Categories
AI Enablement Uncategorized

Curbing AI Overspend via Frameworks

6/26

Situation

Leadership pressure to adopt AI outpaced evidence, risking budget on integrations across ops, marketing, finance, and IT that simpler approaches would solve better.

Task

Evaluate four AI use cases without a shared framework to determine where AI delivers real ROI.

Actions

I built a consistent evaluation lens (using tried-and-true “effort vs. impact”) and ran each use case through it with the rigor the decision required, not vendor messaging.

In the end, presented client leadership (President, COO, Head of IT) with recommendations giving defensible, differentiated answer instead of a blanket “adopt AI” or “don’t”

For the client’s leading storefront-in-a-box platform, I dug into RFM segmentation, Admin API MCP scoping, and the CCPA/GDPR exposure of connecting an AI agent to live customer PII, surfacing scope-limiting as the practical guardrail.

RECOMMENDATION: a qualified “maybe” with privacy guardrails

For the client’s cloud ERP, I weighed scripted automation against LLM/MCP approaches and landed on a Crawl-Walk-Run sequencing rationale instead of defaulting to “AI-first.”

RECOMMENDATION: a scripted-foundation-first path

For the client’s cross-retailer analytics platform, I traced the actual integration constraints (no public API, SFTP-only) to their root cause and concluded Claude Enterprise wasn’t viable as currently configured, then scoped what a custom build would actually cost.

RECOMMENDATION: an explicit no-go on the platform that saved the organization from sinking effort into infrastructure that didn’t exist yet

For org-wide adoption, I flagged a token-economics risk that’s easy to miss: model selection discipline, not just the AI/no-AI decision, determines real-world ROI.

RECOMMENDATION: immediate adoption for org-wide knowledge work using Claude Enterprise

Result

Given a re-forecast, they’ll see cost savings of 45% over the next year.

Categories
AI Enablement

Resolving Slow Personalization via Automation

Jan ’25 – Jun ’25

Situation

Manual content personalization took 2+ hours per asset, from account research to copy-pasting and rewriting messaging, capping output at 2-3 assets per day and causing missed opportunities and team burnout.

Task

Build a pipeline that extracts account context, scores fit, and generates tailored on-brand content.

Actions

Choices & Considerations


I built a TypeScript CLI orchestrating seven specialized AI agents, including an extractor, a scorer, and a content generator, rather than one monolithic prompt, so each stage could be tested, cached, and swapped independently. For ingestion I chose a dual extraction strategy, structured data first (JSON-LD) with an AI fallback for unstructured pages, which kept cost low on clean sources and coverage high on messy ones. I added a configurable criteria-based scoring layer so the team could define what a good-fit account looks like and spend generation effort only where it mattered, and I used Claude for tailoring with two switchable voice modes (leader and builder) to match different buyer personas. A final agent produced structured talking points for sales conversations using the STAR method. The flow was: Source Page → Extract (JSON-LD, AI fallback) → Score (criteria) → Generate (persona mode) → Prep Materials, with logging and caching at every stage.

Tradeoffs

I gave up the simplicity of a single prompt and accepted the overhead of orchestrating multiple agents, plus a CLI-first interface instead of a polished UI. That was worth it at pre-seed, where speed to a working system and the ability to iterate on each stage mattered far more than interface polish. I also accepted some staleness risk from aggressive caching in exchange for a large cut in redundant API calls and cost.

Lessons Learned

I would define the scoring criteria and quality bar with the client’s go-to-market team before writing any agent code, and add human-review checkpoints and an evaluation set earlier, since I tuned output quality by feel before I had a systematic way to measure it.

Result

I cut per-asset personalization time from 2+ hours to 15-20 minutes, a 10x gain that took output from 2-3 to 20-30 tailored assets per day, with a 2-3x lift in response rates.