Client Retention and the Performance Economy

Why is now the time for professional service companies to productize?

And what does it mean to productize your services in the age of AI?

And is the AI bubble about to pop?

Hi, I’m Courtney Baker, and this is AI Knowhow from Knownwell, helping you reimagine your business in the AI era.

As always, I’m joined by Knownwell CEO, David DeWolf, Chief Product Officer and Chief Technology Officer, Mohan Rao, and Chief Strategy Officer, Pete Buer.

We also have a discussion with Dr.

Shea Azulai about what non-technical executives need to know about AI.

But first, let’s take a look at some of the biggest AI news from the last month with our July AI Wrap-Up.

Each month on the AI Wrap-Up, I take a look back at a few of the biggest stories that took place this past month that you need to know about.

So first up, Meta announced the release of its LLaMA 3.1 model, an open source, LLM designed to rival ChatGPT, Gemini, claude, and other major players.

There are a few important things to note about LLaMA.

One is that it is free for the general public to use.

If you’re like me and you’ve been paying $20 a month for ChatGPT, plus it might be worth checking this one out as an alternative.

I certainly am.

The next is that it’s different from other popular LLMs in that it’s open source.

In a letter announcing the update, Mark Zuckerberg compared this dichotomy to the way Linux changed the corporate computing world.

This will make AI capabilities available to a broader range of users and organizations at a lower cost than the closed models.

But keep an eye on OpenAI here.

Shortly after LLMA was announced, they shared the news that their GPT-40 mini model will also allow for free fine tuning and open source accessibility.

Next, according to data from Crunchbase, investment in AI startups in the second quarter more than doubled from the previous quarter.

Investment in AI startups exceeded 24 billion in between April and June, nearly double what it had been in each of the previous four quarters.

This news comes against a backdrop of a growing chorus of skeptics beginning to question when or if we’re going to see the massive promised returns from AI that everyone is expecting.

A Goldman Sachs report stumped cold water on how quickly the benefits of AI will arrive and a number of major outlets from The Economist to The Wall Street Journal to The Washington Post have recently written articles pointing out that the ROI of AI seems to be lagging far behind the hype.

But oh ye of little faith, somewhere I’ve got a case logic full of AOL CDs that says we’ve only just begun figuring out how to make the most out of AI.

Last but not least, is AI actually making us less productive?

Someone may have said that in a recent Hot Take episode.

But an Upwork study of 2,500 full-time workers, freelancers and executives found that nearly 80% of workers who use generative AI in their jobs said it has added to their workload and is hampering their productivity.

The study also found a huge gap between executives’ expectations for AI productivity gains and workers’ expectations.

96% of executives expect AI to boost productivity, and 40% of employees say they don’t know how that will ever happen.

The truth?

Probably somewhere in the middle.

That’s it for this month’s AI Wrap Up.

We’ll see you next month to recap some of the biggest AI stories from August.

We’ve talked on previous episodes about what it means to shift from growth to a performance economy.

I was traveling recently and couldn’t be there for this discussion, but David and Mohan dug into the implications of that shift for professional service companies and where AI fits in.

Mohan, we get to have so much fun today because our babysitter is gone.

Courtney ditched us.

She doesn’t get to frame the conversation.

She doesn’t get to reel us back in.

Oh my God.

Yeah, I’ve been waiting for this day.

Yeah.

Okay.

So here we go.

Courtney is on a plane right now.

I will say we did allow her to see the topic.

So she has told us that we have been getting a lot of chatter and questions about, it was probably six months ago, we had a conversation about the shift in the economy, how we’ve gone from this growth at all costs economy to a performance economy, where we now really need to focus on building true cash flow, building a performing organization that doesn’t just focus on growth, growth, growth, but bottom line results and efficiency and productivity, and those types of things.

It’s really happening because of what’s going on in the economy.

Interest rates have not come down.

And I think organizations are feeling this more.

And really, it’s starting to sink into the culture that there is an expectation of performance from investors, from boards.

