4 GenAI Lessons from Sangeet Choudary’s Book Reshuffle (2 of 2) (Tech Strategy)

Posted on by

In Part 1, I went through three of my take-aways from Sangeet Choudary’s new GenAI book, titled Reshuffle. Here is the last one.

Take-Away 4: GenAI Will Transform How Firms Coordinate Knowledge. And Will Create Brains for Organizations.

I’ve been talking about how GenAI will dramatically increase coordination between a firm and other parts of an ecosystem. But coordination also happens within an organization. That is why every business has tons of meetings, phone calls, and reports. That is the process of internal coordination.

Well, GenAI is also transforming this as well.

Especially in businesses that are knowledge based. Such as software, publishing, consulting, law, medicine, teaching, etc. These are organizations where people with lots of specialized knowledge work together. And a lot of what they do during the day is share their knowledge with each other. Hence all the meetings, reports, feedback, etc.

Sangeet breaks such knowledge coordination within a firm into three steps.

  • First, it is encoded.
  • Then it is classified.
  • And then it is deployed.

An example of encoding is when consultants finish their projects and put all their frameworks and analysis into internal files. To be used by consultants on other projects.

Classification is when the documents are put into the right files. So, they can be later found. Pretty much every business still has file cabinets. That’s classification. In fact, a lot of administrative jobs are about classifying various types of information. Legal, HR, accounting, project history, etc.

Deployment is when you pull the files and past reports. Knowledge deployment is also why there are so meetings, emails, and chats.

Sangeet call all this a big internal coordination tax. We could also call it an internal coordination cost. And coordination is something that GenAI can really impact. Especially in knowledge-based businesses.

For example, GenAI can automatically encode all emails and reports. It can literally listen in and transcribe every meeting and phone call. And it can automatically classify this appropriately. GenAI will be constantly gathering and classifying individual and team-based knowledge.

Then it can deploy this knowledge into new projects, teams and meetings as needed.

For agile teams, GenAI will be like an extra member at the table. Who happens to have all the knowledge of the organization. And it can also share knowledge between teams. And really across the entire organization. GenAI will start to look like an organizational brain.

And this really sounds like an optimal organization structure.

  1. You will have much more effective autonomous teams. That’s great.
  2. But you will also have effective coordination between teams.
  3. And better coordination from the top down from bottom up in the organization.

And this type of coordination is a challenge for most businesses when they get large. Managing hundreds of teams running projects is increasingly difficult.

Sangeet argues that “AI solves the coordination tax at scale”. It can read tacit and unstructured information across the organization. It can classify and deploy it. And it can do this within a firm’s specific language and behaviors.

Here is my take-away:

  • GenAI is potentially an “organizational brain” that is constantly gathering and classifying knowledge within an organization.
    • Over time, it will get smarter.
    • Over time, it will expand into more types of workflows. Which will make it both bigger and smarter.
  • GenAI will also change how teams work. It can deploy knowledge to teams and help them make faster and better decisions. Agentic execution should enable new decision architectures
  • Finally, GenAI will enables new external coordination. Internal aspects of the organization can coordinate with external parties.

This is really interesting to think about.

Technology advancements have been very effective at dropping coordination costs in industries. Which has enabled firms to do things externally instead of internally. Think platform business models.

But previous technologies didn’t change how we “coordinate knowledge” internally. And internal coordination is one of the biggest problems for businesses at large scale.

***

Those are my four main points. I have some other take-aways below. But these are just extra if you’re curious.

Take-Away 5: Coordination Can Be Broken Into 5 Steps

Sangeet provides a good framework for thinking about coordination – and where GenAI will impact it. He outlines 5 steps:

Step 1: Create a Common Representation

You need for a common view of a situation. Say a factory. Or a construction project. The interacting parties need to have a common understanding of a situation or environment. You need a common set of information. So GenAI starts by rebuilding fragmented and tacit data from multipole parties (and other sources) into a unified representation.

Step 2: Make Decisions

Now with a common representation, you can look for information and patterns within in. You can separate the signal from the noise. And GenAI can enable unified decision making across lots of actors. You can get everyone to agree what they are going to do.

Step 3: Execution

This can be a transaction. It can be a construction project. But we need GenAI enables to create a unified execution across disconnected workflows

Step 4: Composition

This is about getting all the different players to work together.

Step 5: Governance, Learning and Adaptation

This is about getting feedback. And getting smarter. And making sure actions are consistent with decisions.
This is a pretty good breakdown of coordination.

And you can start to see where GenAI is going to have a big impact (and not). Definitely Step 1 is important. This is where GenAI can gather information from lots of parties using different data and systems. From different parts of the company or industry or ecosystem. And it can create a common representation.

Take-Away 6: Tasks Can Be Broken Into 3 Types of Value

This was another interesting framework. And it’s a good way to answer the common question:

Will AI take my job?

