For the head
A business case with your numbers
Coordination time, stack utilisation, approvability and budget discipline. Four levers with source, mechanism and metric, plus a worked example you fill with your own assumptions.
The operating system for agentic marketing
Ask, draft, approve. Across your whole marketing stack.
You ask a question or set a goal. batoon reads your CRM, email, ads and analytics together and drafts the next step. Every action stays within your brand, your budget and your approvals.
How many contacts do we have in the CRM, by country?
Read the CRM
batoon
The CRM holds 450 contacts, 262 of them in Germany.
Contacts by country
simulatedTotal: 450 · Source: CRM
Read only. Nothing changed.
Start the NovaBoard product launch.
Run started after your confirmation
batoon
The plan has four tasks. It is waiting for approval by the campaign lead.
Plan approval
requires campaign lead Goal: 500 trial signups4 weeksBudget €25,000
- 1 Build a segment of engaged contacts
- 2 Draft the launch email for the segment
- 3 Send the launch email to the segment
- 4 Check the quality of the segment
Waiting for your approval. Nothing starts without it.
Show me the draft of the launch email.
Read the run: task 2, Emma
batoon
Emma saved the draft in the email system. The brand judge has checked it.
Draft · launch email
simulated- Subject
- Meet NovaBoard
- Segment
- Engaged contacts · 233
NovaBoard is here. Start your trial today.
- Brand judge: 0 violations
- Banned claims: none
- Tone: professional
Saved as a draft. Nothing sent.
Where does the product launch stand?
Read the run
batoon
Two tasks are done. The send, task 3, is waiting for approval by the campaign lead.
- allowed Build segment
- allowed Draft launch email
- on hold Send launch email
Why? Risk level medium reaches your escalation threshold.
Your guardrails
Budget cap €25,000At most 2 touches per weekBanned claims: guaranteed, risk-free, #1
Every decision is recorded in the audit trail, with its reason.
How did the launch email do?
Read the email system
batoon
The launch email went to 180 contacts. 110 opened it, 22 clicked.
Launch email · response
simulated16 contacts without consent left out · Source: email system
Sent in the simulated world, not to real contacts.
What did the launch email achieve?
Compared the outcome with the control group
batoon
An estimated 30.9 additional conversions. The comparison is 37 contacts who got no email.
+30.9 conversions simulated
additional, estimated against the control group
Emailed
99 of 180
55 % converted
Control group
14 of 37
37.8 % converted
Method: holdout comparison
Measured in the simulated world. The control group holds across all channels.
Rendered after the batoon interface. Every number comes from the curated simulated world with the example company NovaTech. Connected real systems are read-and-draft only.
The problem
Agentic marketing fails on control, not on AI.
Every system in your stack is getting its own AI agents. But each one works inside its own silo. The agent in your email tool knows neither your ads budget nor your call centre. Nobody owns the whole picture.
21 %
of companies have a mature governance model. 74 % still plan to deploy agents by 2027.
Deloitte
> 40 %
of agentic AI projects will be abandoned by the end of 2027.
Gartner
15,384
MarTech tools on the market, two thirds of them unused. Of roughly 50 Fortune 500 CMOs surveyed, not one could quantify the ROI of their stack.
chiefmartec 2025, McKinsey
The value
Lead by goals and values.
batoon does the legwork across your stack, bound by your rules. The promise holds for the head and for the heart.
For the heart
Away from legwork in twenty tools
You write down what should be achieved and what must hold. A team of agents works across your stack. You decide at the five points where it matters.
The layer in between
Set a goal. batoon runs it across your stack.
batoon sits between the AI models and your MarTech stack. It stays independent of both sides. You can swap either end without losing control.
AI models replaceable
- Anthropic
- OpenAI
- Mistral
- OpenRouter
- Open weights
- Goals & plans
A measurable goal becomes a plan in parallel waves.
- Agent teams
Assembled like a real team, without a line of code.