And the new thing that we have heard is that this is really starting to hit professional services firms.

And the implications of what does that mean to professional service organizations, and how do they need to think differently is a very real thing, and how can AI offset that and play in that transformation that’s happening?

And I know this is something as a hardcore operator, like you think about all of the time.

And I’m really curious on your thoughts, first and foremost, on how has this trend continued to trickle through this economy and is changing the landscape in organizations that you see, but then getting really practical of what should leaders be doing to help navigate their companies through this, and where is it appropriate to bring AI into that conversation?

Oh, my goodness.

We’ve got to now act like grownups, right?

Build real businesses, right?

As opposed to just kind of ride a wave or whatever we’ve done in parts in our careers.

So, you know, the critical thing is, you know, need to focus on serving your customers better, right?

So offering true value at a cost that they feel is justifiable, that not you feel is justifiable, but they feel is justifiable.

And that changes the dynamics of for services companies to start thinking more like product companies, you know, is or thinks is justifiable.

Can I get some calls from one engagement to the next engagement that truly translate?

Am I building intellectual property that makes things better as I go along?

Right, so there is so much of a productization of services that seems important to me in this performance economy, especially for B2B services.

Yeah, that’s been a passion of yours for a long time, this productization of services, and we’ve really seen it play through the services ecosystem and become a rallying cry, not just of thought leaders like yourself, but you see it as a trend in business, that organizations are moving further and further away from bespoke and more to productized advisory offerings and solutions and those types of things.

One of the things I think of when I think about the productized offering, and it’s interesting because we recently had a roundtable where Chris Barbin, the founder and CEO of Trisera, a private equity fund that invests in B2B service organizations, was talking about how there is this trend in professional services companies where they’re really focused and becoming almost paranoid on constantly delivering value because they have seen a shift from a relational business to more of a transactional business.

And when you talk about the productization of services, one of the things I think is, is there a reality there that as we productize our services more, they actually do become more transactional and it will change the DNA of how we think about services companies today?

What’s your reaction to that intersection of two different trends?

I think it’s inevitable that it happens, right?

So whenever you have to offer higher quality in a repeatable process and maybe for a lower cost, just to keep customers happy and coming back to you, inevitably that means it is productization of services.

So it’s going to take new approaches in services to be able to manage these clients better and to get your engagements, not just humans doing it for you, but humans and machines doing it for you.

So it’s got to be that combination of, we’ve had this category called technology enabled services for a long time, but this revs it up to a much, much higher level of technology and humans offering services together.

Okay, so I want to ask you a really hard question about that.

This is the question I get the most and I’m going to put you on the spot instead of being on the spot for this one now.

In that world where everybody knows machines are doing more, there’s more leverage in the professional services business model because now we have this AI doing more, is it a race to the bottom?

Do you have to pass along that cost savings to your customer?

How do you think about that in this economy and how should organizations be navigating this quote, productization also becomes more transactional, comma also becomes potentially more of a commodity?

Yeah.

It’s no different than the life that we’ve lived in the product world.

You just had to keep providing more and more value for the unit cost.

That’s what it comes down to.

If you don’t do that, then you’re right.

It’s a race to the bottom.

It’s the lowest cost provider, so on and so forth.

It’s where phone companies are.

You don’t care what phone company that you use.

It’s just whatever is the lowest cost that works for you and your family is what you go with.

It becomes hard to find differentiation.

But if you can offer more value because the machines can do more of the work and the humans can hopefully do more creative work.

It really works well with the phrase performance economy to be able to think of it as higher creative work, higher value offered to the customer.

It can put us on a better path here of serving our clients better because you can offer more value for the same unit costs.

When I think that through, there’s just low in your price and you’re saying, hey, that’s not the way to think about it.

You want to increase your value.

I can do something 40 percent faster instead of 40 percent cheaper.

I can do it 40 percent better.

To me, the other aspect of that and the real way you differentiate and prevent yourself from being a commodity is differentiation.

