DataOps Drive
Your documents. Your AI. One private archive.
Connect the ChatGPT, Claude or Grok your team already uses to everything you have ever filed — including the scanned paper.
The problem
The problem isn't the AI. It's that your AI cannot see your documents.
- Returns, objections, audits and correspondence are buried in folders you can only search by file name.
- A large share is scanned paper — invisible to every ordinary search tool.
- Analysts burn their time hunting files instead of advising clients.
- The obvious shortcut — uploading the archive to a public AI — leaks confidential data.
- Nothing is grounded: the same question each time gets an unverified answer.
The shift
Your own AI, reading your own archive, safely.
- Ask ChatGPT, Claude or Grok a question — it searches and reads your documents itself.
- It reads Word, Excel, PowerPoint, PDF, images — and scanned paper.
- It sees only what each person is already allowed to see; never another department's files.
- Every AI connection is a revocable grant an administrator can cut off at will.
- Answers come back with citations that open the exact source document.
How it works
One pipeline, end to end.
- 1
Capture
A small agent runs on each laptop and syncs the folders your team already works in. Nothing to migrate, nothing to change.
- 2
Store
Files land in your own private storage, on your infrastructure, scoped by role and department.
- 3
Index
Every folder is catalogued and kept current. Scanned paper is recognised into text, so it stops being invisible.
- 4
Connect
One secure connector — open standard (MCP) — lets ChatGPT, Claude or Grok read the archive as the signed-in person, and no one else's files.
- 5
Act
Agents research, draft, review and act across your systems. This is the Symbiote Agent Harness, priced separately.
Five stages. Your team touches none of them.
What you get
One system: your files, made answerable.
Why only us
What you get from us that you cannot get anywhere else.
- Inside the AI you already use — no new portal, no training curve, works on phone and desktop.
- Scans are first-class — the paper that blocks every other tool becomes searchable text.
- Permissions inherited, not re-invented — the AI reads exactly as the person, no more.
- You own it — runs on your servers; no per-query vendor tax as you grow.
- Open standard (MCP) — when the next AI app arrives, it plugs in with no rebuild.
- Automatic from your laptops — the archive builds itself; adoption costs your team nothing.
Proof
It is not a concept. It already works.
Built and running against a real archive at a tax advisory practice.
~63,000
documents indexed and searchable
~93%
accuracy reading a scanned letter
3 assistants
one secure connector
Revoked
access cut at the source
- ~63,000 documents indexed and searchable.
- A scanned paper letter was read back, in full, at ~93% accuracy — proof the paper is not a blocker.
- Live with ChatGPT, Claude and Grok through a single secure connector.
- Search answers in one request and downloads no files to find them — fast, and cheap at scale.
- Access is enforced at the source: removing a person ends their AI's access on the next call.
No client is named. Proof is stated, not attributed.
Works with what you have
The AI you already pay for.
Built on MCP, the open connector standard — so the next assistant plugs in without a rebuild.
Your AI subscriptions stay yours.
Pricing
Published pricing. No surprises.
DataOps Monthly
12-month minimumTZS 2,900,000
/ month · ~USD 1,074
- Up to 15 users — start with 10
- Extra user: TZS 60,000 / month
- Setup, onboarding, training, infrastructure and support included
- Upgrades billed at the incremental difference only
DataOps One-off
You own itTZS 48,000,000
one-off · ~USD 17,780
- Up to 15 users
- Extra user: TZS 700,000 one-off
- Instalments: 3 × 17M · 6 × 8.8M · 12 × 4.5M
- After: TZS 850,000 / month for infra + maintenance, or run it yourself
Symbiote Agent Harness
Perpetual, up to 25 usersTZS 39,000,000
one-off add-on · ~USD 14,440
- Extra user: TZS 700,000 each
- Instalments: 3 × 14M · 6 × 7M · 12 × 3.7M
- Optional support: TZS 1,100,000 / month or TZS 12M / year
- AI usage metered, or bring your own key
Take both: 10% off the combined one-off, or your first month free with DataOps Monthly + Harness. Three-year price freeze on recurring fees. Founding-customer terms for the first practice to deploy.
Prices exclude VAT (18%). Full commercial terms are delivered with the proposal.
Scope
What's included — and what isn't.
Included
- Setup and deployment
- Data onboarding
- Staff training
- Infrastructure
- Support and updates
Not included
- Your ChatGPT / Claude / Grok subscriptions
- Storage beyond the included tier
- Bespoke new integrations (quoted separately)
Getting live
Four steps, no surprises.
- 1
Selection
You choose the option that fits — monthly or one-off.
- 2
Agreement
First instalment (or 50%) on signing begins provisioning.
- 3
Deployment
We bring our prior work — harness, tax tools, connector — into your environment.
- 4
Go-live
Staff walkthrough and training; final payment on production delivery.
Future-proof: upgrades available at the incremental difference only.
FAQ
Where the objections die.
Does our data go to OpenAI or Anthropic?
Documents are read only when the signed-in person asks a question that needs them, through a connector authenticated as that person. Access is per-person and can be revoked at any time, at the source.
Do we need a new app?
No. You keep using ChatGPT, Claude or Grok.
What about our scanned paper?
It is recognised into text, so it becomes searchable and citable instead of invisible.
What if someone leaves?
Their AI access ends on the next call — access is inherited from the person, not from a shared API key.
Who owns it?
On the one-off option, you do — you can run it yourself.
Does it work on a phone?
Yes — wherever the AI app works.
The archive is already there. Let it answer.
Send us 20 documents. We'll show you your own archive answering questions about them.
Private by design — your documents stay on your infrastructure