a person as an evangelist?
15:25
Both of you can answer this, right?
15:26
But Sarah, we can start with you.
15:29
What is it that you do to bring that change management in so people are willing to try?
15:37
I think what has worked for us is not pushing it as an enforcement.
15:44
So you must do.
15:45
This is not how we do business.
15:47
It's more bringing, bringing out the stories of look how this helped this person, you know, just by typing this in, they were able to get this result which saved them this much time in their day.
15:59
There is a fear factor associated with.
16:01
I don't know what that is.
16:02
I don't know how it works.
16:04
I don't know what to use it for.
16:07
So to help those people start to use the tools that are there, sharing how other people have found them useful has helped.
16:16
We've also developed a suite of smart assistants.
16:19
So you know, we've we've developed some already ready platforms and tools that people can just go in and they're pre pre loaded with commands that that help people.
16:30
But I think sharing the use cases where people can go off in their own time and give it a go and start to become more familiar with the product.
16:40
So finding those early adopters, because there are, there are people in every team who just always want to be the first one to use, the first one to learn and they get excited about technology.
16:51
And there are always the people who are, Oh no, not for me.
16:54
I like the way things always have been.
16:55
I'll just carry on.
16:57
So using the early adopters to share the use cases with those people who are more nervous really helps us gradually roll out the full capability and get everybody to the point where they're comfortable and becoming evangelists for the the platform.
17:14
So, you know, I think that that's that's the sort of approach that that multiplies and really starts to avalanche in terms of impact because you start with maybe 10 people who are really at it and really love it.
17:27
And then all of a sudden they have an impact on 100 people who have an impact on 10,000 people who have an impact on 100,000 people.
17:34
And you know, that that's it's kind of that cascade type sharing of information and sharing of case studies that that really helps our teams understand how AI can improve their day in work.
17:50
That's excellent.
17:50
That's very much like the cascade effect, which is the start up story, right.
17:54
So we are also looking for early adopters and sharing their their gains and their expertise into everybody else to say, hey, you know, this is what can be done and show them how it can be done.
18:06
So everybody else comes in.
18:09
Anything else you want to add, Rohit, outside of what you had mentioned?
18:12
Yeah, I think the one of the things that I'll actually give a shout out to my company as well.
18:18
At Databricks, there is this culture of, you know, you have all access to all these different tools.
18:23
The platform itself is, is an AI platform.
18:27
At Databricks, the product that we give our customers is something that we consume internally to build a lot of AI stuff, AI tools and everything is accessible to all folks within the company.
18:37
You know, whoever wants to get access to a language model, query the language model, build a tool.
18:41
The culture itself is such that you know, so many seeds and you're going to see something come up from somewhere and you know, bring it all together and try to build something that's more uniform to everyone's access.
18:53
It reminds me of a of a very similar story of, you know, Apple's Co founder Steve Wozniak.
19:00
You know, when he was growing up, his father would leave around capacitors and transistors around him and he would pick them and try to build something.
19:06
And eventually that what that's what motivated him to build the first Apple computer.
19:11
It's, it's a very similar story.
19:13
Like we have access to so many different tools.
19:15
We have the culture of sharing what we build.
19:18
We have a distribution list where we keep, you know, pumping out whatever we, whatever we build, we share it if it picks somebody's interest.
19:25
Even we have the, the CEO of the company, you know, commenting on certain of the certain, certain contributions.
19:30
So I think that culture also matters.
19:33
It's very difficult for, you know, small group to bring a big change if the company itself is providing that culture.
19:39
I think that makes a big, big difference.
19:41
Yeah, You know, we are also, we talked about the use cases along this entire thing.
19:48
So that's excellent.
19:50
You are, you know, the culture of the company and the culture that the company sets out.
19:56
So there is a recent Shopify memo that has been going around.
20:00
Where the culture is set for what we need to do with the AI and everybody else is following through.
20:06
That's amazing.
20:08
And we are actually, even as a company, we, I love talking to, you know, support executives and CI OS who have actually played around with AI.
20:20
Because when they have played around with AI and they have done those PO CS that you both have been talking about, they have a very good idea about what use cases are their priority.
20:31
And it come and they exactly know what kind of issues they will face and what are the, you know, how to evaluate a tool vendor too.
