Latest / The Leading Edge / Beyond Efficiency: Ivalua’s Paul Zammit on What Agentic AI Really Changes in Acquisition

Transcript

Beyond Efficiency: Ivalua’s Paul Zammit on What Agentic AI Really Changes in Acquisition · The Leading Edge

• 0:03
  SPEAKER_00: Well, hello and welcome back to the Leading Edge, a series offering in-depth insights into the key challenges, innovations, and evolving dynamics of the contract and commerce landscape.
• 0:17
  SPEAKER_00: I'm your host, Nikki Mackay, Chief Development Officer of the Commerce and Contract Management Institute.
• 0:23
  SPEAKER_00: Now, in today's conversation, we come to you live from NTMA's World Congress, and we bring in a perspective that we don't often get enough on this podcast, and that is the view from inside a software vendor, building the tools that power acquisition at scale.
• 0:42
  SPEAKER_00: Our guest has genuinely a unique vantage point on where AI is actually headed in the space, not just where the marketing says it's headed.
• 0:53
  SPEAKER_00: So I'm joined by Paul Zamitz, Director of Product Marketing for AI and Platform at Evalua, which provides agentic software for the full acquisition lifestyle and is used by 10 US states and major defense contractors alike.
• 1:09
  SPEAKER_00: So, Paul, thank you so much for joining us.
• 1:12
  SPEAKER_01: Thanks for having me, Nikki.
• 1:13
  SPEAKER_01: Glad to be here.
• 1:13
  SPEAKER_00: Wonderful.
• 1:14
  SPEAKER_00: So let's dive straight in because I want to home in on a big gap that we see between how AI is marketed today and what it actually delivers.
• 1:26
  SPEAKER_00: You sit close to both your chief product officer and chief AI officer at Evalua.
• 1:32
  SPEAKER_00: So from that vantage point, what do you think the gap is that you most often see between how AI is in acquisition, how it's being marketed, and then what's it actually capable of delivering?
• 1:46
  SPEAKER_01: So the big, it's a great question.
• 1:48
  SPEAKER_01: The big word out there right now is autonomous, uh, which sounds both scary and magical, and it's all over everyone's marketing, as if the system is just going to go ahead and take over everything for you without you having to worry about anything.
• 2:02
  SPEAKER_01: And uh what really isn't considered, first and foremost, is well, what is the governance of that autonomy and how do you get there?
• 2:11
  SPEAKER_01: Where where does someone actually start?
• 2:13
  SPEAKER_01: What is a system that is capable of doing autonomous work within an acquisition lifecycle in some capacity, capable of the day I start using it, versus how much work do I have to put into it to actually get achievable results, versus where is this really taking me in the long run?
• 2:31
  SPEAKER_01: And I think the thing you see a lot in marketing is people bragging about how many agents they have to give you, to do this job and that job and whatnot, and then the ability to build your own, which is a a long-term vision of what does three to five years look like if every problem that that comes across the table I'm trying to automate with an agent, how many agents have I built?
• 2:56
  SPEAKER_01: Vis-a-vis, how the heck am I governing and managing all of this autonomy?
• 3:02
  SPEAKER_01: I think the reality is software is capable of autonomously doing a lot of work today.
• 3:07
  SPEAKER_01: The human governance structure around it, and then the ability to actually make sure we're managing this safely in a low-risk fashion is not there for most organizations, and they're figuring out how to bridge the gap because technology is now moving faster than we as people and organizations can keep pace.
• 3:25
  SPEAKER_00: Yeah, absolutely.
• 3:26
  SPEAKER_00: We see that an awful lot.
• 3:28
  SPEAKER_00: So, you you know, evaluate your your work is across state government, you're working with major defense contractors, so you see this quite often.
• 3:38
  SPEAKER_00: Um, do you see meaningfully different appetites or barriers to AI adoption between those worlds, or are there underlying challenges that are more universal than people actually assume?
• 3:53
  SPEAKER_01: I'll I'll actually maybe expand it a step further because we do work with a lot of commercial customers as well.
• 3:59
  SPEAKER_01: Right.
• 3:59
  SPEAKER_01: And historically, there was an understanding that public sector, because it is necessarily a more cautious, regulated, transparent sector, yeah, was lagging behind in terms of adoption.
• 4:13
  SPEAKER_01: And and and defense contracting similarly, because at the end of the day, a defense contractor has to be compliant to their federal customer and whatever obligations slow down, uh, a public sector uh entity is has to be compliant and transparent with the public.
• 4:29
  SPEAKER_01: Technology's moved so fast in the last four years.
• 4:32
  SPEAKER_01: Yeah.
• 4:32
  SPEAKER_01: The gap between the commercial space, the defense contractors, and government contractors in terms of technology adoption is not as significant as it used to be.
