Blog

Insights on custom AI agents and enterprise automation.

Span of Control: How AI Employees Change the Manager-to-Report Ratio
AI Strategy

Span of Control: How AI Employees Change the Manager-to-Report Ratio

Span of control was capped at 5-8 by one bottleneck: a manager's coordination capacity. In 2025 the average crossed 12 as companies delayered - most of them the fragile way, cutting managers but not the coordination work, so the survivors drowned. This piece reframes the oldest lever in org design for the AI era: when routine coordination (status-chasing, re-explaining, reviewing, routine exceptions) moves to AI employees grounded in a Company Brain, one manager can lead a larger, flatter team without cutting a single person. Covers the classic 5-8 rule and Graicunas maths, why span stayed narrow, delayering-by-firing vs widening span with AI, why the Company Brain makes it durable through turnover, a euro-and-output model, a 90-day playbook, and how it differs from an org-chart or workflow tool. Leverage, not headcount reduction.

Henri Jung Henri Jung · 34 min read
The Best AI Deep Research Tools in 2026: ChatGPT vs Gemini vs Perplexity vs Claude - and Why None of Them Knows Your Company
AI Strategy

The Best AI Deep Research Tools in 2026: ChatGPT vs Gemini vs Perplexity vs Claude - and Why None of Them Knows Your Company

An honest 2026 buyer comparison of the deep-research agents (ChatGPT, Gemini, Perplexity, Claude and Grok) on runtime, citation accuracy, synthesis and enterprise pricing - plus the gap they all share: they research the public internet brilliantly and never know how your company works. The durable win is a Company Brain that survives staff turnover plus an AI employee that acts across your real systems.

Henri Jung Henri Jung · 34 min read
The Routine-Work Tax: Why Your Best People Spend Half Their Day Below Their Pay Grade
AI Strategy

The Routine-Work Tax: Why Your Best People Spend Half Their Day Below Their Pay Grade

You pay expert salaries but receive clerical output when skilled people spend half their day on data entry, status chasing and copying between systems. This piece names that hidden tax, models its euro cost, shows why copilots and point tools cannot remove it, and gives a practical pay-down playbook built on leverage, not headcount.

Henri Jung Henri Jung · 38 min read
The Best AI Customer Data Platforms (CDP) in 2026: An Honest Buyer Comparison
AI in Marketing

The Best AI Customer Data Platforms (CDP) in 2026: An Honest Buyer Comparison

An honest 2026 comparison of the AI customer data platforms that matter - Segment (Twilio), Amplitude, Adobe Real-Time CDP, Salesforce Data Cloud, Hightouch, Tealium, mParticle and RudderStack, plus generic ChatGPT and Claude as a baseline - with real capabilities and pricing reality. Every CDP unifies and activates customer data, but none keeps how your company actually interprets a customer - the segment definitions, what counts as a qualified account, the exception and suppression rules - nor runs the last mile end to end across your real systems. The durable win is a Company Brain that keeps your customer-data logic when the marketing-ops or data owner leaves, plus an AI employee that maintains segments, hygiene and activation across CRM, email, the CDP and the warehouse. Includes the DSGVO, EU AI Act Article 50 and US CLOUD Act realities most CDP comparisons skip.

Henri Jung Henri Jung · 35 min read
Your Company's Best Thinking Is Trapped in Personal ChatGPT Accounts
AI Strategy

Your Company's Best Thinking Is Trapped in Personal ChatGPT Accounts

Nearly half of enterprise AI now runs through employees' personal ChatGPT accounts, so the reasoning, drafts and hard-won prompts they produce every day never become company knowledge - they evaporate at logout and leave with the person, while creating DSGVO and EU AI Act exposure. This thought-leadership piece uses verified 2026 data (LayerX 47% of AI conversations on personal identities, Cyberhaven 39.7% of interactions with sensitive data, IBM's $670k shadow-AI breach cost, Panopto knowledge-loss figures) to model the euro cost of the leak, explains why wikis and SharePoint cannot fix it, and shows how a Company Brain that survives turnover plus AI employees acting across email, Teams, SharePoint, CRM and ERP keep the thinking inside the company. Distinct from the shadow-ai-governance article: this is about knowledge loss, not governance risk. Includes a 90-day playbook and a decision framework.

