AI-First Deployment & CS Resources: Public Inventory (2026)

A scan of what AI-first and agentic companies actually publish about how they do post-sales work — playbooks, deployment guides, methodology, kickoff templates, frameworks. Where there's a primary source, I link it. Where there's no public artifact, I say so — and that's often the more important data point.

The headline observation

Almost none of the AI-first companies publish a "Customer Success" playbook.

What they publish instead falls into three categories:

Engineering-flavored deployment guides — how to build, evaluate, and operate AI agents in production. Written for the customer's engineers, not their CS teams.
Adoption / transformation frameworks — high-level enterprise change-management content, usually paired with executive-buyer messaging.
Implementation accelerators for their largest customers — the "first 100 customers" programs that look very FDE-flavored.

If you're looking for the equivalent of Gainsight's Customer Success Playbook or ChurnZero's CSM materials, written by an AI-first company, it largely doesn't exist yet. That gap is itself the content opportunity — and the reason this site exists.

Anthropic

The most prolific publisher of public deployment material among the AI-first cohort.

Resource	What it is	Audience
Building Effective AI Agents	Best-practice guide for agentic system design. Used by Coinbase, Intercom, Thomson Reuters as referenced examples.	Engineers building agents
The Enterprise AI Transformation Guide	Three-step blueprint: foundation → pilots → scaling. References NBIM, Thomson Reuters, Cox Automotive.	Enterprise buyers & program leads
Building Trusted AI in the Enterprise (PDF)	E-book covering planning → deployment.	Enterprise leaders
Claude for the Financial Industry — Practical Deployment Guide (PDF)	Vertical-specific deployment playbook including adoption phases.	Financial services buyers
The Complete Guide to Building Skills for Claude (Jan 2026)	32-page guide on building skills/agents for Claude.	Engineers
Demystifying Evals for AI Agents	Methodology for designing and running evaluation frameworks.	Engineers, DS-types

Posture: technical, transparent, written for builders. No "Customer Success Manager" in the audience targeting. The phasing in the financial deployment guide (Foundation → Pilot → Scale) is the closest thing Anthropic publishes to a CS playbook — and it's still primarily an engineering blueprint.

Why it matters: Anthropic's resources are the gold-standard reference. Their evals piece is particularly load-bearing for any "how to define go-live for AI products" work.

OpenAI

Less publicly available CS/deployment material than Anthropic — most enterprise content is gated behind sales conversations or DeployCo (the new $4B PE-backed subsidiary, May 2026).

Resource	What it is
OpenAI Cookbook	Repository of working code examples for building with the API. Technical, engineer-focused.
GPT-5 / Agentic best practices (docs)	Generic API and agent best practices. Engineer-focused.
DeployCo public posture (May 2026)	$4B PE-backed subsidiary with embedded engineers. No public methodology yet — reference customers via case studies.
Klarna case study	High-profile reference: Klarna's AI assistant doing the work of 700 agents.

Posture: more sales-led than Anthropic. The public artifact is the customer case study; the methodology stays internal.

Palantir

The original FDE/Deployment Strategist model. Surprisingly, Palantir publishes more public methodology than most AI-first companies — because their model is mature and they're recruiting against it.

Resource	What it is
A Day in the Life of a Palantir Forward Deployed Software Engineer	First-person account of the role and process.
A Day in the Life of a Palantir Deployment Strategist	First-person account of the DS role ("Echo").
Foundry Program Overview	Official methodology for the customer program structure (CoE, governance, roles).
Foundry Development Lifecycle	Phased build methodology.
Foundry Technical Overview (PDF)	Architecture and capabilities document.
Guides and Workflows	Official cookbook for common implementation patterns.

The Echo/Delta terminology: Palantir uses internal codenames — Echo = Deployment Strategist (product-manager flavor), Delta = Forward Deployed Software Engineer (technical flavor). Both roles blur into PM/eng/strategist hybrid in practice.

Center of Excellence (CoE) handoff: Palantir publishes the maturity curve explicitly — Palantir's team builds and runs in production first; then helps the customer stand up an internal CoE; then transitions ownership. This is the "champion-out" pattern that lets FDE scale sub-linearly.

Sierra

Smaller public footprint, but their pricing-as-positioning material is the canonical reference for outcome-based AI agent commercials.

Resource	What it is
Outcome-Based Pricing for AI Agents	Sierra's public framing of why they price on resolved outcomes.
Customer stories	WeightWatchers, SiriusXM, Sonos, ADT — all with outcome-volume framings.
Agent OS	Sierra's developer-facing platform for building/optimizing agents.

Posture: "we built the canonical agentic product." Methodology is implicit in the pricing model rather than explicitly published.

Decagon

Resource	What it is
Customer Stories	Chime, Duolingo, ClassPass, Hunter Douglas.
Pricing transparency	Per-conversation + per-resolution hybrid.