20:43
And that it is, I think, very, very, very effective.
20:47
So because you're not teaching them from the basics, but you're actually going and adding in to their learning.
20:56
So that's excellent.
20:58
We do have, you know, Sara, you talked about measuring efficiency, saying hey, the same time for this person and how do you take that expertise and how do you bring everybody up to speed, right?
21:12
And that is a way to measure efficiency.
21:15
If there is any other ways that you are thinking about efficiency, which is, you know, before and after how much time and how much time, if that's the standard, if there is anything else, please feel free to jump in.
21:26
Otherwise, we can go into the next question.
21:31
Yeah, well, I mean, I think it, it depends on the scale, but you know, for internally we measure based on simplification, based on effectiveness and based on efficiency.
21:40
And there's also a qualitative assessment, right, Because some of our AI tools are looking at, well, how good is the output, How good were our coaching forms, our coaching evaluations, for example.
21:50
And we can load what our expectations are and compare the work that our team leaders are doing and to that expectation.
21:58
So that gives us an automated quality report that would have taking a whole bunch of people and a whole bunch of teams to do previously.
22:04
But you know, we can load that data into the tool and it will give us that qualitative feedback.
22:09
So I think there are quantitative measures which are around efficiency, cost reduction and headcount and so on and time saving, all of those sorts of things.
22:19
But there are also qualitative benefits to be had by ways of, you know, measuring consistency measuring and how high quality our our outputs are.
22:28
So there's, you know, I think for me, there's the the two lenses through which we would look at our output from an AI perspective.
22:38
Yeah, consistency is a key and people tend to not, you know, remember that.
22:45
So I'm glad you mentioned it because consistency comes in two forms.
22:49
One is, am I getting the same result every time I check because that's when I can make actions?
22:54
That's one.
22:54
Second is consistency that I've also heard is if Sarah is in Belfast and Kay is in California and Rohit is in North Carolina, can they be consistent in their own actions irrespective of where they are?
23:11
Right.
23:11
So, and that is also a consistency that can be achieved with AI.
23:16
So I'm glad, very glad that you brought that in along with quality.
23:21
Rohit, anything else you want to add to what was mentioned?
23:25
I think it's a big plus one to what Sarah mentioned.
23:27
The only other thing that I could think of is the evaluation set, right?
23:31
This whole concept of how do we evaluate these new systems, these new genea systems, we are talking about building in determinism on top of a non deterministic system.
23:43
The large language models generally generally are known to be highly non deterministic and we are trying to build a layer on top of it to ensure that that we we get deterministic output.
23:53
So the key aspect of that is building an evaluation set for every project, every kind of POC that you're trying to build.
24:03
Am I really aligning to the evaluation I set out to do and on my evaluation sets, really clearing that's that's key.
24:09
And I think probably also to ask the question at every step of the way, is AI really required here, right?
24:18
Because we, we, we, we tend to pick up the new thing that's coming in and try to find, OK, everything is a, you know, I have a hammer, like, let me find a nail and everything is a nail, right?
24:28
So it's, it's very important for us to debate with ourselves to ensure that is AI really needed here?
24:34
Or can this be just automated by a simple traditional way of automation?
24:39
And I think those, those are key things people leaders should definitely remember.
24:45
Well said, very well said.
24:46
And you know, I think the deterministic system, that's a completely separate technical topic and we can talk for hours on it, which is why I always say that is just stable stakes.
24:59
So, so having said that, you know, we're finding all these efficiencies all throughout the organization, all throughout support everything.
25:09
And one of the things, you know, I've never seen that to be a issue so far because support organization, service organizations are so stretched so thin.
25:19
But people always question saying, hey, you know, what do we do with people when there is this additional capacity, right.
25:26
So your perspective on it would be great.
25:29
And Roger, if we can start with you.
25:31
Yeah, absolutely.
25:32
I think just like any other technology, even Jenny is going through an infancy stage, right?
25:39
So you need some guide.
25:41
It's like a child that you're growing along with you.
25:44
So it's first going to augment your work.
25:46
It's not going to replace your work.
25:48
So during the augmentation phase, there is a lot of opportunity for our existing manpower to upskill themselves to ensure that the AI is all sub skilled.
25:56
Accordingly, like Sarah mentioned, data is key.
