Beyond Use Cases: Rami El Debs Discusses AI Strategy for Businesses
This episode features a deep dive into the fundamental questions surrounding AI adoption, led by Adam Davidson and Rami El Debs, who runs AI strategy for Accenture. They discuss the common pitfall of jumping straight into tool selection without first understanding the broader context of AI integration within an organization. Rami emphasizes that businesses should focus on identifying their core challenges and how AI can be woven into existing workflows to create real value, rather than merely chasing use cases. The conversation highlights the importance of empowering employees and unlocking insights from within the organization, as well as the evolving role of consulting firms in an AI-driven landscape. Ultimately, they advocate for a proactive approach to AI, urging companies to embrace the technology as an enabler of transformation rather than a separate entity.
Takeaways:
- AI adoption should focus on fundamental business challenges rather than just use cases.
- Understanding the integration of AI into work processes can lead to real productivity gains.
- Companies often overlook how AI can unlock employee insights and creativity from within.
- A holistic view of business operations helps identify where AI can add real value.
- The future of enterprises involves reshaping organizational structures to leverage AI effectively.
- AI should be viewed as an enabler, transforming how work is done across teams.
Transcript
Welcome to the second episode of the FeedForward podcast.
Adam Davidson:I'm Adam Davidson, one of the Co founders of FeedForward.
Adam Davidson:This week we're going to look into some fundamental questions about how to think about AI adoption.
Adam Davidson:At Enterprise.
Adam Davidson:We so often jump right into tools.
Adam Davidson:Should I get this tool or that tool?
Adam Davidson:This LLM or that LLM?
Adam Davidson:Should I use Microsoft's product or OpenAI's product?
Adam Davidson:But there are some basic questions that precede any of those tool questions.
Adam Davidson:What is our fundamental approach to even thinking about AI?
Adam Davidson:All right, that sounds pretty abstract.
Adam Davidson:We're going to get into it in great detail right now with Rami El Debs.
Adam Davidson:Rami runs AI initiatives for Accenture, the massive consulting company.
Adam Davidson:He's also one of the first members of FeedForward.
Adam Davidson:Rami, it's great to both have you in the FeedForward community and have you on the podcast.
Rami El Debs:Thrilled to be here.
Rami El Debs:Thank you for having me.
Adam Davidson:I wanted to jump in right away because before we started I had asked you, I want to get into use cases, and you said, don't ask about use cases.
Adam Davidson:That's the wrong approach right now.
Adam Davidson:I love that point.
Adam Davidson:But walk me through what was on your mind.
Rami El Debs:Well, the thing with AI that we see that is a bit different than previous technologies is its general purpose capability, right?
Rami El Debs:This, like, emergence of things that it can do, which also make it not a thing that you put in a box and really limit in one place, but understand how to integrate in your work.
Rami El Debs:Right?
Rami El Debs:Even Ethan says, invite AI into what you.
Rami El Debs:The problem with going after use cases only is you end up most of the time with a whole bunch of widgets that might or might not deliver value.
Rami El Debs:And when the work process itself is an afterthought, we usually see companies, including ourselves, struggle with scaling.
Rami El Debs:And I'll give you a very real example.
Rami El Debs:We're working with a company in the oil and gas world and they were really interested in solving problems with AI and they had been following very much of a use case approach for a while.
Rami El Debs:And one of the big problems, problems that they brought to bear is, okay, we have these kind of field folks that work in the field and perform tasks there.
Rami El Debs:And they clearly were screaming and shouting about this problem that they have.
Rami El Debs:It's so annoying.
Rami El Debs:They spend two hours or more a day trying to go after very manual activities in the way they develop plans and so on and so forth without getting into a lot of the detail.
Rami El Debs:And we went and we applied AI and it automated all of that.
Rami El Debs:But then what we ended up with is field Agents that were a bit more satisfied with their work.
Rami El Debs:But we could neither increase the real productivity of the field, neither remove heads.
Rami El Debs:We couldn't do any of those things.
Rami El Debs:Right.
Rami El Debs:So there was no real business value that was translating to the bottom line.
Rami El Debs:And there's this kind of mindset usually in companies.
Rami El Debs:I'll give you a very real example.
Rami El Debs:We're working with one of the biggest logistics firms in the world and the first day we go in, they give us basically a spreadsheet that has 4,442 use cases.
Rami El Debs:And they were very proud that they identified thousands of use cases for AI.
Rami El Debs:So this is why with AI specifically, because it's more general purpose than saying RPA and things like that that came before it and all the other technologies.
Rami El Debs:How do we actually start thinking about the work that we're doing?
Rami El Debs:What are the big hairy, must win business challenges that we want to go after?
Rami El Debs:How do they actually impact real metrics, operational metrics on the ground?
Rami El Debs:What kind of changes in the work do we need to really impact?