The differentiation is not just being better or faster than my customer because that’s arbitrary and that’s just going to be a continual race.

But how do I insert unique value points, have unique insights, have unique aspects of how I deliver, of what I deliver that others simply can’t compete with because their value chain isn’t oriented for that?

Do you have almost a playbook or a roadmap that you can share with leaders of how they should be thinking about that?

How do you find those differentiating points?

How do you find those unique value propositions that others really can’t compete with because it’s not just about faster, better, cheaper?

Yeah, I think David, you know this really well as well.

Every leader has to look at their own company and the value chain of how they offer something, right?

So you can think of it as your front office functions, back office functions.

You can think of it as what you provide to customers, but also what your core capabilities are.

It could be in R&D, it could be creating frameworks that allow your clients to do a much better job.

So implementation is a lot easier, really the thought went into creating these frameworks, having multiple frameworks, having a supply chain of how you are going to find these extremely creative human beings to work on the client engagement, right?

So these types of core processes, it could be in recruiting, it could be in business development, where you really got to look at the company and say, how do you organize your company for this new performance economy that allows you to shine, all right?

So these are various quadrants of how you can look at it.

How can you do serve the clients better?

How can you do things more efficiently within the company?

How can you put out a better product or service out there for your clients?

All right, take all of that and wrap it up for me with the second most frequent question that I get that’s about this is, is the business model of professional service is changing?

Are we moving away from a T&M based capacity, input based model to more outcome orientation, to subscriptions, to what’s the future look like?

And how should professional service firms be thinking about how they may need to shift their fundamental economic relationship with their customers and how they shape that?

You know, I divided into two camps.

There are services companies that offer tangible services, and there are ones that do intangible services.

I’ve heard you say as the ones that are execution oriented, similar concept, right?

If you are in an execution oriented services business, you have to be transitioning to this new world, right?

Without that, it’s going to be a race to the bottom.

At some point, instead of having 100 companies, there’ll be a need for only 10 companies out there, right?

That is the creative destruction that’s going to happen in this process.

It’s not going to be overnight, but it’s going to happen over time.

But if you are in the business where you’re not do it for you, but we help you do it category, it could be something very bespoke.

It’s a particular transaction that you’re seeing through.

It could be very custom software development.

Those types of companies will exist where they’re not going to be threatened, right?

So it’s going to be in the margins where individuals can get more effective there using co-pilots of various shape and form, whether it’s in marketing or software development or whatever discipline.

Those, I think, will stay on as more of the traditional models, but anything that is much more execution-oriented, where you are offering services, where you’re doing it for the clients, we’ll need to look at this new model.

What do you think?

Yeah, I totally agree with that.

I think that the model will change.

I don’t know that we know exactly how it’s going to change or how it’s going to play out, but there is a different expectation, and I actually think it’s to the benefit of professional services firms, because one of the ways you get commoditized is when you’re actually charging for the input versus the output or even hopefully the outcome.

If we can transition away from that, and I think clients are desiring it now and need it, and if we can capture that and figure it out and thread the needle, it’ll be good for all of us.

One of the ways I think we do that, and that last piece I’d add to this conversation is, I do think this intersection of productizing services is going to bring about an even greater emphasis on data.

I think one of the things service organizations have that is to their benefit, but largely have not capitalized on the past, is they have access to information, insights, data across a broad range of customers.

And if we can learn to aggregate that and leverage that data asset in new differentiated ways to fuel our services, I think it is that the epitome of that value we’re talking about, that could be a little bit different.

And if you embed that in, now you can charge differently for your services as well, and it can help shift that business model.

So to me, that’s what’s exciting about this AI economy, is that we now have this ability to truly digitize knowledge work that we do, which is the core of professional services.

And we have the ability to capture and leverage all of these nuanced insights that we see because we work with so many businesses across so many different industries.

And if we can capitalize on the data asset and leverage it to do the things you’re talking about, that becomes the fuel of this transformation.

