X-GenAI

AI automation · Amsterdam → worldwide

Your team has better things to do than copy-paste.

X-GenAI builds custom AI agents and workflow automation, the kind that keeps working after the demo ends.

15 minutes, no pitch deck. Just bring your most annoying workflow.

AGENT ORCHESTRATOR · SUPPORT_TRIAGE_V1

RUNNING

TRIGGER

ticket.created

AI AGENT

classify · route

AUTO-REPLY

docs match

ESCALATE

human review

CRM SYNC

write-back

TICKETS TODAY

1,042

AVG RESPONSE

412ms

ESCALATION RATE

0.18%

ticket #1042 received · classifying intent

matched to docs · replied in 4s

ticket #1043 flagged · refund, needs sign-off

Illustrative run: shows how X-GenAI structures an agent pipeline, not a specific client's live data.

One agent · one Tuesday morning

While your team grabbed coffee before standup

  1. 8:4114 support tickets came in overnight

    sorted and prioritized before anyone logged in

  2. 8:4711 routine ones answered

    straight from your docs, in your tone of voice

  3. 8:52One refund case set aside

    policy says a human decides, so a human will

  4. 8:58Summary posted to #support

    one tidy message instead of forty pings

By the time the coffee kicked in, the queue was already empty. Nobody noticed. That's the point.

Model-agnostic by design. Built with:

Claude
OpenAI
Gemini
Mistral AI
Meta AI
Hugging Face
LangChain
PostgreSQL
Supabase
n8n
Zapier
Make
NVIDIA
Python
TypeScript
Docker
GitHub
Vercel
Claude
OpenAI
Gemini
Mistral AI
Meta AI
Hugging Face
LangChain
PostgreSQL
Supabase
n8n
Zapier
Make
NVIDIA
Python
TypeScript
Docker
GitHub
Vercel
FIXED SCOPEFLAT PRICEGDPR BY DEFAULTHUMAN REVIEW GATESREMOTE BY DEFAULTAMSTERDAM → WORLDWIDEFIXED SCOPEFLAT PRICEGDPR BY DEFAULTHUMAN REVIEW GATESREMOTE BY DEFAULTAMSTERDAM → WORLDWIDE

What would your week look like if the boring parts ran themselves?

What the research says

The shift to AI automation, in numbers.

What research firms and the companies building this technology are finding, on where AI automation is already paying off.

92% of companies plan to increase AI investment over the next three years, but only 1% call their deployment mature.

· McKinsey, Superagency in the Workplace, Jan 2025

Companies that align AI, platform, and business strategy see 2.2x the revenue growth of their peers.

· Accenture, Platform Strategy in the Age of Agentic AI, Dec 2025

82% of CEOs are now more optimistic about AI's ROI than they were a year ago.

· BCG, AI Radar 2026

Median time-to-value on agent deployments is about 5.1 months.

· BCG · Forrester, 2026

Only 25% of AI initiatives deliver their expected ROI, and just 16% have scaled enterprise-wide.

· IBM Institute for Business Value

88% of enterprises say AI has already increased annual revenue in some or all parts of the business.

· NVIDIA, State of AI 2026

NVIDIA agreed to acquire Hugging Face — 18M+ developers, 3M+ models, 200K+ enterprise customers — for about $13 billion.

· NVIDIA / Bloomberg, Sept 2026

Anthropic's annualized revenue run rate hit $65 billion in July 2026, up from $30 billion just three months earlier.

· Anthropic, via CNBC, Aug 2026

52% of executives say their organization has already deployed AI agents in production.

· Google Cloud, ROI of AI Study 2025

Google shipped Gemini 3.8 Flash, its third Flash model in six weeks, built for coding and agentic tasks.

· Google DeepMind, Sept 2026

Meta's business AI tools went from 1 million to 10 million weekly conversations in a single quarter.

· Meta, 2026

Meta launched Muse, its first personal AI agent, across web, iOS, Android, and WhatsApp.

· Meta, Sept 2026

OpenAI released GPT-6 Astra, built for computer use, browsing, and software engineering, rolling out across ChatGPT, the API, Azure, and Bedrock.

· OpenAI, Sept 2026

By 2026, 40% of enterprise apps will ship with task-specific AI agents, up from less than 5% in 2025.

· Gartner, Aug 2025

$234 billion in enterprise application software spend is now at risk from agentic AI.

· Gartner, July 2026

Workforce access to AI grew from under 40% to around 60% of workers in a single year.

· Deloitte, State of AI in the Enterprise 2026

Uber burned through its entire 2026 AI budget in just four months, as Claude Code adoption spread across engineering faster than finance had modeled.

