Author
Shikhar Mishra
Cofounder and CEO, EGI
Shikhar Mishra is a cofounder and the CEO of EGI, the company behind Bruce.
Shikhar was VP of Engineering at MainStreet and at Scalable, cofounded the ML hiring platform Simppler, and built products at Eventbrite and World Wide Technology.
Shikhar holds an MS in Mathematics and Computer Science from the University of Missouri, St. Louis, with research in genetic algorithms.
Outside work, Shikhar makes time for the small things in life, and has been a Warriors fan for life, through every season.
Posts by Shikhar Mishra
19 posts
While you debate agents, operators are compounding
The window is open. Fence-sitters lose quietly. Operators compound quote speed, follow-up, and oversight while others stay in the pilot meeting.
Industry
The future of agent UX will not be chat-first. It will be intent-first
NL-native, not chat-only. One session, three surfaces — finance, purchasing, and GTM with agent-driven GUI escalation and human commit at the boundary.
Architecture
Meeting booked. Now the real work begins
Jordan Greene replied on email six. Bruce runs Harbor Goods from Calendar discovery through Sheets close — Google Workspace, not a parallel stack.
Deals
The anatomy of effective outreach: why BDRs should think like Netflix
Reply on email six, not email one. Bruce ranks person-person, business-business, and persona-pain similarity — then runs outbound autonomously in Gmail.
Pipeline
Why long context is hard for agents
Long context is working memory, not storage. Why agents struggle with relevance and time — and why converging frontier models make architecture the bottleneck.
Architecture
The command pattern
Heads of Sales and Customer Success run teams from Command — shared reality, automation on the motion, leader judgment on exceptions only. Manage by exception, not inspect everything.
Leadership
Targeting beats messaging
For net-new pipeline, targeting effectiveness dominates messaging personalization — across SaaS verticals, trade distributors, retail, CPG, and services businesses. Who you reach matters more than copy polish.
Pipeline
GTM build vs buy
Teams building GTM tooling underestimate maintenance — agents, guardrails, eval, observability. Build where product is unique. Adopt where GTM is standardizing.
GTM
Why enterprise AEs dread cold calling
The cold call tax — psychological load, admin grind, and prospect pain. AI-native outbound removes the grunt layer with real-time briefs: stack, objections, and peer KPIs per persona.
Pipeline
MEDDPICC before discovery
Call-time MEDDPICC extraction is backwards. AI-native GTM pre-contextualizes deal intelligence before the first meeting — discovery validates hypotheses, not extracts them.
Deals
White-gloved at scale
White-gloved CS at 1:60+ account ratios: personalized motion on every account — via AI-native awareness, automation, and empowerment, not more headcount.
Success
POC is the new funnel
Enterprise buyers commit in proof-of-value, not at demo. Why the funnel's center of gravity moved to POC — and what GTM must measure and execute in evaluation.
GTM
What is an AI-native GTM motion?
AI-native GTM is not just automation. It is unified awareness, policy-driven execution, and seller empowerment — awareness + execution + human leverage.
GTM
Execution leakage
We modeled a $100M distributor. 300–700 bps of EBITDA leave through the desk. Ten places. None of them are a line item.
Industry
RevOps from CRM police to GTM architects
How RevOps teams are evolving from CRM administrators enforcing data hygiene to strategic GTM architects driving revenue operations with AI-powered autonomous systems.
RevOps
Multi-agent orchestration: building reliable, auditable GTM systems
How Bruce coordinates specialized agents to handle complex GTM workflows with reliability, auditability, and operational excellence.
Architecture · with Chien Nguyen and David Shugert
CodeGen for RevOps: generating production Python from natural language
How Bruce generates and executes Python code for RevOps analysis. Natural language queries become production-ready code for forecasting, pipeline analysis, and deal velocity.
RevOps · with Chien Nguyen and David Shugert
ReAct: running the full GTM cycle as a closed-loop execution system
How ReAct-style agents can run a full GTM motion end-to-end as a self-regulating system, not as automation.
Architecture · with Chien Nguyen and David Shugert
RAG vs neuro-symbolic execution: building AI systems that can execute reliably
RAG limitations and how neuro-symbolic execution enables reliable AI systems for operational tasks.
Architecture · with Chien Nguyen and David Shugert