• 4:42
  SPEAKER_01: And I think that's really interesting from a barriers to adoption perspective, because we actually see the same patterns playing out even across the um the commercial space to the to the government space, to the things like the things like just genuinely uh ensuring that we're reducing risk, things like have we actually provided the right foundation for AI to succeed?
• 5:08
  SPEAKER_01: Have I handled all of my fragmented data before I go ahead and take care of this?
• 5:12
  SPEAKER_01: Um and then similarly, in terms of uh appetite to actually go ahead and adopt it, I think what's interesting about both the government contracting and and and and defense space and public sector, particularly since since we're at NCMA and and and in the US, yeah, uh what we see is in the US, in public sector, especially in the States, you have an entire workforce about to retire.
• 5:39
  SPEAKER_01: You have a whole baby boomer generation about to retire with a whole lot of institutional knowledge that'll walk out the door.
• 5:45
  SPEAKER_00: Yeah.
• 5:45
  SPEAKER_01: And in defense contracting, you consistently have a workforce shortage because it's easy for someone in A and D to go into automotive or industrial manufacturing.
• 5:56
  SPEAKER_01: They they almost have a higher skill set.
• 5:58
  SPEAKER_01: It's harder to come back in.
• 5:59
  SPEAKER_00: Yeah.
• 6:00
  SPEAKER_01: And so the appetite isn't often just around how do we, you know, automate more, work smarter, faster, better.
• 6:09
  SPEAKER_01: What we're seeing is a legitimate how do I encode institutional knowledge into a system before it walks out the door so that my new hires, my kids who just got degrees using Chat GPT, right?
• 6:22
  SPEAKER_01: Yeah.
• 6:22
  SPEAKER_01: Have something that can help guide them to the best practices that that my my 65-year-old who just retired knew like the back of their hand.
• 6:30
  SPEAKER_00: Yeah, that's such a good point.
• 6:32
  SPEAKER_00: Um, I mean, we've just touched on it at the beginning, agenc has been has become one of the buzziest words out there right now, hasn't it?
• 6:40
  SPEAKER_00: What does agentic software, what does it genuinely change um about how acquisition professionals work day to day versus where it's just automated with a new name?
• 6:52
  SPEAKER_01: Yeah, so so probably a year ago, a lot of the marketing I saw was people calling things agents that were secretly old-world automation, things that RPA long could have solved.
• 7:06
  SPEAKER_01: Uh, what an agent really is capable of is automating an outcome, not just a task.
• 7:11
  SPEAKER_01: Yeah.
• 7:11
  SPEAKER_01: And so something that really doesn't necessarily have a predetermined path, uh, maybe is a bit ambiguous.
• 7:19
  SPEAKER_01: To put that in in really concrete terms for acquisition professionals, for procurement professionals, contract professionals, uh, I think an agent can help you across three three lanes that that ultimately, and I think no one really knows, but ultimately transform jobs.
• 7:37
  SPEAKER_01: Not just the way you work, but but what the job will be altogether.
• 7:40
  SPEAKER_01: The first is uh agents change the way that people can access information.
• 7:45
  SPEAKER_01: So if I have an agent that I can ask to search across a hundred contracts to find terms that are similar or dissimilar, I'm operating at a pace that I would have I would have sent 10 interns to go scour through, hey, this law in California just changed.
• 8:01
  SPEAKER_01: We need to make sure that we're ready for it.
• 8:04
  SPEAKER_01: So so I'm able to interact with my data in a much different way.
• 8:08
  SPEAKER_01: So I'm able to ask and be more intelligent.
• 8:10
  SPEAKER_01: Secondly, people are able to delegate work.
• 8:12
  SPEAKER_01: I think the real world is realistically very messy, and a lot of low-value tasks often pop up out of nowhere or or disruptions occur, any anything that that throws your day off.
• 8:24
  SPEAKER_01: And with a gentic systems, because you can hand over a whole task within the right governance boundary, so so making sure you have the final stamp of authority, um uh this ability to delegate work changes the way someone lives.
• 8:40
  SPEAKER_01: If I can, if I can say, hey, this contract's expiring, and and the the contracting officer who was responsible for it changed jobs, like now we have we're in a bit of a bind, someone has to work more.
• 8:52
  SPEAKER_01: Can I delegate the market research to an agent so I can go take that meeting that I can't miss?
• 8:58
  SPEAKER_01: Yeah.
• 8:59
  SPEAKER_01: So I'm not working late tonight, right?
• 9:01
  SPEAKER_01: There's a capacity.
• 9:01
  SPEAKER_01: And then the final legitimate way things do change is there is the ability to automate tasks.
• 9:08
  SPEAKER_01: Yeah.