· Uber, via Fortune, May 2026

Early results from Booking.com's new agentic messaging tools showed a 73% increase in partner satisfaction over previous tooling.

· Booking.com Newsroom, 2026

ServiceNow customers running agentic AI in production grew 9x in a single quarter compared with nine months earlier.

· ServiceNow, Q2 2026 earnings

Salesforce's Agentforce reached about $800 million in annual recurring revenue by the end of FY26, up 169% year over year.

· Salesforce, Q4 FY26 earnings

JPMorgan Chase's AI initiatives are already delivering about $2 billion a year in cost savings.

· JPMorgan Chase, Jamie Dimon, 2026 shareholder letter

Amazon now runs more than one million robots across its warehouses — nearly one for every 1.5 human workers.

· Amazon, Andy Jassy, Q1 2026 earnings call

92% of companies plan to increase AI investment over the next three years, but only 1% call their deployment mature.

· McKinsey, Superagency in the Workplace, Jan 2025

Companies that align AI, platform, and business strategy see 2.2x the revenue growth of their peers.

· Accenture, Platform Strategy in the Age of Agentic AI, Dec 2025

82% of CEOs are now more optimistic about AI's ROI than they were a year ago.

· BCG, AI Radar 2026

Median time-to-value on agent deployments is about 5.1 months.

· BCG · Forrester, 2026

Only 25% of AI initiatives deliver their expected ROI, and just 16% have scaled enterprise-wide.

· IBM Institute for Business Value

88% of enterprises say AI has already increased annual revenue in some or all parts of the business.

· NVIDIA, State of AI 2026

NVIDIA agreed to acquire Hugging Face — 18M+ developers, 3M+ models, 200K+ enterprise customers — for about $13 billion.

· NVIDIA / Bloomberg, Sept 2026

Anthropic's annualized revenue run rate hit $65 billion in July 2026, up from $30 billion just three months earlier.

· Anthropic, via CNBC, Aug 2026

52% of executives say their organization has already deployed AI agents in production.

· Google Cloud, ROI of AI Study 2025

Google shipped Gemini 3.8 Flash, its third Flash model in six weeks, built for coding and agentic tasks.

· Google DeepMind, Sept 2026

Meta's business AI tools went from 1 million to 10 million weekly conversations in a single quarter.

· Meta, 2026

Meta launched Muse, its first personal AI agent, across web, iOS, Android, and WhatsApp.

· Meta, Sept 2026

OpenAI released GPT-6 Astra, built for computer use, browsing, and software engineering, rolling out across ChatGPT, the API, Azure, and Bedrock.

· OpenAI, Sept 2026

By 2026, 40% of enterprise apps will ship with task-specific AI agents, up from less than 5% in 2025.

· Gartner, Aug 2025

$234 billion in enterprise application software spend is now at risk from agentic AI.

· Gartner, July 2026

Workforce access to AI grew from under 40% to around 60% of workers in a single year.

· Deloitte, State of AI in the Enterprise 2026

Uber burned through its entire 2026 AI budget in just four months, as Claude Code adoption spread across engineering faster than finance had modeled.

· Uber, via Fortune, May 2026

Early results from Booking.com's new agentic messaging tools showed a 73% increase in partner satisfaction over previous tooling.

· Booking.com Newsroom, 2026

ServiceNow customers running agentic AI in production grew 9x in a single quarter compared with nine months earlier.

· ServiceNow, Q2 2026 earnings

Salesforce's Agentforce reached about $800 million in annual recurring revenue by the end of FY26, up 169% year over year.

· Salesforce, Q4 FY26 earnings

JPMorgan Chase's AI initiatives are already delivering about $2 billion a year in cost savings.

· JPMorgan Chase, Jamie Dimon, 2026 shareholder letter

Amazon now runs more than one million robots across its warehouses — nearly one for every 1.5 human workers.

· Amazon, Andy Jassy, Q1 2026 earnings call

How one real workflow actually runs

Input → workflow → review → result

Not illustrative — this is the enquiry-to-CRM automation built and tested for real, synthetic-data proof. See the full run →

01 / Input

An enquiry comes in

A contact-form submission or lead-gen webhook — messy, sometimes incomplete, sometimes a duplicate delivery.

02 / Workflow

Validated, deduped, classified

Schema-checked before anything else happens. Already-seen enquiries are skipped, not re-processed. The rest gets classified — with the enquiry text always treated as data, never as instructions.

03 / Review

Anything uncertain goes to a human

Missing fields, ambiguous intent, or a message that tried to manipulate the pipeline — all routed to a review queue instead of guessed into a record.

04 / Result

A clean record, or a logged reason why not

Everything else becomes a CRM record with a drafted (never auto-sent) acknowledgement. A provider failure is visible and logged, not silent.