Rami El Debs:And then understand, okay, if those are the changes in the work, where does AI fit into this equation and how does it help?
Rami El Debs:And this way you'd end up with three, four big ticket items that you're using AI for by design linked to real business value that you know will be generated instead of doing it the other way around.
Adam Davidson:I love this.
Adam Davidson:I mean, one thing it makes me think of is if I can define a use case, it's inherently backwards looking.
Adam Davidson:I'm assuming my company is exactly the same, except this one thing we used to do, maybe we won't do anymore, but AI is really an enabler of so many new use cases, most of which we don't even know what they are.
Rami El Debs:100%.
Rami El Debs:You said it very well.
Rami El Debs:The one thing also to clarify is at the end of the day, once we really understand where AI will apply with this other methodology that we were just talking about, when we decompose the work and go and implement, we will end up again with use cases, if you will.
Rami El Debs:But it's just the approach of don't start there and say, what are 15 little things?
Rami El Debs:But more like get there by thinking about the work itself.
Rami El Debs:Exactly what we're saying.
Rami El Debs:A lot of companies do end up saying, okay, applying AI for knowledge management or talking to my documents or something like that is good.
Rami El Debs:And you will end up with a lot of those basics, I think.
Rami El Debs:But it's also like, for what exactly and to do what?
Rami El Debs:To change what in my work, in my ways of Working, that matters most, right?
Adam Davidson:And there's a subtle point in there, which is you're not saying you're not going to use AI.
Adam Davidson:Obviously, there will be use cases, there will be tools.
Adam Davidson:My sense is increasingly we won't think of AI as a separate thing.
Adam Davidson:It's already embedded in the iPhone.
Adam Davidson:The idea that there's this category of AI, like there's a category of CRM or category of Excel spreadsheets or something like that, some people use it sometimes, some people don't use it, sometimes it's just going to be everywhere.
Adam Davidson:So I know you can't tell us a specific company's name or anything, but walk me through.
Adam Davidson:What is a healthier approach.
Adam Davidson:Let's say FeedForward, the giant enterprise of FeedForward, calls you up and says, hey, Rami, we want to hire you in Accenture to help us figure out how to use AI.
Adam Davidson:Maybe we come to you and say, hey, we have all these new members onboarding.
Adam Davidson:We'd love to automate their onboarding process.
Adam Davidson:We'd love to have a bunch of tools that reach out to them if they haven't posted on the Discord server in a few weeks.
Adam Davidson:In fact, we have 4,428 use cases that we've listed.
Adam Davidson:And we want you to help us hit all of these and decide if we should design our own or buy already existing tools.
Adam Davidson:Walk me through how that conversation then goes.
Rami El Debs:I'll walk you through a real example without naming a lot of specifics, but we're working with a company in the chemical sector and they have some big business objectives that they want to realize.
Rami El Debs:This is one of the companies that are thinking about it in our perspective in the right way.
Rami El Debs:And what they came and said is, look, we have billion plus revenue generating goals that we're trying to achieve on top of the current plans that we have.
Rami El Debs:We have hundreds of millions of costs that we need to take out.
Rami El Debs:And we also want to give our field service agents a day back in their lives.
Rami El Debs:So these are three big problems that we started with Sync or you worry about where to apply.
Rami El Debs:I tell me, like, what are you trying to do from a business perspective?
Rami El Debs:It's also one of the companies that are, I think, more advanced from a clarity of vision.
Rami El Debs:They knew that this is what they wanted to achieve.
Rami El Debs:They didn't know that they could think all the way from that to get to AI, right?
Rami El Debs:So we came and said, okay, if we were to take your company and try to dissect it and understand what are those streams of Value in your business.
Rami El Debs:So a bunch of processes unconstrained by your organization, different functions, all of these things that each one in a contained way delivers end value to your bottom line, to your customers, and so on and so forth.
Rami El Debs:What would those be?
Rami El Debs:So we took the most important one for them, which was from lead to cash.
Rami El Debs:So all the way from pre sales to sales to proposal development to fulfillment and servicing in the field, and all the order to cash back office, all of that value stream.
Rami El Debs:And we tried to say, okay, what kind of levers can we pull from business perspective to either unlock the target that you have from a revenue perspective or from a cost perspective, or help your field agents get their one day back.
Rami El Debs:And when we look at that and put all of the work that happens in this end to end value stream, which is pretty complex at the table, and look end to end horizontally across things and not every function by itself, you immediately start seeing.
Rami El Debs:Whenever they were thinking about field service agents, all they were trying to look at is what does a field service agent do?
Rami El Debs:And how do I potentially automate things they do or augment things they do.
Rami El Debs:But in reality, when we looked at the end to end thing, we, we discovered that if I can actually solve problems that are way downstream from the field service agents, things that have to do with billing disputes with their customers, things that have to do with fulfillment problems, this one day back is solved because a lot of the topical problems that they spend their time doing, those field service people who talk to their own clients, their own institutional clients, spend a lot of their times trying to solve billing issues that happened with the billing department down the line.