And so I think that’s the last thing I would leave with listeners, is to not to overlook data, start to look at that data asset in the same way that product organizations look at their data assets.

Excellent.

Well said.

All right.

Did we do it?

High five.

We proved to Courtney we could do this without her.

I think we survived without her.

All right.

Here we go.

We can’t wait to have her back.

Thanks Mohan, it was a lot of fun.

Yep.

Same here.

I’ve got a secret for you.

One way that you can take advantage of this shift from a growth to a performance economy is to utilize AI platforms that help drive client satisfaction and retention.

That’s exactly what we’re building here at Knownwell, and we’re now accepting the first cohort of companies into our early access program.

If you’re an innovator and an early adopter who wants to lead your company into AI, go to knownwell.com to sign up and learn more.

Dr.

Shea Azulai is a cognitive neuroscientist who’s been consulting in areas including AI, data science, and education for more than two decades.

He sat down with Pete Buer recently to talk about what non-technical executives need to know about AI.

Dr.

Shea, so great to have you on the podcast.

Thank you for being here today.

It’s a pleasure to be here.

Thank you for having me, Pete.

Just to give listeners some context for the conversation, could you share a little bit about your background and how AI features in it?

Sure.

Well, I’ve been in this game for quite a while now.

I did AI and computer science back in undergrad, specializing also in graphics processing, which just so happens to relate directly to the way a lot of AIs work.

Then I was in industry doing IT and business consulting before I shifted back into academia.

I went and got PhDs in cognitive science and psychology, where I specialized in perception and learning, education, and how we process and understand things has always been a lifelong passion of mine, and I’ve taught continuously throughout that time.

During that stretch, I’ve also been doing tons of data science work.

Most recently, I was developing some new AI analytics tools for a data science company I co-founded with a friend.

These days, I focus on science communication and teaching people about how their minds work, how they process information with the focus on perception, cognition, learning, and even emotional intelligence.

I really love it.

Awesome.

Thank you.

I’d love to pick up where you just left off around this notion of how the mind works.

Where would you start us there?

Yeah.

Well, usually where I like to start people is to give them a little bit of context for how to think about things.

One of the most common misconceptions we have about the mind is that it’s one thing that has a lot of specialized systems.

That’s not really the way it works.

The brain is many, many different neural systems that evolved at different times, under different evolutionary pressures to solve different problems and they all work in fundamentally different ways.

They also work together and allow us to process things in ways that we couldn’t otherwise.

One of the key takeaways we can get from this is that we are capable of thinking in far more modalities and different ways than we often realize.

Most of the time when we are taught or we’re engaging with people, it’s linguistic thought.

We’re always chatting around and that is probably the most common way that we think about what cognition is.

It’s also the basis for logical thinking, step-by-step reasoning.

On the other hand, we can also think visuospatially.

And the difference between thinking linearly versus visuospatially is, you know, if you’re thinking with linguistics, if I have a really long story I’m telling you, maybe by the time I get over here, you’ve forgotten where we started.

On the other hand, visuospatially, even if the position of two objects relative to one another, tells you something about their relationship.

And the shape of the information that our brain uses to process those two types of information is also fundamentally different.

Which means that while there are really wonderful things we can do with linguistic thought and logical thinking, there are other ways to think and other parts of our mind sometimes are better at processing information in a different way than we’re normally used to.

And that includes, by the way, kinesthetic thinking.

If you’ve ever spoken to an athlete or a dancer and asked them, do you know what it means to think with movement?

They’ll immediately know what you’re talking about.

So we really do have all these different ways of processing and making sense of things.

And that’s one of the things I try to highlight to people is you’re far smarter than most people give, most of us give ourselves credit for.

Awesome.

Well, as you know, this is a podcast about AI and where AI features in the work of humans in business.

With that explanation as starting point, can you help draw the distinction between the ways the mind works, as you’ve just discussed, and the way AI works in terms of learning?

And I think that’s a really good question.

One of the things, I’ve been doing this for ages, and one of the things a lot of people don’t realize is they see a system like these modern LLMs, ChatGPT, claude, and it’s amazing.

