01 / 05
Board Review 2026
OGSM AI
Review
Compounding gains across Cognitive, Generative, Agentic & Decisioning AI
Decisioning in Action
Airline Disruption Recovery
Weather event at DFW — 14 flights impacted, cascading crew & gate constraints
Flight Route Status
CF 1247
DFW → ORD
Disrupted
CF 892
DFW → LAX
Disrupted
CF 3051
DFW → JFK
Disrupted
CF 610
DFW → MIA
Disrupted
✓
Crew Analysis23 pairings across 3 hubs
✓
Gate ReallocationORD, LAX, JFK optimized
✓
Pax Rebooking1,840 re-routed, tiers honored
✓
Cost Optimization$2.1M saved vs. manual
✅
4.2m
Resolution
$2.1M
Savings
0
Crew Violations
14/14
Flights Resolved
⚙️
Initializing decisioning engine...
Progress
Our Progress from Last OGSM
Winning the sprints while staying focused on the marathon.
Strategy
Bimodal: Cognitive + Generative AI
→ Cognitive + Generative + Agentic + Decisioning
LLM Landscape
GPT-3 was coming of age
→ Now GPT-5, Opus, and beyond
People Trained
300 people trained in AI
→ All Employees (SPARK), All Developers (GitHub Copilot), All CSG (M365 Copilot)
Client Projects
75 Clients (100 Projects)
→ All projects — no project without AI element
Partnerships
Primarily Hyperscalers
→ + HPE/Nvidia, Arize, Cursor, Anthropic, Poetiq
CoE Lab
Lab was created
→ Producing Agentic AI artifacts & training other teams
Software Supply Chain
Quasar introduced
→ Unified stack: Quasar, ForgeX, EvolveOps, Cosmos, AIVA
Experimentation
Exploring Small Language Models (SLM)
→ SLMs shipped, Agentic workflows everywhere, Edge inference
Landscape
How the World Has Changed
The technology race is continous, the execution race has begun.

It Is All About Journey, Not The Product

  • Core inference will continue to get cheaper, but it is about the ecosystem
  • It is not just about automating task, it is about building new composable services
  • With rapid innovation, clients are looking for trusted partners

AI Innovations Outpacing Adoption: 95% Pilots Are failing

  • Most enterprises still struggling to move beyond AI pilots
Alex Karp (Palantir) — “Biggest adoption failure is 'LLM-on-the-stack' — buying off the shelf models and expecting regulated, precision work”

Agentic AI: A Fundamentally Different Architecture

  • From single-task inference (request/response) to Agentic AI chains — multi-step reasoning, tool use, context and memory
  • Agent Lifecycle & Orchestration is the new middleware — Human-in-the-loop → Human-on-the-loop → Human-out-of-loop
  • Agents becoming enterprise-ready and personal first-class citizens (OpenClaw — 289K GitHub stars)

Client Expectations: Hype Cycle → Operating Model

  • Clients buying outcomes instead of capacity
  • Technical blocker is data, not models — 80%+ of enterprise data sits in legacy
  • Client spend increasing, but growing void for talent
Learnings
What Are Key Takeaways
AI brings new value pools and opportunities.

Value Pools are Up

  • Expanded demand across new and redefined roles such as AI Engineer, AI Governance, Red-teaming, Tech-debt reduction, etc.
  • A completely new operational stack around agents with Agent Creation / Orchestration / Monitoring / Management — pools that didn't exist 12 months ago
  • Build repeatable platforms and practices — because clients don't yet have internal muscle memory for any of it.

Blend Functional Domain with AI

  • Knowing client business workflows, processes gives us opportunity to be proactive and be first to engage
  • Functional expertise guides model selection, prompt engineering, and workflows to deliver business value
  • Regulated industries need AI governance, transparency, auditability and other gaurdrails from their own data and processes

Client Intimacy Matters The Most

  • Knowing Client Systems and Application - Gives a leg up for Proactive Proposals
  • Clients will only trust AI when it is grounded in its own properietary workflows, decision login and tacit knowledge

Unify Assets and Platforms

  • Create a unified asset and IP platform that monetizes progressively
  • Land with core asset and then expand through adjacent IP and capabilities
  • Create an Agentic model operating system for clients based on knowledge graph
Forward Strategy
OGSM Going Forward
Compounding gains with Workforce, Assets & IP.
Create New Workforce

AI as a "Religion"

SPARK, GLT (Agents), CSG (Copilot + SFDC), SPT (RFI/RFQ/Proposals)

Hybrid Forward Deployment Engineers

Grounded in first principles

AI Engineers and AI Test Engineers

GHCP Engineers, Claude Code / Anti Gravity Engineers, Specialized in Test Validation and Quality

Agentic & Responsible AI Engineers

Building and governing the next generation of autonomous systems

Create New Differentiated GTM

Legacy Rewrite

Modernization & data engineering (prepping data for AI). Rebalancing tech ledger from debt to value is an outsized opportunity for Coforge

BIG BET

Custom Off The Shelf Software (COTS) To Custom Software

Most Enterprises use only 50% of COTS but pay huge in license, maintenance and customization costs.

BIG BET
Thank You
OGSM AI Review 2026
← → or scroll to navigate