What X-GenAI does

Services

Five ways X-GenAI helps, matched to what your team needs, not a fixed package.

001 / 005

AI Agents & Automation

Custom agents and workflow automation that take real work off your team: scoped, built, and shipped to production. You own the result.

DevelopmentAutomationsMonitoring
Let's build something
x-genai · agent

$ scope workflow --client onboarding

> 3 automatable steps found · 2 review gates

$ deploy --env production

> shipped · logging on · eval set attached

$

002 / 005

AI Consulting

Before anything gets built, what's worth automating in your workflows gets mapped, along with what each piece would cost, so budget goes where it pays off.

Opportunity mappingPrioritisationROI
Get a Discovery Audit

Opportunity map · draft

Ticket triage

20–40 hrs to build

Automate now

Onboarding emails

15–25 hrs to build

Automate now

Contract review

N/A

Human only

003 / 005

AI Education & Content

The blog and newsletter track what's changing in AI: real advancements, not hype, so you can tell the difference before you spend on it.

TrendsCertificationsNewsletter
Read the blog

This week's issue

AI trends

How to actually keep up with AI (without the doom-scrolling)

Responsible AI

The environmental cost of AI (and what gets done about it)

AI trends

What's production-ready in AI automation right now

004 / 005

Responsible & Climate-Conscious AI

The smallest model that reliably does the job, not the biggest one that impresses in a demo. Better engineering and a smaller footprint, by default.

Model efficiencyGovernanceTransparency
Read the mission

Model choice · default

Industry defaultbiggest available
X-GenAI defaultclears the eval, no more

005 / 005

AI Evaluation & Monitoring

Independent evaluation for AI systems: yours or ours, already in production or still a prototype. Benchmark suites, drift monitoring, and eval harnesses that catch a regression before your users do.

Eval harnessesBenchmarkingDrift monitoring
Get an eval audit

Eval run · nightly

Factuality96.2%
Safety100%
Latency p951.2s
Cost / 1k calls€0.41

Illustrative run: shows the kind of eval harness X-GenAI builds, not a measured result from a specific client system.

Proof, not a pitch

Two demos, actually running

Self-built, not a case study — X-GenAI has no clients yet to claim results from. Both outputs below are copied verbatim from an actual run, not written by hand.

01 / For B2B SaaS teams

Onboarding Agent

Every new signup gets triaged the moment it happens: persona classified and activation risk scored by plain, readable rules, then a personalized welcome email drafted by Claude. If an account goes quiet, the agent drafts a Slack alert for CS — it never contacts an at-risk customer itself. A human always makes the save call.

Rule-based risk scoringClaude-drafted copyHuman makes the call
#

cs-activation-alerts

🔴 Nordfleet Logistics (Scale) has gone 91h with zero activation — flagging before this reads as churn. Suggest a personal outreach call re: “enterprise rollout / SSO,” not another automated email.

DRAFTED · AWAITING HUMAN REVIEW

02 / For digital agencies

Client-Reporting Automation

Every metric — CTR, cost per conversion, week-over-week deltas — is computed deterministically from raw campaign data, never touched by the model. Claude only narrates the facts that computation already flagged, so the summary can't invent a number that just sounds plausible.

Deterministic mathThreshold-based flagsClaude narrates only verified facts

Marlowe & Finch Law

🔴 At risk

“This week's clear win: average tracked keyword rank improved from 9.6 to 8.2. The thing worth your attention: conversions dropped -35.3% week-over-week (34 → 22).”

⚠ cost per conversion€65 → €109
✅ avg. keyword rank9.6 → 8.2

“Northstar Analytics” and “Marlowe & Finch Law” are fictional, invented for these demos — same disclosure the rest of this site uses for illustrative work. Every number above is computed for real from sample data; only the surrounding prose runs from a template without a live Claude API key. The source lives in X-GenAI's working repo, which also holds unreleased pricing and outreach material, so it isn't public — happy to walk through the actual code on a call.

See the most rigorously tested one — hardened against bad input & prompt injection →

Why it holds up

Built for what happens after the demo.

Concurrency & latency engineering
Observability & tracing
Human-in-the-loop error handling
Compliance-aware architecture
See the engineering behind it →

Working together

Clear scope at every stage, no lock-in. Five steps, and you can stop after any of them.

Start here

Discovery Audit

X-GenAI maps what's worth automating in your workflows and what each piece would cost.

Stage 1 of 5: Discovery Audit

See how each stage works →

Most “AI automation” looks great in the demo and falls apart the first week it meets your real data. X-GenAI would rather ship you something boring, the kind where it just… works.