Skip to content
knowbl
Request a demo
Resources

Everything dated.
Everything sourced.

No generic blog. The open scorecard, the CISO packet, the General Counsel FAQ, practitioner notes, and the market numbers we cite with their sources.

Open scorecard

The Four Requirements for Enterprise AI

Twelve questions procurement can put to any vendor, with Nexus's answers printed beside each.

Read it
Free, two minutes

Score your company

Your scorecard from your own answers, then a free 30-minute strategy jam that turns it into your AI Strategy Roadmap.

Start
For the CISO

The CISO packet

Hosting architecture in your own cloud account, the OAuth model, encryption, retention, MCP allow-lists, the audit trail, SOC 2 Type 2.

Request it
For the General Counsel

The General Counsel FAQ

What stays yours, what happens at termination, and whether anyone trains on your data. Plain English, reviewed by counsel.

Read it
Practitioner notes

Notes from engagements. One a month.

Anonymized where required: tier, industry and timeframe, never a name without sign-off. The first three are in progress.

In progress

What a scope model looked like at a wealth manager

Where Personal ended and Team began, and why memory recall stayed scoped to the project.

In progress

How a broadband provider consolidated its point solutions

Which tools became skills, which became agents, and which the department kept.

In progress

What a General Counsel asked for in the IP clause

Outputs, prompts, skills, memory, termination. The questions, and the architecture that answered them.

The numbers we cite

Every number, sourced and dated.

Verified against the primary report before it goes on this site. Where the research comes from a vendor, the source column names the vendor.

SignalThe numberSource
Shadow AI is mainstream, on personal accountsRegular AI use on corporate devices rose from 15% to 45% of employees in one year. 67% of that access runs through non-corporate accounts.Verizon 2026 DBIR, via BleepingComputer · May 2026
What leaks is IPThe most common data type submitted to external AI tools in 858,440 DLP events was source code.Kiteworks DBIR brief · May 2026
Shadow AI has a price$670,000 more per breach with high shadow AI use. 97% of organizations with an AI-related breach lacked AI access controls.IBM Cost of a Data Breach 2025 · July 2025
Analysts expect it to get worseMore than 40% of organizations will suffer security and compliance incidents from unauthorized AI by 2030.Gartner, via Infosecurity Magazine · November 2025
Point solutions multiplyAn average of 66 GenAI apps detected per organization; about 76% of shadow AI tools fail SOC 2.Teramind Shadow AI Behavior Report · June 2026
Budgets follow governance58% of organizations have AI embedded in enterprise-wide strategy, up from 26% in 2025. 47% plan AI budget increases of 20% or more.Info-Tech AI Trends 2026, via Continuity Insights · November 2025
Questions, by seat

The FAQ.

CIO: We already have a model vendor's enterprise plan. Why add a layer?
Keep it. Nexus plugs that model in as one of several and moves the interface, the context and the controls into a workplace you govern, on your own cloud. A repricing becomes a setting, not a migration.
CIO: Will governance slow adoption?
Only if the governed tool is worse than the shadow one. Nexus is designed to be the tool people would pick anyway, and skills and agents built in it are shared across teams instead of rebuilt in silos.
CEO: Why not build this ourselves?
Many companies intend to and most are behind. Nexus is what you would build, running in your account, with practitioners who have done it before. Spend your engineers on your agents, not on auth screens.
General Counsel: What stays ours in Nexus?
Your data, your prompts and your outputs are yours under the agreement, and they sit in your own account. Final wording follows counsel's review against the customer agreement.
General Counsel: Does Knowbl or any model vendor train on our data?
Not under the standard Nexus architecture. Your prompts, documents and model interactions are processed without being used to train public AI models or shared with underlying model providers for their own use. Your data powers your AI experience, not someone else's model.
CTO: Which model providers can we use?
Whatever you register: AWS Bedrock, direct provider APIs, any OpenAI-compatible endpoint (including self-hosted vLLM or Ollama), Google Vertex, or a custom HTTP endpoint. You map them to Fast, Balanced and Powerful tiers so the rest of the configuration never changes.
COO: What happens to our agents when the person who built them leaves?
In Nexus, agents, skills and connections are workspace assets, not personal projects. Each has an owner, a scope and a run history, and ownership can be reassigned. The skill that documents how the work is done stays with the company.
CFO: Will token costs run away?
Token limits per user and workspace, and every message records its usage. Routine steps run on the Fast tier by default; only the hardest decisions escalate to Powerful. Compaction keeps long conversations from ballooning. Thinking budgets are capped by the admin.
Everyone: Are we locked in?
No. Nexus runs in your own cloud account on the models you choose. Your data, skills, agents and connections live in your workspace, and switching a model is a setting, not a migration.