Tejas is a system of 129 AI agents organised across 11 specialist teams — covering marketing, sales, product, engineering, design, advisory, and operations. It is operated personally by two founding partners and functions as an AI-native growth department for mid-market companies. Tejas replaces the need to hire multiple senior executives or separate agency retainers across functions.

Is the 90-Day Growth Blueprint really free?

Yes. The Blueprint is fully funded by Appetals Solutions. It includes a Diagnostic Audit, a half-day Leadership Bootcamp, and a 90-Day Action Plan — all at zero cost and with no obligation to proceed to a retainer.

Who actually does the work — the partners or AI agents?

Both. Ishwar Jha designs the operational strategy and operates every engagement cycle personally. Tejas — 129 AI agents across 11 specialist functions — executes the operational workflows under his direction. Neither delegates down.

Are we too small for this?

The Blueprint is designed for companies between ₹50 crore and ₹500 crore in revenue — mid-market businesses where growth has stalled and the board or investors are watching results every quarter.

What does performance-linked pricing mean?

Retainer pricing scales with your results. You do not pay a fixed monthly fee regardless of output. If Tejas misses targets, you pay less. Tejas earns when you grow.

How is Tejas different from hiring a CMO or an agency?

A CMO hire takes 4–8 months to recruit and 3–6 months to ramp, at ₹1.2–2.4 crore per year before results. Three separate agencies cost ₹80L–1.5 crore per year and take months to align. The Blueprint gives you a running growth system in 90 days at no upfront cost, operated by a founding partner with 30 years of experience.

Do we need an internal team to make this work?

No internal build-out is required. Tejas operates as your growth function. We handle execution; you handle decisions and approvals.

What happens after the 90-Day Blueprint?

Founders who want to continue move to a Growth Acceleration retainer — a performance-linked engagement where Ishwar continues operating the system on your behalf across five playbooks: growth stack architecture, sales system engineering, demand generation, retention operating system, and investor narrative and board reporting.

Chapter 1: What Lumen Does

Lumen is a Claude Code plugin that orchestrates 18 specialized AI agents across six product management workflows. You run a command, answer a few context questions, and get a structured, evidence-graded report.

That is the whole idea. No new tool to learn. No dashboard to maintain. You type in the same terminal where you already work.

The Problem Lumen Solves

Most PMs know what good PM work looks like. Discovery interviews, PMF scoring, prioritization frameworks, experiment design, and launch readiness audits. The frameworks are not the hard part.

The hard part is doing all of it rigorously under time pressure, without a research team, without forgetting a step, and without your recommendations looking like gut calls dressed up as analysis.

Lumen runs the framework. You answer the questions and make the calls.

Six Workflows, Six Product Problems

Lumen maps to six distinct problems that come up repeatedly across a product's life.

W1: PMF Discovery (/lumen:pmf-discovery) You have a product and users. You need to know if you have product-market fit, which segments have it, which don't, and what to do about it. This is the starting point for most Lumen users.

W2: PMF Recovery / Churn Crisis (/lumen:pmf-recovery) Something has broken. Churn is up, activation is down, or a core segment is slipping. W2 runs a crisis workflow: diagnoses the cause, classifies churn type, designs an intervention, and produces a leadership briefing.

W3: Quarterly Strategy (/lumen:strategy) It is the start of a quarter. You need a North Star metric, a cascaded set of OKRs, a prioritized roadmap, a pricing model, and a narrative your team can actually use. W3 produces all of it in one run.

W4: Feature Validation (/lumen:feature) You have a feature idea. You need to know whether to build it, buy it, or test it first. W4 includes an AI ethics checkpoint, an experiment design, an interaction spec, and a decision memo.

W5: GTM Launch (/lumen:launch) You are launching a product, a new tier, or a new market. W5 audits readiness across seven dimensions, writes launch messaging for every audience, and produces a Day 1/7/30 execution plan.

W6: Churn Analysis (/lumen:churn) NRR is soft. You need to decompose it, find the at-risk accounts, rank them by ARR, and design win-back campaigns. W6 does that and sets up 30/60/90-day outcome tracking.

What Comes Out

Every workflow produces a structured report. Reports include:

A summary of findings
A set of recommendations
An evidence quality grade (HIGH / MEDIUM / LOW) for every section
A decision memo from DecideWell (for workflows that require one)
An experiment brief, roadmap, or launch plan, depending on the workflow
A Limitations section that tells you exactly what was missing

The evidence quality grade is the most important part. It tells you how much data backs each conclusion. A LOW-evidence recommendation is still useful. It just tells you which gaps to close before acting.

What Lumen Is Not Trying to Do

Lumen does not replace customer conversations. We have DiscoveryOS agent that synthesises interview data, but you still conduct the interviews.

Lumen does not tell you what your product should be. It tells you what the data says about what your users need, and it runs the frameworks to help you decide. The judgment calls are yours.

Lumen does not integrate with your project management tool. Reports live in the Claude Code session and, if Supabase is connected, in the Knowledge Graph. They do not automatically create Jira tickets.

Why It Runs Inside Claude Code

Lumen requires four things that only Claude Code provides: slash commands, agent files read at runtime, MCP server connections, and local file system access.

Claude.ai chat and Claude for Teams are conversation interfaces. Claude Code is a programmable agent runtime. Lumen is built for the latter.

If you have not installed Claude Code yet, start there. The next chapter walks through everything.

We also found Claude Code to be a powerful tool providing the capabilities perfectly fit for running Lumen without the hassle of setting up hosting, configuration, etc.