Managed AI infrastructure & operations

Stop guessing with AI. Start running an engine.

An AI assistant that read the internet is a novelty. One that read your pricing policy, your quarterly reports and your actual CRM records is a colleague. We build that inside your tenancy, and prove every month it still works.

Built in your tenancyZero vendor lock-inNZD, excl. GST
The fragmented AI trap

Disconnected bots are telling your customers things that stopped being true.

Right now, well-meaning teams across your business are building their own prompts, skills and standalone bots. Marketing built a tool for copy. Sales pointed an assistant at a folder of old bids. HR set up a bot to answer policy questions. None of them talk to each other, nobody maintains them, and they are failing quietly in production.[1,2]

The soundproof office

Imagine hiring 15 brilliant executives, locking each in a separate soundproof room, handing each an outdated operating manual, and asking them to run your company without ever speaking to one another.

That is how most enterprise AI is deployed today. You aren't building leverage. You're building a digital Tower of Babel.

1.8 hrsLost every day to searching

Knowledge workers spend 19.8% of the working week hunting through drives, chats and inbox threads for answers the business already has.

McKinsey Global Institute[1]
78%Bring their own unmanaged AI

Three quarters of knowledge workers now use generative AI at work, and most bring their own tools to corporate files with no IT oversight.

Microsoft & LinkedIn Work Trend Index[2]
30%Retrieval accuracy decay

Ungoverned retrieval pipelines lose up to 30% of precision within six months as documents and schemas drift. No error is ever thrown.

Nature Scientific Reports & IEEE RAG research[3,4]
The enterprise operating system for AI

One brain. A departmental fleet. Self-healing assurance.

You don't need 20 disconnected AI experiments. You need one governed Company Brain powering every system, department and workflow — enclosed inside your own enterprise boundary.

Your enterprise tenancy — Microsoft 365, Google Workspace or Claude
Live system connectors
CRM — Salesforce, HubSpotERP & financeSharePoint & DriveInboxes & calendarsBids & contractsInternal wiki
Read-only access · approval gates

The Company Brain

Core engine
Canonical knowledge set25–50 governed truths, written once and owned by a named steward.
Reality vs. strategy splitOperational fact kept strictly apart from strategic intent.
Compiled truth & evidenceDated, audited history so every answer can be traced.
4-layer governanceIdentity & permissionsPlatform controlsConnector boundariesInstruction sets
A single governed source of truth
Sales & bids
  • RFP bid engine
  • Pitch generator
  • Pricing verifier
Ops & delivery
  • SOPs & manuals
  • Project sync
  • Quality controls
HR & people
  • Policy advisor
  • Onboarding copilot
  • Talent Q&A
Finance & legal
  • Contract audits
  • Invoice matching
  • Compliance checks
Exec & strategy
  • Board packs
  • Risk drift
  • KPI audit
Measured monthly
Brain Assurance — the self-healing loop
  • Monthly accuracy benchmark
  • Policy drift remediation
  • Model migration testing
  • Safety gates
01

Live connectors, zero duplication

We don't replace your systems. We connect to CRM, ERP and SharePoint read-only, behind your identity provider.

02

Reality separated from strategy

Operational facts stay apart from strategic intent, so day-to-day drift can never quietly corrupt policy.

03

One truth, an entire fleet

Update knowledge once in the Brain and your sales, ops, HR, legal and executive agents all update with it.

04

You own the tenancy

Built on your Microsoft 365, Google Workspace or Claude licences. The infrastructure and the IP remain yours.

A predictable roadmap

From unmanaged chaos to a governed estate.

No massive upfront lock-in, and no unmaintainable code delivered over the wall. We establish the foundation, build capability at a controlled rate, and keep the estate alive and accurate.

01

Standardise & audit

We inventory what you already have: every bot, permission, integration and cost. You get runbooks and a baseline.

Month 1 · $7,500 one-off, credited back on a plan
02

Build your Company Brain

Tribal knowledge rewritten into 25–50 canonical documents in your tenancy, with live connectors and governance.

Weeks 2–8 · from $12,000 one-off
03

Attach capabilities

Agents that reason, or low-code tools that remove friction — built from a backlog you control.

Ongoing · from $4,500 a month
04

Run, protect & assure

One layer keeps the platform healthy. The other keeps the answers accurate. Two jobs, one invoice.

Monthly · from $2,500 a month
1 · Managed Plans

Keeps the platform healthy

The engineering layer that stops what we built from quietly falling over.

  • System monitoring and error capture
  • SLA-backed incident response, down to one hour
  • Security patching and connector maintenance
  • Dedicated monthly engineer change capacity
2 · Brain Assurance

Keeps the knowledge accurate

The measurement layer that catches wrong answers before your people or your clients do.

  • Monthly 50-question benchmark runs
  • A score for how often the right source is found, and how much of it comes back
  • Knowledge drift detection and source fixes
  • Model migration testing before provider cutovers
Choosing what to build first

Agent Studio builds the mind. App Factory builds the muscles.

The Company Brain is your enterprise memory. Agent Studio adds digital colleagues that read, reason, draft and synthesise. App Factory adds the forms, approvals and routing that take friction out of the day.