Rami El Debs:So this kind of horizontal view, it's nothing new.
Rami El Debs:It's really just taking a step back and looking at the business end to end.
Rami El Debs:I can start uncovering patterns where I don't need to just look at it in a very deterministic way and go and automate every step.
Rami El Debs:But I can actually create a more dynamic system where AI can make smart decisions.
Rami El Debs:Okay, for this type of situation, I can bend the process here a bit, do this, do that within a certain obviously like risk framework, et cetera, that we set for the company.
Rami El Debs:So it's this kind of approach.
Rami El Debs:This is just one real example.
Rami El Debs:And this actually ended up with nine big, you can call them agents or whatever you want to do, but like nine big roles for AI to play that are much closer to an expert human that you would think about in this situation, or super expert human versus saying, I'm going to add a widget to Salesforce here, to do only this, to do only that.
Rami El Debs:But really they're contained packages of work that actually deliver value directly, this kind of flow.
Rami El Debs:By doing that now, I can immediately go and translate this thing that I did to.
Rami El Debs:How does the work of a field service agent itself, like that Persona, that work?
Rami El Debs:How does that look like in the future?
Rami El Debs:Because they don't have to do all of these things because AI is doing so.
Rami El Debs:That's one example of where this is applying.
Adam Davidson:So on the one hand, whoever might have been doing that kind of process engineering, or whatever you would call it, 20 years ago, 30 years ago, you talked about a widget in Salesforce.
Adam Davidson:Maybe I would have had to hope a vendor had created such a thing.
Adam Davidson:Find it, research it, deploy it, train it.
Adam Davidson:And now for at least some types of agentic tools, it's essentially trivial.
Adam Davidson:Once you've identified the problem, it can be an afternoon or maybe a weekend or maybe a week to just custom design for your particular process.
Adam Davidson:So you get to solve these problems.
Adam Davidson:Like whoever had your job 30 years ago, you now have a much wider palette of colors to paint with than that person did.
Rami El Debs:That's definitely a way to think about it.
Rami El Debs:I don't want to oversimplify.
Rami El Debs:What it takes to actually implement, to actually go and apply AI is not super trivial.
Rami El Debs:It's not just a kind of a weekend job.
Rami El Debs:If we look at the hyperscalers of the world, et cetera, Microsoft just released a bunch of agents.
Rami El Debs:The domain of those players is much more the cross company productivity, things that can apply anywhere, that's really the business they're in.
Rami El Debs:When it comes to much more specific workflows for a company, at least from what we're seeing right now, there still needs to be a lot of custom work to be done, but that doesn't mean building everything from scratch.
Rami El Debs:This is also the power of AI and these agents where you can take those, what we usually call utility agents from hyperscalers from other companies.
Rami El Debs:I mean we even develop our own and then stitch them together to do something for your process.
Rami El Debs:Right.
Rami El Debs:It's a mix of both.
Rami El Debs:But yes, definitely the palette is much bigger.
Rami El Debs:But the general purpose of the capability that exists is also much bigger.
Rami El Debs:There are a lot of examples like the Bloomberg GPT1, we used to think that we needed something very specific to solve a specific problem.
Rami El Debs:And then as frontier models come up, it's very clear that they can do much more.
Rami El Debs:So it's really understanding how to use those, but integrate them in the work process, I guess.
Adam Davidson:I want to get at this point because for me I'm part of this four person, five person team that runs FeedForward.
Adam Davidson:And then I have some other things I do.
Adam Davidson:I can create amazing tools in a weekend or an afternoon.
Adam Davidson:I do feel like I have a whole new range of superpowers.
Adam Davidson:And of course at a large enterprise I don't have to worry about the kind of safety and security and IP protection and all of those issues.
Adam Davidson:But the other piece that I think about with AI that feels relevant is unlocking knowledge.
Adam Davidson:Those field agents, just using that example, every company has people on every surface of the company.
Adam Davidson:There are people dealing with clients, there are people deep in the financial engine, aware of all sorts of inefficiencies.
Adam Davidson:And in a traditional corporation it can be hard to get that intelligence to senior leadership.
Adam Davidson:You described a very nice top down view of how to think about AI.
Adam Davidson:What about that bottom up piece where you're empowering workers, you're using AI to unlock insights, creativity that already exist.
Rami El Debs:It is both, right?
Rami El Debs:Top down is really just to understand where do I invest in the right way and then once I understand that it needs to be bottom up work to understand you're empowering at the end of the day with this and all of that stuff.
Rami El Debs:To give you a very real example around that is another company we're working with and it was much more of a traditional cost reduction exercise that they were going after.
Rami El Debs:They spend tens of billions of dollars in capex every year.
