Coordinate agents with shared context, persistent memory and traceable actions. We gave a team of PortableMind agents one job — improve our own production platform every night — and let them run for 104 nights. Every action is on the record. No per-seat fees.
Every number on this page comes from our own production records, June 20 – September 27, 2026.
One orchestration run
Specialist agents hand work to each other through defined stages. They read the same project, write to the same memory, and none of their changes lands without a person approving it.
Scans for security advisories, outdated packages, duplication and design problems. Reads the full task history first.
Writes a short backlog (4 tasks max) into a tracked project, with exact files, risks and acceptance checks.
Work in parallel, each on its own branch with its own merge request, from the same shared context.
Runs the full suite (20,000+ tests). A regression sends the work back to the developer that made it.
One merge request and one test report are waiting for your review. Nothing merges without you.
Why it doesn't turn into chaos
Disconnected AI tools each start from zero and nobody can say what they did. Here every agent works from the same memory, and every action leaves a trace you can read.
Defined stages, bounded scope, budget limits and a human approval gate at the end. Multi-agent, not multi-chaos.
Each night's audit reads the full task history and carries leftovers into tomorrow's plan. One agent learns; all benefit.
Every one of the 391 tasks lives in one project; every run saves its audit, merge request and test report. See what your agents actually did.
“Collision check: the previous run's integration branch is not yet merged … Tonight's tasks avoid all of those files and that bump. … Candidates cross-checked against the prior ~380 backlog tasks to avoid repeats.”
The auditor agent, in its own report, September 27, 2026 — an agent reading the shared memory before it acts
Before and after
The first audit scored our backend 4/10 and listed critical findings. The latest one found no security advisories, only small, precise cleanups like a 15-line helper copied into three files.
Your agents, your workspace
Code was our first case because we could measure it exactly. The same orchestration — audit, plan, act, verify, approve, remember — runs any team of agents on any recurring work, with the same shared memory and the same audit trail.
Re-check documents against the latest rules every week.
Find stale pages, draft updates and route them for approval.
Reconcile records, flag anomalies and open tracked tasks.
Triage, draft replies and escalate with full context.
Build your first orchestration on PortableMind. Agents share one context, every step is tracked, and nothing ships until you approve it. No per-seat fees.
Start AI orchestration freeno credit card needed to start · 1M free AI tokens · flat pricing
Copied from Russell's /nightly-case-study/ and re-aimed at agent orchestration. Case-study numbers and the auditor quote are exactly as on his page; the founders section is dropped for length (it can come back). Hero, section leads, pillar copy and finale are onb's draft. Ad language in quotes is from the Sept-26 export.
01 › AI Agent Orchestration (20 kws, 3 live) — “Stop agent chaos”, “Run every agent in one place”, “One source of truth for agents”. 01 #2 › AI Agents for Teams (10 kws, 3 live) — “Your agents in one place”, “AI agents with shared memory”, “Audit every agent action”. 01 › Shared AI Memory (19 kws, 3 live) — “One memory layer for your AI”, “Give every agent shared memory”, “One agent learns. All benefit”. Flag: Shared AI Memory's team-memory keywords (“ai memory for teams”, “enterprise ai memory”) read as cluster B buyers; the agent-memory half belongs here. Worth its own page if the split holds in the numbers.
Segment agents (exists) → highlighted card “Hire your first AI teammate” (proposed re-aim from “agent chat”); row adds “Give the agent a memory”. Open the Welcome audition.
Orchestration ads: promise “Coordinate agents with shared context, persistent memory and traceable actions” / “Stop the chaos of disconnected AI tools”; CTA “Start AI orchestration free”, “No credit card needed to start”. Honoured by: hero (same words), “One orchestration run”, “Why it doesn't turn into chaos”, CTA “Start AI orchestration free”. Agents-for-teams ads: promise “All your AI agents in one place”, “Audit every agent action”, “AI teammates that remember”; CTA “Start free now”. Honoured by: the swap (below) puts “Your agents in one place” in the hero; the audit pillar and the compare panel are the proof. Shared-memory ads: promise “One memory layer for your AI”, “Stop re-explaining to AI”; CTA “Start free. No card needed”. Honoured by: the swap; the memory pillar and the auditor quote (an agent reading shared memory). Welcome: segment agents, first card hires an agent with a job description and a memory.
Per-ad-group swap (draft mechanism): ?utm_content=ai-agent-orchestration · ?utm_content=ai-agents-for-teams · ?utm_content=shared-ai-memory.
Internal annotations, generated from the campaign map (2026-09-29T14:25Z). This panel is the static form of the notes overlay on the local viewer; nothing here is customer copy. ← Campaign index
ClusterA · Agent orchestration & agents for teams — “Stop agent chaos. Run every agent in one place.”
Campaign angleorchestration · base /nightly-case-study/
Match vs this variantmatch the case study is the proof of orchestration, shared memory and audit trails; hero/pillars/finale in the ads' words; CTA 'Start AI orchestration free'; swap echoes each ad group.
