# Replacing Planners With AI Increased Errors by 27%: Steve Morris on Strategy-First Automation

> TechRound spotlights NEWMEDIA.COM CEO Steve Morris on enterprise AI: fix the model before automating. Strategy-first use, guardrails, and human review beat blind scale.

- Type: AI news
- Page: https://completeaitraining.com/news/replacing-planners-with-ai-increased-errors-by-27-steve/
- Published: 2025-10-14
- Updated: 2025-10-14
- Author: Jeroen Erne
- Jobs: Executives and strategy
- Reading time: 2 min

### TechRound Features NEWMEDIA.COM CEO Steve Morris on Enterprise AI Strategy: Caution Over Blind Automation

TechRound published "Can Companies Automate Their Way Out of Decline?", featuring insights from Steve Morris, Founder & CEO of NEWMEDIA.COM. The piece challenges a common executive impulse: scale automation to fix performance decay. The message is clear-automation amplifies whatever exists in your operating model, good or bad.

**"Don't automate a broken business model,"** Morris warns. **"Automating a broken business model just makes it decline faster. The winners are the people who begin by looking hard at the business they're in. Let the business model drive technology."**

#### Why over-automation backfires

Morris cites an enterprise case where automation replaced senior planners and fulfillment errors jumped 27%. The issue wasn't the tool-it was misaligned design and the removal of tacit knowledge. AI made decisions faster, but in the wrong direction.

- Objective functions mis-specified: the system optimized speed, not accuracy or margin.
- Tacit expertise removed: critical edge cases were handled worse than before.
- No guardrails: exceptions lacked human review, so small errors compounded.
- Process debt ignored: broken workflows were scaled instead of fixed.

#### Strategy-first AI: a practical approach for executives

- Start with unit economics: identify where decision quality moves margin, churn, or working capital.
- Map the decision, then the model: define inputs, constraints, error costs, and escalation paths.
- Pilot with stage gates: run A/B or canary releases with rollback triggers and pre-agreed kill criteria.
- Design for human-in-the-loop: route high-impact or low-confidence cases to experts.
- Tie incentives to outcomes: owners are accountable for cycle time, quality, and customer impact, not "model accuracy."

#### Where AI delivers right now

- Forecasting and allocation with planners reviewing low-confidence calls.
- Anomaly detection across finance, logistics, and security with clear escalation flows.
- Lead scoring and journey modeling tied to conversion and LTV, not vanity metrics.
- Content and operations workflows that pair automation with editorial or QA checkpoints.

#### Governance that prevents costly mistakes

- Risk policy: define acceptable error rates by decision type and customer segment.
- Controls: canary deploys, circuit-breakers, and automatic rollback on KPI breaches.
- Observability: monitor model drift, data lineage, and decision logs for auditability.
- Standards: align with frameworks such as the [NIST AI Risk Management Framework](https://www.nist.gov/itl/ai-risk-management-framework?ref=completeaitraining.com).

#### Metrics that prove value

- Quality: error rate, defect density, returns/chargebacks, complaint volume.
- Throughput: cycle time, SLA attainment, backlog velocity.
- Economics: gross margin, LTV/CAC, cash conversion cycle.
- Customer: CSAT/NPS on affected journeys, time-to-resolution on exceptions.

#### Executive checklist before green-lighting automation

- Have we fixed the process and clarified decision rights, or are we scaling process debt?
- Is the cost of a false positive/negative quantified and reflected in objectives?
- Do we have baseline metrics, a clean control group, and a defined win threshold?
- What is the exception path and who owns it (with response SLAs)?
- Are security, privacy, and compliance obligations documented and tested?
- Is there a training plan for operators and reviewers, not just the data team?

#### NEWMEDIA.COM's stance

NEWMEDIA.COM blends technical depth with commercial focus across SEO, performance marketing, infrastructure hardening, enterprise UX, and advanced automation. Their approach is simple: translate algorithms into measurable revenue and resilience. **"AI is powerful,"** Morris said, **"but without the right scaffolding, it can make bad decisions at scale."**

For the original feature, visit [TechRound](https://techround.co.uk/?ref=completeaitraining.com). If your leadership team needs structured upskilling on AI use cases and governance, explore executive-focused programs at [Complete AI Training](https://completeaitraining.com/courses-by-job/).

---

This is the Markdown form of https://completeaitraining.com/news/replacing-planners-with-ai-increased-errors-by-27-steve/. The whole site is mapped at https://completeaitraining.com/llms.txt.