Rami El Debs:It's also an asset intensive industry and they wanted to understand where do I deploy my capex?
Rami El Debs:And this is really like more of the bottom up, right on the ground, finance team and asset owners and things like that.
Rami El Debs:Wanted to understand where do I deploy my capital better, where is my capital actually performing on the ground and things like that.
Rami El Debs:And this is always a very, very data intensive exercise.
Rami El Debs:It used to be done in a traditional consultative way where you get all of this data, try to analyze as much as you can with a bunch of humans and it's always a 80, 20 rule.
Rami El Debs:And you end up leaving a lot of value at the table because it's just impossible, right?
Rami El Debs:Especially with a ton of unstructured data that exists in processes like that.
Rami El Debs:A lot of PDFs scanned from the field and I don't know what.
Rami El Debs:And it's not just a matter of having very structured tables all the time.
Rami El Debs:So the power of AI coming to something like that, which was very much bottom up, is also its ability, especially with generative AI, to consume unstructured data in A very fast way.
Rami El Debs:So what we end up doing with this company is to say, look, why don't we actually take these 15,000 PDFs that you have that are actually picture scans.
Rami El Debs:They're not even like searchable text with all of this structured data that you have in your systems that have to do with cost and things like that.
Rami El Debs:Also very technical data that has to do with the performance of those assets and things like that.
Rami El Debs:These are like data modalities that are almost never joined together before.
Rami El Debs:Not because people are stupid, everybody knows that there could be value in doing that, but it was very hard to do that, right?
Rami El Debs:Time consuming and so on and so forth.
Rami El Debs:And to our surprise, because even us as AI providers, we also discover what AI can do as we do stuff with clients.
Rami El Debs:We ingested all of this data.
Rami El Debs:It was a bit of an advanced drag model and vectorize it and all of these things.
Rami El Debs:And what we ended up discovering is that it wasn't only good at saying, do this analysis for me or do that.
Rami El Debs:It was also great at saying, look, I actually gave you, I don't know, a gigabyte of data from all of these different things.
Rami El Debs:Look at it and tell me what kind of analysis I can do.
Rami El Debs:What could I do?
Rami El Debs:Ethan has a bunch of those cool experiments that he does online, but this is almost that at an enterprise level, right?
Rami El Debs:Saying what can you look at?
Rami El Debs:And the type of stuff that it came up with was completely astonishing even for the team.
Rami El Debs:The level of specificity around try to actually correlate this kind of operational metric from the asset with this kind of cost here.
Rami El Debs:Okay, let me try to do that.
Rami El Debs:Can you do it for me?
Rami El Debs:And it can actually do something for you.
Rami El Debs:And now what kind of insights can you get from that?
Rami El Debs:And we quickly saw that it was much more than just accelerating the work of people, but also unlocking to your point, a lot of those insights that they were never able to raise to their management because of that.
Rami El Debs:So this is just one example of them talking to their knowledge, if you will, or talking to their enterprise data in a way that was not possible before.
Rami El Debs:And this was done not in a weekend, but in a couple of weeks.
Rami El Debs:Six month piece of work.
Rami El Debs:It was done just like an innovative thing on top of this process they were already running.
Rami El Debs:Right?
Rami El Debs:So to your point, we're seeing quite a bit of those things.
Rami El Debs:It's important, I think, for companies to, from one side know when they can apply those things and when not, but also be courageous to try.
Rami El Debs:This client was actually pretty courageous to say this looks really promising.
Rami El Debs:Let's try it out within the right obviously controls.
Adam Davidson:I want to get into the range with feedforward.
Adam Davidson:I talked to a bunch of different companies, but there's sort of a selection effect.
Adam Davidson:By the time I'm talking to them, they've expressed some interest in AI.
Adam Davidson:Although even within our membership, I think there's a range of levels of engagement.
Adam Davidson:You said 9% are doing it in the Accenture approved way.
Adam Davidson:Is that recommended way, not approved?
Adam Davidson: companies or just top: Rami El Debs: % statistic is from the G: Rami El Debs: So the largest: Rami El Debs:It's that that we're focused on the most.
Rami El Debs:The main thing they're doing that is special is that they're thinking about it and a bit of a holistic way in three different manners, three different components.
Rami El Debs:One is this approach Accenture approved.
Rami El Debs:Actually we learned it from the market.
Rami El Debs:It's not just Accenture approved.
Rami El Debs:It's really this approach of looking at the work and leading with value.
Rami El Debs:The second important point that they're doing those types of companies is that they are really being very proactive about unpacking the impact on operating model ways of working so basically on the people, because they understand it's a people thing, not about the tool, it's about how do we empower our people.
Rami El Debs:It's really understanding what is the impact on my workforce and being pretty proactive about it.