Wow, it’s doing all this stuff.

I can ask it these crazy questions.

It’s able to give me clear answers and teach me something in a way that I hadn’t been able to learn before.

But really, if you consider, in this case, these LLMs, fundamentally, they’re about word prediction.

And so, and that’s where it all started.

And you’ve seen this stuff on your phones for ages, where you start typing a text message and it says, oh, do you want to say this?

The big improvement, the big shift that came about a year and a half ago, was the context window for these LLMs grew, so that now it was able to look farther back in the conversation and make connections that it wouldn’t have been able to in the past because it was a much smaller context window.

But even with these large windows and even with the most modern AIs, like ChatGPT 4, Omni just was released yesterday, I think, it still fundamentally is based on word prediction.

And you can see really, there’s one of my favorite examples of where it can get things wrong, is if you just ask it how many words are in this sentence.

There’s seven words.

Sometimes it figures it out.

Sometimes it thinks there’s six or eight or 10 or five.

It’s a little bit of a mushy guess.

And the reason is because it’s not actually counting.

It is predicting what word would come next, and that is based off of the corpus of data that it consumed in order to learn what it learned and make those connections.

So that’s probably the biggest place I would start.

We see the illusion of true intelligence.

We’re not quite there yet.

So how about as we translate this into a framework for business people to think about as they, as they look at all the operations they’re trying to transform and thinking about the role of human versus machine.

Are there kind of rules of thumb about the kinds of thought work or learning that is better left to humans versus better deployed with AI?

A hundred percent.

And that actually ties in pretty nicely with what we were talking about earlier in regards to how the mind works because we do have all these different systems and they work together.

And there are higher order abstractions that occur in our mind in between these systems in the multimodal areas of the brain.

AIs can’t do that.

Even the new versions like GPT-4 Omni that does have a lot of multimodal capabilities, it really is different systems that are put together and talking to each other.

But there’s not a layer on top of that that’s really giving us that same type of abstraction and connecting everything the way humans might.

So what we see is that certain types of things like strategic thinking, being able to have a clear idea of what you want to explain and explaining it to somebody, like just clear communication skills.

Honestly, some of the most basic and fundamental things that we look for, those are the things humans are still able to do better.

And with AIs, the more specific you try to get, unless you’ve got a model that is really trained on a subset of information that it knows to be true, the more specific you try to get, the more you’re going to get errors like that, how many words are in the sentence being counted wrong.

On the other hand, what AI can do quite well is anything that is a little bit softer.

So AI is really good at picking up sentiment in these LLMs, in statements that we’re making.

It’s getting better at picking up emotion and emotional content, not just in the words we type, but also now when we’re speaking to it.

because that is a different type of information and a lot of the emotion is carried in the tonality of our voice.

So some of the surprising places AI does quite well, actually, is in helping somebody, for example, who may be neurodivergent and not quite able to parse social situations as well as they might like.

They can go to an AI and give it some context and the explanation of what’s going on and just ask for another opinion.

And oftentimes, that’s actually really helpful for someone like that because they feel safe with the AI.

They don’t get the same type of emotional peeking that you do when you talk to a human.

So for the humans, one of the big distinctions I draw from your explanation is that the abstraction layer above the individual thought exploration streams, I’m probably using the wrong language, but that’s sort of still the domain of the human.

But you did use the word still, which kind of implies the potential for that not to be the case in the future.

So is that just a function of where we are in the evolution of the technology, or is there any aspect of thought and learning that will ultimately remain the domain of the human?

That’s an excellent question.

I think if you look far enough out into the future, I don’t think that we would necessarily be able to tell.

If there’s a black box between us and another individual, whether it’s a human or an AI, I think eventually they’re going to be fairly indistinguishable, and there’ll be things that AIs can do, actually, that humans can’t do, because in some ways we’re limited by our biology.

Humans are not good at thinking in more than three dimensions.

AIs aren’t limited by that necessarily.