Notable customer outcomes:

Chime: 70% chat and voice resolution
Duolingo: 80% deflection rate
ClassPass: 10× deflection increase
Hunter Douglas: $1M in revenue attributed to AI-handled conversations

These are the most-quoted reference numbers in the AI agent category. Worth knowing by heart.

Cresta

Published more operationally-focused material than most.

Resource	What it is
Cresta AI Agent / Agent Operations Center	Live oversight model — AI conversations monitored by Cresta tooling, flagged for human intervention.
Decagon vs Sierra vs Cresta Buyer Guide	Cresta's own competitive positioning document.

Notable: Cresta is closest among the AI agent vendors to publishing CS-style operational content. Their Agent Operations Center is essentially a "post-go-live" framework with a public name.

Dust, Mistral, Glean

Smaller public footprints. The most informative artifacts are their job postings, which reveal the staffing model:

Company	Most-informative public artifact	What it tells you
Dust	Founding AI Deployment Strategist, Post-Sales	Explicit Post-Sales label; DS hybrid role.
Mistral	AI Deployment Strategist, Cybersecurity (Paris, April 2026)	Vertical-organized DS function.
Glean	Customer stories + Gleanvocates community	Best-documented community-led adoption play among AI-first companies.

Salesforce Agentforce — the hybrid case

Salesforce is the most prolific publisher of agentic deployment material among legacy SaaS companies — because they have the most to defend.

Resource	What it is
From Pilot to Playbook: First Year Using Agentforce	Salesforce's own "Customer Zero" account of deploying internally.
Agentforce Contact Center 100 program	The first 100 enterprise customers get embedded engineering + executive support to scale fast.
Get Ready for Agentforce (Trailhead)	Salesforce's training path.
Agentforce in Action: Customer Success Stories	Reference customer outcomes.

Notable: Salesforce explicitly used forward-deployed engineers for the Agentforce Contact Center 100 program — a hybrid SaaS company adopting the AI-first deployment model for their AI-first product line. That's the most direct evidence that the FDE model is migrating into legacy SaaS, not just defining new entrants.

Intercom — Fin / hybrid case

The richest public CS-adjacent content from any AI-first-leaning company.

Resource	What it is
2026 Customer Service Transformation Report	Survey of 3,000+ leaders, role evolution, AI adoption data.
Inside the AI-First Support Team	The 4 new roles: AI Ops Lead, Knowledge Manager, Conversation Designer, Support Automation Specialist.
The AI Deployment Gap Is Widening	Why CS leaders are losing time.
How AI Is Evolving Support Careers	Practitioner perspective on role transitions.
Automating 81% of Customer Service	Specific customer transformation account.

This is the closest thing in the industry to a published "CS in the AI era" playbook. Intercom has both the incentive and the product to take this position publicly. Worth reading in full.

The pattern across the industry

If you organize the public CS material by company type:

Company Type	What they publish about CS/deployment
Pure AI-first model providers (Anthropic, OpenAI)	Engineering deployment guides, eval methodology, enterprise transformation frameworks. Not CS playbooks.
AI agent vendors (Sierra, Decagon, Cresta)	Pricing-as-positioning, customer outcome case studies, Agent Operations frameworks. Light on CS playbook content.
Enterprise platform vendors with FDE practice (Palantir)	Mature role descriptions, methodology (CoE handoff), program structure. Most public methodology.
Legacy SaaS adding agentic (Salesforce Agentforce, Intercom Fin, ServiceNow)	The most CS-flavored content. "How to transition your CS team to AI" is their content wedge — because their customers are exactly asking this.
AI-native SMB/mid-market (Dust, Glean)	Job postings reveal staffing model; public case studies show outcomes.

Templates and downloadable resources — what actually exists publicly

Most "templates" you'll find from AI-first companies are:

Eval rubric examples — Anthropic's evals guide, OpenAI's evals documentation.
Pricing calculators — Sierra and Decagon publish per-resolution math.
Implementation accelerators — gated, customer-only (Salesforce Trailhead is the most accessible exception).
Onboarding flows for the model itself — API quickstarts, not customer-onboarding playbooks.

What's notably absent publicly

Pre-kickoff customer intake forms
Deployment SOWs for AI agents
Eval co-design templates (customer-facing)
Outcome-attribution worksheets
Renewal review templates for outcome-priced products
Quality regression incident playbooks
Org-design frameworks for AI-first post-sales

This is the template inventory backlog. Each one is a downloadable artifact that doesn't exist anywhere else publicly — and that we'll be releasing here, one at a time, alongside this working notes series.

Full source list

This inventory will be kept updated as new resources are published. If you spot a missing one — or run post-sales at an AI-first company and want to add yours to the list — let me know.

What AI-First Companies Actually Publish About Deployment & Customer Success — A Public Resource Inventory