Rami El Debs:The third piece is what we usually call the digital core and all of that stuff, which is really the data and AI backbone and infrastructure that you need to be able to leverage those agents or all of those AI use cases, et cetera, properly in your company.
Rami El Debs:And the main point there is maybe for AI is two things.
Rami El Debs:The first one is the ability to choose your models and curate those models and switch among the models, et cetera, in a smart way to reduce costs.
Rami El Debs:Because now inference costs are.
Rami El Debs:Even though the unit cost is becoming smaller with how much we're using AI, this becomes a huge cost equation.
Rami El Debs:So how that from one perspective, but also the performance of the model being fit for purpose, et cetera.
Rami El Debs:And the second important thing I think is contextualizing it.
Rami El Debs:When the Gen AI boom happened after the release of ChatGPT, if you really think about the way companies used to pass their data and their context started with rag, Advanced rag, then context Windows became very big.
Rami El Debs:But to really be able to keep this AI that you're using specific to what's happening in your company, your domain knowledge, the context of your stuff.
Rami El Debs:This is where we call it.
Rami El Debs:The brain is just a fancy name.
Rami El Debs:But this is where there needs to be a fresh approach to how you look at the data platform.
Rami El Debs:So these three things together are what those 9% of companies are looking at holistically.
Rami El Debs:The second archetype that we're seeing are companies that are still starting from use cases.
Rami El Debs:It's not a terrible thing to do.
Rami El Debs:We don't think it's the most efficient.
Rami El Debs:But there are companies that are looking at that and trying to understand, let me reduce the number of use cases, choose three or four things to go after and try to scale those.
Rami El Debs:The good thing about that is even if it's a point use case, when you're starting from that and looking at the stack down, you're able to solve what you need for this to work versus the third type of companies that are still looking at, oh my God, I can't do this because I need to fix all my data first because my data is a mess, right?
Rami El Debs:Almost every company in the world will tell you their data is a mess.
Rami El Debs:Even though there are levels of maturity, nobody's ever fully satisfied with their data.
Rami El Debs:There are some companies that are still thinking about, let me go fix all my data first and then I'll go think about AI.
Rami El Debs:Now, that group and also the fourth group, which are least mature, we're seeing this group obviously become smaller and smaller with time.
Rami El Debs:But one of the main issues that those two groups usually face is that their core enterprise systems beyond AI, like their actual ERPs and CRMs and all of that, are still pretty legacy.
Rami El Debs:And in this case, I think it is the right thing to do to not immediately go AI all the way, but actually fix your stuff because you will still need those transactional systems, et cetera.
Rami El Debs:Even in the age of AI, it.
Adam Davidson:Is fewer as time goes on.
Adam Davidson:But I still see companies that are pretty much stuck at we're not comfortable with AI and very often what you see is the C suite.
Adam Davidson:Typically the CMO is gung ho.
Adam Davidson:They want to use it as soon as possible.
Adam Davidson:Their staff is overwhelmed.
Adam Davidson:But very often a general counsel or chief legal officer is extremely cautious and maybe doesn't fully understand it and is tapping the brakes.
Adam Davidson:I think you see CTO CIOs can play a range of roles.
Adam Davidson:Some of them are quite resistant because it really is not the old model of the IT department does.
Adam Davidson:An RFP picks a vendor, decides who gets access, trains it, IT'S more wild and chaotic than that.
Adam Davidson:So I don't know what the number is, but I definitely am aware of some major household name giant companies that are pretty much stuck.
Adam Davidson:Although what we know, and Ethan has shown the data, their employees are using it all the time.
Adam Davidson:They're just using it on their personal devices and they're not telling their bosses.
Adam Davidson:So the company's not learning.
Adam Davidson:Do you think?
Adam Davidson:Is that also around 9%?
Adam Davidson:Is that a bigger group?
Rami El Debs:It would be in the 10, maybe maximum 20% range of companies that are still not really using it.
Rami El Debs:But you brought a very good point around the CMO wanting to move fast and then people using it like on their own.
Rami El Debs:What we're seeing in terms of the resistance to use AI are really two big patterns.
Rami El Debs:The first one is from a people adoption perspective, what you said is 100% right.
Rami El Debs:Usually the top C suite is all about, yes, we want to do this.
Rami El Debs:The people right on the field usually are also very eager to use that.
Rami El Debs:The middle management is usually the biggest disconnect that needs to be solved.
Rami El Debs:And that's for a wide range of purposes.
Rami El Debs:We don't need to go into all the details, but incentives can be one.
Rami El Debs:The type of work that middle management does is another, and so on.
Rami El Debs:But that is a more structural people issue that gets solved with more awareness, more understanding and learning about what AI is, how you use it.
Rami El Debs:And by the way, we're no different as a company.
Rami El Debs:We do a lot of AI for ourselves.
Rami El Debs:Obviously, I think we're more advanced than the average company in our own use, but we're not immune to these kind of behaviors.