Henri Jung Henri Jung · 34 min read
The Best AI Tools for Compliance and Audit Management: An Honest 2026 Buyer Comparison
AI in Compliance

The Best AI Tools for Compliance and Audit Management: An Honest 2026 Buyer Comparison

An honest 2026 comparison of AI compliance and audit tools - Vanta, Drata, Secureframe, Sprinto, AuditBoard/Optro, ServiceNow GRC, OneTrust, LogicGate and Hyperproof, plus generic ChatGPT and Microsoft Copilot - with real capabilities and pricing tiers. Every tool tracks controls and stores evidence but none keeps how your company actually assesses compliance - control rationale, exception decisions and past-audit reasoning - nor runs the routine evidence collection end to end across your real systems. The durable win is a Company Brain that keeps that reasoning when the compliance owner leaves, plus an AI employee that runs the routine evidence and questionnaire work across cloud, identity, HR, ticketing and email. Includes the ISO/IEC 42001, EU AI Act and DSGVO realities most comparisons skip.

Henri Jung Henri Jung · 35 min read
Process Debt: The Hidden Tax of Undocumented Workarounds
AI Strategy

Process Debt: The Hidden Tax of Undocumented Workarounds

Process debt is the accumulated cost of every undocumented workaround, tribal shortcut and off-book exception - it compounds silently and comes due catastrophically the day the one person who understood the shortcut leaves. Why static wikis, SharePoint and SOPs never pay it down (they capture the idealised outcome, not the reasoning and exception-handling), and how a Company Brain that survives turnover pays it down structurally, with AI employees able to run the concentrated person's routine work across email, Teams, SharePoint, CRM and ERP. Borrows Ward Cunningham's technical-debt metaphor and applies it to process, with verified 2026 data (Panopto knowledge-loss research, Gallup turnover cost, McKinsey tech-debt, Gartner agentic-AI), a euro model of process-debt exposure and a practical pay-down playbook.

Henri Jung Henri Jung · 33 min read
The Best AI Tools for Supply Chain Planning and S&OP: An Honest 2026 Buyer Comparison
AI in Supply Chain

The Best AI Tools for Supply Chain Planning and S&OP: An Honest 2026 Buyer Comparison

An honest 2026 comparison of the AI supply chain planning and S&OP tools that matter - SAP IBP, Blue Yonder, o9, Kinaxis, ToolsGroup, Flowlity and John Galt, plus ChatGPT and Claude as a baseline - with real capabilities and pricing reality. Every suite forecasts and optimises, but none keeps how your planners actually decide - the driver assumptions, override reasoning and exception rules - nor runs the S&OP cycle end to end across your ERP. The durable win is a Company Brain that keeps your planning logic when the demand or supply planner leaves, plus an AI employee that owns the routine S&OP cycle across ERP, email and the planning tool. Includes the EU AI Act Article 50 transparency line, DSGVO and CLOUD Act realities most comparisons skip.

Henri Jung Henri Jung · 35 min read
The Bus Factor: When One Person Leaving Stalls the Whole Company
AI Strategy

The Bus Factor: When One Person Leaving Stalls the Whole Company

The bus factor is the minimum number of people who would have to disappear before a critical process stalls - and for most teams it is uncomfortably close to one. Why static wikis, SharePoint and handover docs never lower it (they capture outcomes, not the reasoning and exception-handling), and how a Company Brain that survives turnover raises the bus factor structurally, with AI employees able to run the concentrated person's routine work across email, Teams, SharePoint, CRM and ERP. Includes verified 2026 data (Deloitte's knowledge exodus, the GitHub truck-factor study, key-person-risk surveys), a euro model of key-person exposure, and a practical playbook to raise the bus factor.

Henri Jung Henri Jung · 33 min read
Building better software Contact us together