So just as an example, but when we look at what’s happening right now, we…

it’s funny, I see a lot of people talking about AGI, or artificial general intelligence, and we’re so close.

It’s just a couple years around the corner.

And I just fundamentally disagree with that.

I think we are far away from it still.

And for two very specific reasons.

The first is that we do not have any kinds of structures that are combining these modalities and getting this abstraction that we’re talking about.

And you can see that even still in LLMs, if you ask questions that rely on another sense or another way of thinking about things that’s non-linguistic, the AIs aren’t able to do that.

Sometimes occasionally because people have written about things, and so it’s in their knowledge, but not really that well.

So that’s one of the challenges.

The second is just data.

Think about the amount of textual data that exists in the world.

Now think about the amount of data that exists for images and movies.

There’s a ton of all of that.

But those are actually still very limited spaces, and we don’t have data for so many of the other different things that would be required before a AGI could really have the kinds of capabilities that humans have.

They don’t have the senses we have.

They don’t have the ability to interact with the world in the way that we do.

But fundamentally, right now, it’s really a data problem.

I can’t help but try to roll the movie forward in wonder 20 years from now.

As a leader in a business with such rich data gathering, data processing, thought organization learning capability at my disposal, how does my work look different then versus how it looks today?

How am I spending my time?

One of the analogies I like to use is akin to thinking of an AI as a tool like a camera.

How did the invention of the camera change visual art?

Suddenly, somebody that needed to spend thousands and thousands and thousands of hours learning how to paint and mix colors and use the correct brushstrokes, you can still do that, that’s still a great skill, but I can also just take a picture and now I’ve got it in just a few seconds.

So the big thing it’s doing is just speeding up a lot of the work that we want to do and eliminating the need for as many specialized skills.

So even now we can, you know, have an AI very quickly build a presentation for us.

It might not be exactly what we’re looking for, but then we can go in and tweak it.

So we can think of AIs today as a really enthusiastic high school graduate, who’s going to help you the best they can, isn’t going to get everything right, but is a great place to start and to speed up what you might want to do until you’re ready to take a look at it and put some polish.

As we move forward 5, 10, 20 years into the future, those are going to continue to get better.

And AI is, I think, in a lot of ways, are going to become invisible.

In the same way that a camera is sort of invisible to us, it’s just so ubiquitous now, they are going to be systems that we interact with and talk with and reach out to for, oh, you know, somebody was telling me that I should invest in this new technology.

What can you tell me about what’s the basis for that tech and where it’s going and what are the industries that are working on it?

So, I see a lot of specialized systems sort of being the next step, and then those specialized systems continuing to merge and just become better at what they’re doing until we have assistants that are supporting us with the things AIs not as good at doing, building relationships and strategizing.

And figuring out what kind of world do we want to live in and what kind of world do we want to build.

Thank you for rising to the challenge of that question because it’s very much crystal ball, and so I appreciate you lending your expertise to it.

I do think it’s important for leaders to be thinking about.

They have to shape the future of their businesses and how work gets done, and they have to think about their own careers.

And so all of this is fantastic input.

Dr.

Shea, it’s been a pleasure.

You are so steeped in your areas of expertise and so interesting to talk to.

So thank you very much for spending time with us today.

Really my pleasure to be here.

Thank you, Pete.

Thanks as always for listening and watching.

Don’t forget to give us a review on your podcast player of choice.

And we’d really appreciate it if you’d also leave a review or share this episode.

At the end of every episode, we like to ask one of the LLMs to weigh in on the episode at hand.

Hey LLAMA, welcome to the show.

Excited to hear your take.

This episode we’re talking about how AI can be used by companies in the performance economy.

Any recommendations?

Hi, I’d recommend AI driven data analytics and automation tools to boost company performance, enhancing efficiency and decision making.

They can unlock new levels of productivity and innovation.

And now you’re in the know.

Thanks as always for listening.

We’ll be back next week with more headlines, roundtable discussions and interviews with AI experts.