Jobs are a collection of tasks. And Sangeet breaks the value of a task into 3 buckets:

1.Customer / User Value

What is the utility of the activity? How is it valued by the user? Books have high customer value (i.e., readers really like them). But you can’t charge much for them. Sangeet calls this intrinsic value.

2.Contextual Value

Does the task impact a wider system? In what context does it take place?

The value of security guards depends on what they are actually guarding. Security guards at a parking lot have less contextual value than security guards at a jewelry store.

Similarly, a notary that signs off on basic government documents is not that valuable. But a notary that signs off on a big business deal has big contextual value.

You can see that GenAI is going to have a big impact on contextual value. Because it changes how certain tasks interact (i.e., coordinate) with wider systems.

3. Economic Value

This is about how much you can charge for something. This is what people are usually thinking about for this question. As mentioned, books have high user value but little economic value.

Sangeet says economic value follows from scarcity, risk and/or coordination.

For example, doctors get paid a lot because they are scarce (relatively). In fact, lots of highly trained professionals get high fees because of their scarcity.

Anesthesiologists, in contrast, get paid a lot because they are both scarce and take on significant risks. They are the one that keeps the patient under and alive. As Warren Buffett says, nobody argues price with their heart surgeon.

Consultants actually do well because they can get a complicated organization to start operating in coordination fashion (around a specific project or goal).

You can see GenAI is wreaking havoc with scarcity. It is making entire domains of knowledge common and available to anyone. So, professionals that are mostly providers of knowledge (like certain types of consulting) are dropping in economic value. But professionals who take on risk and take an active role in coordinating behavior do ok.

Take-Away 7: GenAI Is a Lot About the Changing Relationship Between Tech Tool Providers and Customer Solutions Providers

Sangeet had some nice thinking on this. Basically, we see businesses adopting new GenAI tech tools from companies such as OpenAI and Google. That’s powerful. And it can also be a trap for businesses that provide customer solutions.

Customer solutions providers (such as lawyers, accounts, SaaS businesses) are all adopting these GenAI tools. They have no choice. The tools will give them a big boost in their performance. We see content creators and advertising companies using image generators. We see design firms using new generative design tools. We see law firms using ChatGPT. We see software providers using Cursor for coding. It’s a huge boost in both productivity and effectiveness.

But it also puts them at risk. They can easily become dependent on the tool. And they are basically training the tools in their knowledge. At an exponential scale. These tools are not just gaining their expertise. They are also then enabling lots of other businesses to do what they do.

So, you really don’t want to outsource your core activity to a GenAI tool provider. If it’s something like accounting that’s fine. But if it’s your core activity, this is a major problem.

Because here is what is going to happen.

  • The tech tool companies are going to get better and better. Thanks to the solutions providers using their tools. They will get smarter and better. And their rate of improvement will be much faster than the customer solutions providers.
  • They are amassing power. They are increasingly going to be able to shape and architect industries. Watch for GenAI powered law firms to start to rewrite how law is practiced.
  • They are then going to expand horizontally into more tasks and workflows. They will get involved in more parts of the solution provider workflows.

Overall, power is going to shift from ownership of the customer to the performance layer. They will absorb your expertise. And commoditize your skills. You really don’t want to end up being just a wrapper for ChatGPT.

So, what do you do?

That depends on the industry. First, you need to decide what part of your business is going to be hit.
If it’s not core (like Starbucks), you’re fine. If it hits your core (like lawyers doing document drafting and review), you need to act. You do not want to outsource your core expertise.

So, you need to build an internal tool (maybe a good idea). Or you need to pivot to a new focus in your industry (such as lawyers focusing on high-risk situations or complicated M&A). As a professional, you want to focus on where there is economic value: scarcity, risk and/or coordination.

Overall, the evolving relationship between the GenAI tool providers and customer solutions providers is really important. As a knowledge worker myself, I have a pretty clear plan on how I’m going to take advantage of increasingly ubiquitous knowledge.

***

That’s it. Cheers, Jeff

———–

Related articles:

From the Concept Library, concepts for this article are:

  • GenAI
  • Coordination and Transaction Costs

From the Company Library, companies for this article are:

  • n/a

———

I am a consultant and keynote speaker on how to accelerate growth with improving customer experiences (CX) and digital moats.

I am a partner at TechMoat Consulting, a consulting firm specialized in how to increase growth with improved customer experiences (CX), personalization and other types of customer value. Get in touch here.

I am also author of the Moats and Marathons book series, a framework for building and measuring competitive advantages in digital businesses.

Sign up for My Email List Here

This content (articles, podcasts, website info) is not investment, legal or tax advice. The information and opinions from me and any guests may be incorrect. The numbers and information may be wrong. The views expressed may no longer be relevant or accurate. This is not investment advice. Investing is risky. Do your own research.

Leave a Reply