Road to Agentic Era

A Weekend View of what the path to agentic could look like for Large Organizations

Last year I had liquidated all my Indian IT Stocks looking at the way Each one has been reacting to agentic Trends. I was dissapointed by all of them.

This weekend I was doing a review of IT Companies in India to check if I should revisit and still found everyone seems to be stuck in the old factory mindset. If this continues then this whole Industry is going to crash badly very soon. So what could be the alternative… What am I looking for?

The Lack of Thought Leadership

Fundamentally I’m looking for at least one leader to accept that the existing structure is broken in the agentic era and articulate how they intend to reorient but all I see is copilot english with no clear direction.

So what could be one of the option.. I’m jotting out a top of mind account of what I have in mind. Obviously you could have alternatives. Happy about that.. Go ahead and blog about it please. Hopefully the Leadership folks get some good ideas to chew.

So here goes what is running in my head

Phase 1 : Envision the North Star

Agentic Solutions are still evolving but the key patterns have already started to emerge and show maturity. Addressing them with the old factory model of just throw some FTEs wont work. Fundamentally the whole reporting interms of Financial Statements need to be augmented with “Value Delivered Statements”.

Let me illustrate

If the company has executed a project and it has saved the business with Sales Improvement of $X then that needs to be consildated and reported as part of the “Value Statement”. Right now the focus is on Sales (Booking and Revenue) which do not give the full picture in terms of the health of the accounts. It only says that the company is getting or going to get paid but not confident that the services rendered make sense for the account! In the agentic era with AI being able to track performance and re-iteratively think about alternatives, Third parties have a new competitor. Without the Value statement there is no reason that the Account Leadership turns pragmatic and starts questioning the purpose of the partner being engaged! This as a Investor makes me to know if my money going to the right place.

The Value Delivered Statement would vary and have multiple dimentions nad would be a topic for a completely new Blog but I will stop at this level for now.

Phase 2 : Determine the Key Value Engine of Your Organization architecture

The Organization Stucture’s Architecture is a key deciding aspect of whether the Value to Customer is Understood, Measured, Articulated and Tracked. If every team is going to do its bit at random then good luck in getting a corporate level statement. Instead The Value Statment needs to be owned by a Single Team but streamed via multiple support teams.

Let me explain: Lets take an example of the following Organization Architecture… Your structure may vary and you would need to figure out how to extract this in your structure.

The SS Team (Sales & Solutioning)

These could be Bunch of Vertical Sales with Sales Support inbuilt with Delivery across all People, Cloud and Independent Patterns. They are directly responsible for the Booking Numbers.

The Infrastructure Teams

Infrastructure Team Owning Indirect Sales for Independent Technology that is independent of Vertical (Network, Cloud, Database, etc) This could in turn be of two types

Internal Billed Team : Not billed directly to customers [Overhead]
External Billed Team : Billed Directly to Customer [Have a Margin Target]

The Product Teams

Shared Products which have Process that span multiple verticals (Accounts Payable, Payment Management, Infrastructure Management, DevOps, Cloud Ops, CX Ops etc.) [White labeling would classify and treat team as Ecosystem than Core SS] These teams have indrect Sales Targets on Booking and Revenue

The Ecosystem Teams

Consist of Sales responsible of influencing the Hyperscale ecosystem or External Product Ecosystem through Training, Solution Support, R&D Support and 360 degree relationships. Again theyu have indirect Sales Revuenue Targets and Margin Targets as they would be owning Expenditure to run the products.

Other Support Systems

HR, Recruitment & Retirement, Finance, Admin etc. [All Overheads]

Now here the Value Articulation and Accounting would obviously be owned by the SS Team with Updates from the directly related teams working to infuse Agentic AI to deliver services faster, at lower price points and more importantly impact the customer Revenue and Expenses.

I’m stopping here as my coffee is wearing out

Phase 3 : Know thy people

I’m a bit oldschool when it comes to systems thinking and I belong to the class of people who beleive “If it cannot be measured, then it cannot be improved”.

Look at your existing workforce from a few different dimentions and you will find the assets are actually of differnt types. Each Asset will have different levels. I’m illustrating a few below:

Parameter	Levels
Knowledge Levels	Regimented(Trained), Experienced(Low, Med, High), Creators(Ad hoc, Project Level, Company Level, Country Level, Humanity Level)
Working Style	Regimented(Extreme Process Adherence - process first, customer next) , Flexible (Customer First process adherence), Artistic (No process Adherence)
Managerial Style	Micro Managers(Extreme Ego Maniacs), Pacifist (Flexible/Adaptive but pushy), Anarchic (No Control banana republic)

Phase 4 : Know Thy Systems

Types of Existing Data

Parameter	Levels
Usage	Realtime, Batched, Sluggy, Dark
Change	Volatile, Pin in Strawpile, Stagnant

Types of APIs

Parameter	Levels
Stone age	Good old VT3270 Green Screen,VT100 Terminals, Java/Windows Apps
Metal Age	Bit Mapped like ISO 8583, SOAP and many closed undocumented Legacy APIs
Cloud Age	REST, gRPC, GraphQL etc
AI Age	MCP, A2A, Web MCP, A2UI, AG-UI

Types of Processes

Parameter	Levels
Awareness	Well Known, Documented/tooled but not Known/used, Undocumented but Tooled and Hardly Known, Undocumented/Untooled/Unknown
Complexity	Simple, Hard, Complex, Insane
Typical duration	Realtime, Near Realtime, Batched, Days, Months, Years
Effort Scale	Mobile, Server, Human, Gov, Civilization
Impact Scale	Person, Group, Company, Region, Country, Civilization

Phase 4 : Know Agentic

In the Agentic era the Agentic Workforce Could be of the following types:

Parameter	Levels
AI Free	Hardly use AI, AI also cant impact (Hard Labor.. for now till robotics mature, doctors till telemedicine becomes normal, Politics , Business Management, Entrepreneurship, Entertainment)
Builder	Build AI agents and related Agentic Systems
Orchestrator	Orchestrates Multiple AI Agents as part of Job
Assisted	Uses AI to assist in Job but still owns the job primarily

My coffee is really running thin so I’m going to parapharase the further phases for now… Will Detail in future weekends…