Rule: Message-match rule (Rick, 2026-09-28): there is not necessarily a one-to-one mapping from ad group to landing page, but no mapping should bring a person from an ad to a landing page that doesn't match the ad's expressed value proposition, CTA etc. It extends through the Welcome segment: promise → page → Welcome.
1Campaign01 Portablemind Search
Suggested name (ours)Search | Agents, Knowledge, AI-native PM & Automation (themed) | US (market not in the export — US is our guess)
2Ad groupAI Agent Orchestration 20 kws · 3 live of 4 ads
Keywordsai agent orchestration system · ai agent orchestration tool · ai agent orchestration platform · agent orchestration · multi agent orchestration tools · multi agent orchestration software …
3Ad
HeadlineAI agent orchestration
PromiseCoordinate agents with shared context, persistent memory and traceable actions.
CTA— · Match vs today's pagepartial
3Ad
Headline{KeyWord:Your team's workday}
PromiseRun AI agents with shared context and persistent memory, without constant re-explaining.
CTAStart AI orchestration free · Match vs today's pagepartial
3Ad
Headlineportablemind.ai
PromiseKeep multiple AI agents coordinated with shared context. No per-seat fees.
CTANo credit card needed to start · Match vs today's pagepartial
3Proposed ad (ours)pull-up · for Manuel’s review — ours, not in the account
HeadlineRun every agent in one place
PromiseRun all your agents in one workspace: shared memory, full audit trail. Start free.
CTAStart AI orchestration free · Match vs this variantmatch points at this page
Landing page (today)/team/ partial agents appear in section 4 ('Meet Nova'); orchestration, shared agent memory and audit trails — the ads' promise — are not on the page. The nightly case study, which shows them, is unused.
Hero swapthe live page would put this ad group’s words in the hero on ?utm_content=ai-agent-orchestration; pre-rendered here instead (the hero above shows the cluster default):
Headlineeyebrow AI agent orchestration · real results, 104 nights
Stop agent chaos. Run every agent in one place.
Coordinate agents with shared context, persistent memory and traceable actions. 104 nights of a real multi-agent orchestration, every action on the record. No per-seat fees.
CTAStart AI orchestration free
1Campaign01 Portablemind Search #2
Suggested name (ours)Search | AI Agents for Teams (themed) | US (market not in the export — US is our guess) · appears to be a clone of 01 (same 12 ad groups, same ads) in which only AI Agents for Teams is live (3 ads, 10 kws); everything else is paused, and the paused Shared AI Memory / Workflow Automation / Knowledge Base / Agile / 1.0 copies point at the homepage — or retire it: move AI Agents for Teams into 01 (it belongs beside the agents groups there) and delete #2, so the account has one themed campaign instead of a clone with one live group
2Ad groupAI Agents for Teams 10 kws · 3 live of 4 ads
Keywordsai agents for businesses · ai agents for work · ai agent management platform · ai agents for enterprise · enterprise ai agents · ai business agents …
3Ad
HeadlineShared AI agents for teams
PromiseStart free with 1M AI tokens. No credit card, no per-seat fees. Bring the whole team.
CTA— · Match vs today's pagepartial
3Ad
Headline{KeyWord:Your agents in one place}
PromiseBuild and ship faster, without re-explaining context. Share agent memory with your team.
CTAStart free now · Match vs today's pagepartial
3Ad
Headline{KeyWord: Keep agents in one place}
PromiseGive your team one place for agents, files and shared context. No per-seat fees.
CTA1M free tokens · Match vs today's pagepartial
Landing page (today)/team/ partial agents appear in section 4 ('Meet Nova'); orchestration, shared agent memory and audit trails — the ads' promise — are not on the page. The nightly case study, which shows them, is unused.
Hero swapthe live page would put this ad group’s words in the hero on ?utm_content=ai-agents-for-teams; pre-rendered here instead (the hero above shows the cluster default):
Headlineeyebrow AI agents for teams · real results, 104 nights
Your agents in one place. AI teammates that remember.
Bring your AI agents, projects, knowledge and team into one place. Agents keep context across tasks, chats and docs, and you can audit every agent action. No per-seat pricing.
CTAStart free now
5Welcome segmentagents
Highlighted cardHire your first AI teammate
highlight moves from agent-chat to hire-ai-teammate; row adds agent-memory
Card click hands off“Let's hire your first AI teammate. What role would help most?”
6Mindy opening (proposed)
“You came to run agents with shared memory instead of a pile of disconnected tools. Let's hire one, give it a job description and a memory, and you'll see every action it takes.”
Mindy opening (catalogue today)“Hi Ada — I'm Mindy. You wanted an agent helper in your chats. Shall we set one up? Tell me what the chat is for and I'll bring the right helper.”
a job description, a memory and an audited first task are the three promises the ads make.
Welcome screenshots for segment agents → · This ad group on the index →