25:59
I mean, if you don't have the right data, no matter how sophisticated your agent, AI agent is, it's going to give you, you know, garbage results if the garbage data exists.
26:10
So it's very important for these existing engineers, existing folks, existing personnel to work alongside with AI to make AI do some of their mundane tasks, automate some of their routine tasks so that they can focus on tasks that require creativity, high skill involvement domain, the ever changing domain importance, right?
26:34
So your product is constantly evolving and the AI is going to take some time to catch up to it to get, I'm getting complex.
26:41
Yes, absolutely.
26:42
And and the demands are also increasing.
26:44
So let's say you now have AI agents working for you.
26:47
The expectation from customers are also going to say, hey, now you have that you should be faster in your responses.
26:52
So you know what, what is now stopping you?
26:55
So there's a lot of opportunity for the for engineers to upskill themselves from where they were and, you know, offload the, the mundane, the time task, time taking tasks to systems like AI and over time become specialists in their areas.
27:11
I think there are so many avenues.
27:14
I just feel that it's, it's a revolution just like how things start improving from people doing manual work with the Internet coming on, you know, people said they don't, you know, Internet is going to take over.
27:25
Of course it took over, but it took over alongside a highly skilled set of engineers who upskill themselves to utilize the the Internet as it should be.
27:35
And I'm, I'm going to I'm, I'm very positive that the next set of engineers that are going to come up or grow into are going to be highly skilled engineers that are going to team AI to use it to the best possible way, just like how we see the Internet being used today.
27:52
Sarah, Yeah, very, very similar.
27:55
You know, I think having been in this industry for 20 years, you know, we started with telephone calls.
28:00
And when e-mail came along, everybody said, oh, nobody will need a phone because we've got e-mail now.
28:05
And then when messaging came along, everybody said, oh, we'll not need e-mail or phone anymore because we've got messaging and we still have phones and we still have e-mail and we still have messaging.
28:14
So now we have an additional capability and I think every through all of those evolutions, we've seen the role of, of our frontline team become more and more advanced to, you know, exactly as you said, the skills have increased and people have had to learn more.
28:31
We've had to teach them more, they've had to develop different skill sets.
28:34
And I think that this is just, we're still on that continuum.
28:37
So, you know, we, we have time to develop people's capability in terms of what are what are the requirements now to keep our customers satisfied given the world of AI.
28:49
So it may be we need to respond quicker.
28:51
It may be that we don't have different methods of accessing a human contact if that's what's required.
28:56
And they're yeah.
28:57
So I think it's the upscaling of our staff and understanding where, where do we fit in and where does AI augment and improve what we're doing and what does that leave us that we need to do as the, the human interface behind support.
29:14
So, you know, I, I think it's, it's opportunity for people to learn and grow.
29:19
And I think it's exciting.
29:22
It is exciting.
29:24
So if you're just going to add one more aspect, like probably usually gets unnoticed is how quickly academia is responding to this change.
29:32
If you look at all the things that are happening in colleges and courses and stuff, they're already implementing so many of these newer tech technology to be to part be part of the curriculum.
29:42
So the next work, you know, the workforce that comes out and comes out and graduates, you're going to see them already at a level where you know, the highly skilled able to use these tech to the best of its possibility, being extremely enterprising and building out newer tech along with this.
29:59
So I think we are at a phase and are, you know, very lucky to be in this generation where, you know, Internet AI, all of them are really flourishing and academia is actually responding it to it very, very quickly.
30:13
So I'm very bullish.
30:15
Yeah, that actually the support needs for this generation will be different to the support needs for me because my reaction is I want to pick up a phone and talk to somebody.
30:24
My 16 and 13 year old children do not have that response whatsoever.
30:28
You know, so how consumers respond to, you know, kind of how they engage with that technology will will drive us to to how the service industry then needs to respond.
30:39
Yeah, so this is awesome.
30:42
I will, I think it's a perfect way to end in this bullish and optimistic tone that you both mentioned for the how the next generation is getting prepared, how the existing people are getting prepared and how we are all marching towards yet another change where change is the constant.
30:58
So, Rohit, Sarah, thank you both very much for your time that you took to share your experience, your background to the rest of the service leaders and the rest of the service and the support community, thank you.
31:10
Awesome.
31:11
Thanks for having us.
31:12
Really appreciate it.