Rami El Debs:In fact, just to give you a bit of a point, one of the biggest tools that we've invested in to boost our own consulting, advisory type of services, we've put a lot of investment behind it.
Rami El Debs:It has so much data and it's a really powerful tool.
Rami El Debs:But we're seeing a bit of a limitation in how much adoption we're getting.
Rami El Debs:Not what we wanted.
Adam Davidson:Is it internal tool or customer facing?
Rami El Debs:It's more an internal tool.
Rami El Debs:So it's really to accelerate how we provide advice to clients and things like that.
Rami El Debs:Think basically about a tool that can go and analyze the profile of the company, what are their value pools, where to go, what are the trends in industry, things like that.
Rami El Debs:So we've created very specific focus groups to go understand, like why are you guys not using it?
Rami El Debs:And one of the questions we're asking to the people that are using it enough is do you actually use something like chatgpt or something out of the box.
Rami El Debs:And we found that the overwhelming answer was yes, obviously every day.
Rami El Debs:So I think it's less of a resistance of people not wanting to use AI.
Rami El Debs:I think it's really understanding how to use AI, where it can help in their day to day jobs and companies not trying to reinvent the wheel around what they can get from the market that their people are already used to doing.
Adam Davidson:That makes a lot of sense.
Rami El Debs:One kind of very interesting statistic or data point I would bring here is maybe I'll ask you this question.
Rami El Debs:Do you know how much us as humans hallucinate, quote unquote, out of the stuff that you hear in general from human mouths?
Rami El Debs:What do you think is the percentage of things that are inaccurate?
Adam Davidson:So hallucinate in the machine learning way of just very confidently asserting something, my wife would say that I do it 73% of the time.
Adam Davidson:I think it's, I don't know, 15%.
Rami El Debs:Well, it's 50%.
Adam Davidson:50%.
Rami El Debs:Okay, it's closer to 50%.
Rami El Debs:And our chief Learning officer keeps giving this example.
Rami El Debs:The reason why planes don't fall from the sky is not because we don't hallucinate.
Rami El Debs:It's because we've put the systems around us.
Rami El Debs:The right control, the right processes, the right systems, the right oversight around our own human capabilities to make sure that the system's performance is satisfactory.
Rami El Debs:AI actually hallucinates less than us humans in general.
Adam Davidson:That's interesting.
Rami El Debs:We're always afraid of that thing that it hallucinates.
Rami El Debs:Whereas what we should be thinking about is can we actually design a system around it where there's a human in the loop, on the loop, the right risk tolerance thresholds, the right controls in general, and so on and so forth to make sure that the system as a whole is performing in a satisfactory way?
Rami El Debs:Instead of saying no, don't use AI?
Adam Davidson:This is kind of shifting gears a little.
Adam Davidson:What is the role of a consulting company in the age of AI?
Adam Davidson:This is something Ethan talks about all the time that we all now have access to.
Adam Davidson:Let's call it a B minus to B plus consultant.
Adam Davidson:Right.
Adam Davidson:Like I could come up with a business idea, I could go to any LLM and say, use an Accenture like approach to evaluating this business model.
Adam Davidson:And it's not going to be as good as you, but it's going to be better than some people who work in Accenture.
Adam Davidson:It's going to bring a lot of value and it's going to be essentially free and take five seconds.
Adam Davidson:So how do you personally And Accenture, how do you continue to exist in this era?
Rami El Debs:That's a great question.
Rami El Debs:I will not divulge all of our strategy for the future of consulting, but we recognize that it is, I think, and I think all other big consulting companies recognize that we are one of the industries that would be the most disrupted because we are knowledge industry at heart.
Rami El Debs:However, I think the value of consultants in the future will be very specially relevant in three different categories, I think where AI will not, at least in the short term, in 20, 30 years, who knows if we get AGI, what happens.
Rami El Debs:But I will not speculate a lot about that.
Rami El Debs:But I think now it's just like the other domains where we say it's more like the human plus AI that will displace the human without AI versus AI displacing a human.
Rami El Debs:And by the way, even before AI, in the like most recent years, the industry of consulting has shifted in a way where very little of the projects we used, we do today.
Rami El Debs: this to what we used to do in: Rami El Debs:Is the pure, pure kind of question around evaluated business models or things like that.
Rami El Debs:Companies I think already have a lot of those capabilities internally.
Rami El Debs:They still use consultants for that, but it's not the dominant part of the business that used to be before in the age of AI.
Rami El Debs:I think the three big things that consulting companies would still be needed for if they're able to morph into this kind of advisory.
Rami El Debs:The first one is understanding from a high level is this a good business idea or not?
Rami El Debs:Is one thing.
Rami El Debs:I think really understanding how to implement it with the complexities of the world, of the company, of the industry dynamics, all of that stuff.