Agent Studio vs. App Factory
ScopeAgent Studio — cognitionApp Factory — workflow
Primary focusReasoning, Q&A, proposal drafting, policy synthesis.Forms, approvals, data entry, routing, dashboards.
Underlying enginePowered directly by your Company Brain.Built natively on Microsoft 365 or Google Workspace.
Build velocity3 to 5 days per standard agent.1 to 2 days per standard app.
Sample deliverablesRFP bid engine, HR policy advisor, competitive watch.Equipment inspection app, expense approval flow, team dashboard.
Delivery rhythm1, 2 or 4 agents a month, from $5,000/mo.2, 4 or 8 apps a month, from $4,500/mo.
What it is notNo autonomous action on money or customers without human sign-off.No multi-month ERP builds or date-bound regulatory projects.
Transparent pricing

Build products, then a plan that keeps them true.

Build work is fixed and one-off. The managed layer is metered on the number of artefacts we look after, so the price tracks the size of your estate rather than a guess.

Build & pipeline products
Build productInvestmentEngagementKey deliverable
Standardise$7,500one-off3–4 weeksReadiness audit, permissions overhaul, secret vaulting, runbooks, baseline eval suite, cost baseline. Credited back in full on a Managed Plan.
Company BrainFrom $12,000one-off6–8 weeksCanonical knowledge set in your tenancy, working connectors, 4-layer governance, 50-question baseline score. Essential $12k · Standard $22k · Enterprise $40k.
Agent StudioFrom $5,000per month1, 2 or 4 agents/moHigh-impact agents connected to your Brain, with working sessions and a custom evaluation set per agent. Studio 1 $5k · Studio 2 $9.5k · Studio 4 $18k.
App FactoryFrom $4,500per month2, 4 or 8 apps/moInternal Power Apps, Google Workspace tools and workflow automations, built from a backlog you control.
Managed operating plans
Managed planMonthlyArtefactsChange capacitySLA & support
Watch$2,500Up to 5Monitoring only5 business-day issue response, continuous patching, monthly health reports.
Run$5,000Up to 102 engineer-days2 business-day response, continuous regression testing against your eval suite.
OperateMost chosen$8,000Up to 203 engineer-days4-hour P1 response, on-call escalation, model migration management, prompt version control.
Scale$16,000Up to 4010 engineer-days2-hour P1 response, named delivery lead, fortnightly triage sessions, 12-month maintained roadmap.
Programme$28,000Up to 8020 engineer-days1-hour P1 response, named onshore NZ architect, quarterly governance packs, ISO 42001 and audit support.
Brain Assurance add-on
From $1,500 / month

Every month we put the same 50 real business questions to your Brain and score the answers two ways: how often it surfaced the right source, and how much of the relevant material it actually returned. When an answer fails, the fix is included — we correct the underlying document before a wrong answer reaches your staff or your clients. You get the score, the failures and what we changed, in a report you can hand to your board. Essential $1.5k · Standard $3k · Enterprise $6k per month.

Prices in NZD, excluding GST. Platform inference, hosting and telephony are passed through at cost with a stated handling margin, itemised on your invoice. Microsoft, Google and Claude seat licences are held directly by you, with no markup.

Why organisations choose possibl

We ran this on ourselves for a year before we sold it to anyone.

What you leave with

A governed knowledge asset in your own tenancy, a fleet that stays in sync with it, and a monthly score that tells you whether it is still true.

  • Zero lock-in. Everything is built in your tenancy on your licences. If you stop working with us, the infrastructure stays 100% yours.
  • Pass-through platform costs. Inference, hosting and telephony at cost plus a stated handling margin, itemised.
  • Human in the loop. We refuse to build agents that move money or contact customers without an explicit approval gate.
  • No retainer waste. Unused change capacity is reserved for you. We sell active measurement, not insurance against our own code.
  • Clear boundaries. Data remediation in inaccessible systems is reported during the audit and quoted separately. We produce the evidence your board or auditor needs, but we don't act as your auditor.
References

Every number on this page has a source.

  1. [1]McKinsey Global Institute. The social economy: unlocking value and productivity through social technologies. Knowledge workers spend 1.8 hours daily, around 19.8% of work time, searching for and gathering internal information.
  2. [2]Microsoft & LinkedIn Work Trend Index. AI at work is here. Now comes the hard part. Global benchmark of 31,000 workers across 31 markets: 75% AI adoption, 78% bringing their own unmanaged AI tools.
  3. [3]Nature Scientific Reports. Temporal degradation in machine learning models. Peer-reviewed study across 128 model-dataset combinations: 91% of production models degrade over time as real-world data shifts.
  4. [4]IEEE and Atlan research. Vector index staleness and retrieval degradation in enterprise RAG systems. Ungoverned retrieval pipelines lose 20–30% of precision over six months without continuous benchmarking.

Ready to build your Company Brain?

Give us 30 minutes. We'll review your current estate, pinpoint the unmanaged shadow AI already running in your business, and map a governed architecture you can take to your board.

Book your 30-minute scoping session