Policy gate before every action
- Approvals
- Audit trail
- Conflict resolution
- Impact measurement
- CRM
- CDP
- Ads
- CMS
- Analytics
- Creative
- Call centre
Your MarTech stack replaceable
Everything connects through the open MCP standard. Your company knowledge and configuration stay with you as version-controlled YAML and Markdown. Company as code, no data lock-in.
Four promises
Control you can verify.
01 in code, not in the prompt
Check every action before it runs
The policy gate checks budget caps, frequency caps, forbidden claims, consent, suppression lists and risk classes. Each agent may only touch what belongs to its own task. Every verdict comes with a reason: allow, flag, escalate or deny. A brand judge also checks against your brand guide, and it can only tighten a verdict.
02 enforced server-side
Route approvals by role
Five situations pause a run. Plan approval, an escalation, a conflict, variant sign-off and a scheduled start. Who may decide follows from role, risk and budget size. Multi-stage approvals follow the four-eyes principle. Who actually decided is read from the session, not from the browser.
03 why this decision
Back every decision with evidence
The audit trail is append-only per run. It records every action, every verdict, every human decision and every model call with its cost. Anything that changes something carries a checked rationale with numbered evidence. An argument without evidence halts the run.
04 customer boundary
Measure impact across every channel
Your control group stays clean in email, in ads and in the call centre. No single vendor can promise you that, because none of them sees the other channels. And it is calculated without a single contact ID ever reaching batoon.
Why now
Four curves cross in 2026.
- 01
Agents reach scale
Gartner expects a third of all enterprise applications to include agentic AI by 2028.
- 02
Every system opens up
The open MCP standard sees roughly 97 million SDK downloads per month. HubSpot, Adobe and Salesforce ship their own servers.
- 03
Governance decides the purchase
Only 21 % of companies have a mature governance model. 74 % still plan to deploy by 2027 (Deloitte).
- 04
Traceability becomes mandatory
Since August 2026 the EU AI Act requires AI decisions to be evidenced. In regulated industries that decides whether a project starts at all.
The bottleneck is moving from integration to orchestration and governance.
Where batoon stands
We show you the measurement, not a top grade.
batoon measures itself repeatably against a catalogue of 96 enterprise capabilities, eight layers and 13 industry scenarios. This is where it stands in September 2026, after 105 documented rounds.
66.8 %
Capability coverage. 625 of 936 points, 13 of 13 industry scenarios feasible by our own catalogue. In August it was 51.2 % and 2 of 13.
By layer
- Decisioning 72 %
- Orchestration & governance 71 %
- Platform & operations 71 %
- Design 69 %
- Human–agent interaction 68 %
- Distribution 65 %
- Measurement 63 %
- Data layer 55 %
The 18 critical gaps from August are closed. 24 capabilities remain below maturity 3, all in the should and could classes. We set the assessment ourselves.
What is built
- 3
- use cases end to end in a curated simulated world
- 5
- real MarTech systems connected, read-and-draft only
- ~56,000
- lines of code in Python and TypeScript, excluding tests
- ~3,900
- test functions
- 105
- documented rounds in nine weeks
- 1
- founder, built solo
What batoon cannot do
- No live execution. The five connected systems run read-and-draft. Nothing is sent, no campaign launches.
- No paying customers. We are looking for our first design partners.
- The three use cases and every per-contact decision run in a curated simulated world, not on real customer data.
- We set the assessment ourselves. External review is still open.
What we are looking for
Two things.
Two to three design partners
Mid-size and large enterprises with a real marketing stack, ideally in a regulated industry. Paid pilots where you help shape the product.
Pre-seed, around €450,000
As a convertible loan, funding the validation phase and the first hires. The founder holds more than 75 % after the first round.
Contact
Write to me.
Straight to the founder. No form, no layer in between.
Email: florian@batoon.ai
LinkedIn: Florian Stompe
What belongs in the first email
Your industry, which systems sit in your stack, and when you want to introduce agentic marketing. That is all a first conversation needs.