Rami El Debs:We don't have huge open source wargaming models and industry that can model the world for you.
Rami El Debs:Right.
Rami El Debs:You still need to understand those experts that have been doing this for a while, a lot of data from different companies, et cetera.
Rami El Debs:I think that's one which is basically move from the pure is this a good idea?
Rami El Debs:Kind of type of 6 week, 8 week study to more of how do I construct a business, a true value proposition around that?
Rami El Debs:That's one thing, I think.
Rami El Debs:The second thing is around the human element.
Rami El Debs:Although we have seen research Ethan has published, I think around this in the fact that AI, even in the medical field turned out to be more empathetic than humans and doctors.
Rami El Debs:But in reality on the ground, being able to truly move and change how people work.
Rami El Debs:I think will still be a huge role where we will need humans next to them to train, to work with, to observe, to advise on that angle.
Rami El Debs:And I think the third piece is the whole infrastructural part of that.
Rami El Debs:We use AI today already in our own services.
Rami El Debs:When we talk about designing architectures better, writing code faster like everybody does.
Rami El Debs:But I think being able to truly envision and implement and really bring to life complex systems.
Rami El Debs:When it comes to implementing AI, I don't think AI is going to replace that at all.
Rami El Debs:I think where it is today is still pretty limited.
Rami El Debs:There will be a big role for consulting companies to do that.
Rami El Debs:I think we will lose the pure, pure, pure strategy firm.
Rami El Debs:Pure, pure, pure implementation firm.
Rami El Debs:And it will be more the few companies that can bridge the gap.
Rami El Debs:And you can see this not only in us, you can see everybody moving a bit toward the same trend to try to cover that.
Rami El Debs:And we're very, very intentional ourselves around how we are addressing the AI age for our industry.
Adam Davidson:That does make a lot of sense to me.
Adam Davidson:I've never hired McKinsey or Accenture or BCG.
Adam Davidson:I've never personally been in a position to do that.
Adam Davidson:Now I can.
Adam Davidson:Maybe it's a diminished version.
Adam Davidson:I feel like an economy where everyone has access, even if it's a B minus version.
Adam Davidson:That is a richer economy.
Adam Davidson:I think in terms of medicine, I've spent time in some very poor countries around the world.
Adam Davidson:A country like Haiti, the vast majority of people will never see a medical professional in their entire lives and.
Rami El Debs:But they have access to the best AI that you, me or anybody else has access to.
Adam Davidson:But they, in their devices, they can have access to the best AI and they can get.
Adam Davidson:Even if it's again, B minus medical advice, it's still an entire sea change.
Rami El Debs:It's even becoming more than a B minus.
Rami El Debs:I think it's starting to become like a, A minus consultant maybe.
Adam Davidson:Yeah, fair enough.
Rami El Debs:There's a report that will come out pretty soon from us where we basically examined the future of enterprises in the age of AI because I truly believe, and I know you guys like @feedforward collective also shared that I don't think companies are thinking bold enough.
Rami El Debs:Like everybody goes back to adding, oh yeah, AI.
Rami El Debs:It's not AI replacing humans or anything like that, but it's about augmenting.
Rami El Debs:And I don't believe that it's going to be a small step change, like a small incremental change.
Rami El Debs:I think there is a big, big change that would happen to the fabric of how enterprises work and I'll touch on just a few key points around why that is.
Rami El Debs:The first one is the amplified intelligence of people and of the firm is definitely important.
Rami El Debs:So this idea of unlocking the full potential of intelligence definitely be one of the biggest game changers.
Rami El Debs:But I think there are others as well.
Rami El Debs:I think we will start seeing a lot more T shaped people.
Rami El Debs:So from I shaped experts and functional people to more like T shaped, right, they can really span more of the process and look like a bit more end to end.
Rami El Debs:This will make organizations more dynamic and people will be able to learn more, learn new things, et cetera.
Rami El Debs:Because the AI can take care of the basics or the details of a specific domain that we train.
Rami El Debs:I don't know exactly what this will mean to specific people, but it's definitely a big trend we're seeing.
Rami El Debs:There's a third point which is all around org structures, because beyond the fact that AI will augment humans to make decisions, I think AI will also do itself.
Rami El Debs:A lot of the back office, and I don't mean just enterprise functions from a back office perspective, but the repeatable things, et cetera, even that include taking action beyond just providing advice, will be very automated.
Rami El Debs:So whereas organizations today spend, I don't know, 80% or something like that of their time running processes, I think a lot of the running processes will happen with AI.
Rami El Debs:So there will be also a lot of rewiring from that angle.
Rami El Debs:In terms of teaming models.
Rami El Debs:What are humans really focused on versus the big static orgs that are looking at running the processes every day?
Rami El Debs:I think also with the democratization of a lot of the knowledge that you're talking about, there will be a premium on company specific information that your people have access to and know that defines who you are.