• 9:08
  SPEAKER_01: To actually, it in in a governed sense, but to set up systems that do autonomously send signals and then actually do things.
• 9:18
  SPEAKER_01: Like if a vendor, if a contractor hits a sanctions list, why can't the system fire off and immediately take a few prescribed steps or prescribed outcomes before coming back?
• 9:31
  SPEAKER_01: And and that whole mixture of ask, delegate, automate sums up really we don't know what the world's gonna be like in five years because four years ago we never expected this.
• 9:45
  SPEAKER_01: Chat GPT was writing silly poems for people and and songs, right?
• 9:48
  SPEAKER_00: Yeah.
• 9:49
  SPEAKER_01: But but the way they work does will significantly change and the people will be capable of more.
• 9:55
  SPEAKER_00: Yeah.
• 9:56
  SPEAKER_00: And as you say, the the world is a messy place, and certainly uncertainty has just become a status quo that we exist in.
• 10:03
  SPEAKER_00: So individuals are having all kinds of things thrown at them about what's in my contracts because you know, constant uncertainty, constantly things are changing geopolitically, and that really impacts so AI is giving people that power back to get that contract intelligence and to um to really take advantage of what they have in their contracts.
• 10:29
  SPEAKER_00: Um so you mentioned AI reshaping acquisition management across many dimensions, and it's not just efficiency.
• 10:38
  SPEAKER_00: What's the dimension of change, Paul, you think is being under discussed right now?
• 10:44
  SPEAKER_00: Something beyond, you know, that speed and cost that this profession, you know, really should be paying more attention to?
• 10:52
  SPEAKER_01: So the I think when and and this is what we see out of our customers who are early adopters.
• 10:57
  SPEAKER_01: We have US states using AI, we have defense contractors using AI, commercial customers using AI globally.
• 11:05
  SPEAKER_01: When when it's being applied appropriately, really all the all the value drivers that procurement and acquisition lifecycle teams are are measured upon, those KPIs are the ones that that do move.
• 11:18
  SPEAKER_01: So yes, efficiency cycle time does move.
• 11:21
  SPEAKER_01: Yes, um savings, uh contract leakage goes down, right?
• 11:26
  SPEAKER_01: What people don't necessarily recognize is if you have a computer system that is actually doing work, it's capable of doing contract work, acquisition work, that's a capacity jump.
• 11:38
  SPEAKER_01: And there is a whole sum of jobs not done, like a backlog of work that people don't consider.
• 11:43
  SPEAKER_00: Yeah.
• 11:44
  SPEAKER_01: Um post con contract post award.
• 11:47
  SPEAKER_00: Yeah.
• 11:47
  SPEAKER_01: Uh an invoice comes through the door.
• 11:50
  SPEAKER_01: Do we have to pay it?
• 11:52
  SPEAKER_01: Do I actually know if there was an SLA?
• 11:55
  SPEAKER_01: If that SLA was actually met in the field, um, if it wasn't met in the field, is this invoice actually reflecting any kind of liquidated damage as a result of missing that SLA?
• 12:07
  SPEAKER_01: The fact of the matter is there's a lot of value that is allegedly created during the sourcing and contracting process that never gets realized because there's a whole backlog of there's just not the capacity to do it.
• 12:21
  SPEAKER_01: In the acquisition world, uh in the US federal government, they audit according to samples of contract files.
• 12:29
  SPEAKER_01: They take a sample because there's only so many people.
• 12:32
  SPEAKER_01: And so the real thing that that I think isn't fully being understood or considered is it's not an efficiency, it's not a cost, it's not a speed, it's a capacity jump.
• 12:41
  SPEAKER_00: Yeah.
• 12:41
  SPEAKER_01: And it's a capacity jump at each person.
• 12:44
  SPEAKER_01: Can each person really deliver more than they've ever been able to?
• 12:48
  SPEAKER_00: Yeah.
• 12:48
  SPEAKER_00: Uh Paul, I think that's a great place to leave it.
• 12:51
  SPEAKER_00: That capacity.
• 12:53
  SPEAKER_00: Can each individual deliver to its full capacity?
• 12:57
  SPEAKER_00: Um, Paul, thank you so much for bringing that vendor's eye view into this conversation.
• 13:04
  SPEAKER_00: Um, it's genuinely useful complement to the government and industry perspective we usually hear.
• 13:09
  SPEAKER_00: And it's clear this shift is about a lot more than shaving time off a procurement cycle.
• 13:16
  SPEAKER_00: So thanks for listening to Leading Edge.
• 13:19
  SPEAKER_00: If you like what you heard, stay tuned for our next episode where we will continue to explore the ideas, innovations, and leadership shifts shaping the future of commerce and contracting.
• 13:31
  SPEAKER_00: Until then, keep challenging, keep adapting, and keep leading.