Rami El Debs:A combination of those things, in my perspective, will reshape companies in the next three to five years.
Adam Davidson:Yeah, it has the potential to be a very exciting time.
Adam Davidson:There will be disruption.
Adam Davidson:There will be winners and losers.
Adam Davidson:Some of my friends think I'm too Pollyanna, but I find it easier to picture a world where more people are more engaged with their work with AI than the previous world.
Adam Davidson:As we wrap up, we've been talking on offensive meaning.
Adam Davidson:What are the proactive things I can do to use AI?
Adam Davidson:I did want to talk about defense.
Adam Davidson:We talked about that a bit with consulting.
Adam Davidson:Is AI going to render my whole industry defunct or is it going to render my company defunct?
Adam Davidson:I want to underscore your point that most people are not bold enough.
Adam Davidson:I don't think we're there yet.
Adam Davidson:But at some point soonish, there will be dominant players and it will be just harder and harder to catch up.
Adam Davidson:If you haven't started to incorporate AI, how do you think about that issue?
Rami El Debs:This question, I think has a bit of, maybe not philosophical angle, but an angle around, like when we reach AGI.
Rami El Debs:I think that's the, the biggest, biggest question.
Rami El Debs:There's a lot of unknown around how AI is going.
Rami El Debs:I think if that happens, we reach AGI in the next five years, six years, or whatever it is, I think that would have a completely different, much more drastic change to a lot of different industries.
Adam Davidson:And I don't even know how to think about AGI, to be honest with you.
Adam Davidson:It's like saying, what if aliens invade and just become CEOs of every company?
Rami El Debs:But in the more realistic world that we're living in and what we're seeing and talking to companies about, I have not seen yet.
Rami El Debs:Entire industries at risk of being displaced by AI being disrupted.
Rami El Debs:Yes, with the caveat of using the word disruption, not like Harvard wants with you.
Rami El Debs:But this idea of companies will really feel this change.
Rami El Debs:A massive change for sure.
Rami El Debs:We're starting to see that in a lot of functions, customer service, things like that.
Rami El Debs:Case in point is we're working with a lot of private equity firms who are coming and saying, should I do something?
Rami El Debs:Even early on in the due diligence process, as I'm looking at a company and they're understanding how much it's worth, how do I think about the impact of AI on this industry and this company in the future to understand will it be worth more or less?
Rami El Debs:So definitely there's a big angle there.
Rami El Debs:The vast majority of places and of industries, it's much more back to the same point of this industry plus AI, so their people plus AI versus their people alone, I think that will stand out and completely take off from the rest of the competition.
Rami El Debs:It's really understanding.
Rami El Debs:Back to our first point, right?
Rami El Debs:Just to close on the same thing, really understanding how do I use it, where, what is really competitive advantage for me in the future.
Rami El Debs:And there needs to be some scenario analysis to understand what does my industry look like if this happens, if this happens, if that happens.
Rami El Debs:We work a lot with life sciences companies where we're doing some of that beyond just drug discovery, which is I think now the big thing, but what would happen in terms of access to market, in terms of discovery, these things where AI can change the game.
Rami El Debs:So it's more like which consumer goods company will be able to go to market much faster Understand what to market, what would actually work in the field versus not by using AI for scenario analysis and stuff like that.
Rami El Debs:Right, right.
Adam Davidson:I talked to a consumer goods person who was at a big, well known company who said they were able to do an R and D in an hour.
Adam Davidson:They did what normally takes months or years and had a richer array of new products.
Adam Davidson:And that's not the kind of thing where in one day they become a monopoly and all the other companies go out of business.
Adam Davidson:But you can see how over the years they're just going to have a steady stream of hits that other companies don't have.
Adam Davidson:And that will be a game changer.
Adam Davidson:My default message is at a minimum, now's a time to be learning.
Adam Davidson:Like, if you are wondering about AI, you should @ least be playing around with it A lot.
Adam Davidson:A separate conversation is how do you deploy it at your enterprise?
Adam Davidson:But I do think we're at the point where it's not Bitcoin.
Adam Davidson:It's not a hypey thing that might go away.
Rami El Debs:It's metaverse.
Rami El Debs:Yeah.
Rami El Debs:100%.
Adam Davidson:100%.
Adam Davidson:Get in.
Adam Davidson:All right, Rami, this was fabulous.
Adam Davidson:You're in the discord so people can reach out to you.
Adam Davidson:Are you happy to chat with folks if they reach out, yes.
Adam Davidson:So, Ramiel Debs, what is your proper title at Accenture?
Rami El Debs:I lead our global data in AI strategy practice.
Adam Davidson:Fabulous.
Adam Davidson:All right, well, this was a great conversation.
Adam Davidson:Thank you so much.
Rami El Debs:Rami, thank you so much.
Rami El Debs:Thank you for having me.
