Our references

Thinkers and builders.

What the client needed, what we built with their teams, and what it changed.

AdoptionFinancePrivate markets

For a global private markets firm, we built a secure drafting assistant for investor relations that the team actually uses daily.

The objective: Investor relations and product teams spent much of their week retrieving answers the firm had already produced. Leadership wanted the right use cases, built within strict confidentiality and European regulatory constraints, and adopted by the teams.

Mapped the work through interviews from junior analysts to partners, crossed with systems records
Co-created use cases with staff from every level, filtered on value and willingness to adopt
Built a model-agnostic drafting assistant on the firm's validated answers, inside its secured environment

>50%of the target team uses it daily, unprompted

Several thousand investor questions answered through the assistant, and the same secure foundation now supports broader research and due-diligence assistance.

An illustrative engagement profile drawn from our work with private equity firms and their portfolio companies; details are representative and not attributable to a single client.Read the case study ↗

GEOTrade organization

A six-month GEO program measuring and improving what AI engines tell consumers about a global trade organization's category.

The objective: Leadership asked what the AI engines were actually telling consumers about its category, and whether that could be measured and improved.

Tracked 500 consumer prompts weekly across major AI answer engines, with sentiment and cited sources
Found one online community platform cited in 46% of tracked prompts, then baselined its sentiment
Selected 94 priority threads where credentialed contributors engage factually, within each platform's rules

36% → 14%negative citation share in AI answers, in eight weeks

67 → 3critical threads cited by AI models

Positive sentiment in AI answers rose from 19% to 33%, past the six-month target, with 4 of 6 KPIs beaten by week 8. The organization now treats category reputation as a managed, instrumented channel.

Program metrics are actual; the organizational example in the final section is a composite drawn from our board-level work.Read the case study ↗

AI auditLuxuryRetail

A 60-day AI opportunity audit for a PE-owned global luxury retail group, cross-checking systems data, public data and interviews.

The objective: The board wanted to identify where AI would create measurable value and present investable options within 60 days. Transformation initiatives kept losing momentum between decision and execution.

Classified 7,612 support tickets into a ten-category taxonomy, validated at 94% accuracy
Analyzed 236 public reviews, the 203-style live catalog, BI usage and AI answer presence
Interviewed the full chain of command, from chief executive to customer service team

33AI use cases ranked by return and effort

45.6%of the support queue was genuine demand

The board funded the first three use cases within the quarter and adopted the audit's re-scan rhythm for its reviews.

Read the case study ↗

GEOLuxury

A three-month GEO program that rewrote what five AI engines said about a luxury jewelry brand's after-sales service.

The objective: The brand had fixed its after-sales problems operationally, but ChatGPT, Claude, Gemini, Perplexity and AI Overviews still drew on old complaints, while two competitors were positioned as service benchmarks on weaker evidence.

Mapped every forum, review site and article shaping AI answers, model by model
Ran the same prompts for direct competitors to quantify the gap across all five engines
Placed authoritative content in GQ, luxury media and trade publications, and displaced legacy forum sources over three months

13% → 30%brand website citation rate

11 → 0after-sales negatives in AI answers

Heritage citations rose from 15% to 50%, resale negatives went from 6 to neutral, and AI visibility monitoring now runs alongside traditional media monitoring.

Read the case study ↗

AI ImplementationPrivate marketsIndustrial services

A Wharton field study valuing five AI use cases in a PE-backed industrial services contractor, then the first build: AI bid generation and outreach.

The objective: The sponsor's thesis was margin expansion through applied AI. The pricing logic lived in one estimator's head, and the back office ran on spreadsheets and institutional memory.

Mapped workflows end to end on site and valued five use cases standalone and in combination
Captured the pricing logic into a structured, auditable knowledge base reviewed by the senior estimator
Built AI bid generation and a CRM-driven outreach loop for the business developers

1.45 monthspayback on the first build

836+ hrssaved per year on bids

Four of five use cases NPV-positive standalone, five of five together, with ROI from 50% to 342%. About $800K of projected incremental revenue, with no additional hires.

Study and implementation figures are actual; the incremental revenue is a projection.Read the case study ↗

AI ImplementationMediaServices

An LLM-agnostic AI platform with four custom agents for a creative agency, rolled out in four weeks and handed to the team.

The objective: Pricing a proposal took two to three hours and a first-draft deck took a day. The team used ChatGPT and Claude individually, with no shared knowledge and no governance.

Connected the platform to the messaging workspace, email and ten years of proposals
Built pricing, pitch deck, daily briefing and master agents from the team's own workflows
Launched with power users, ran two training sessions and fifteen days of iteration

2-3 h → minutesto price a proposal, team estimate

15users live on one governed platform

A first-draft deck now takes twenty to thirty minutes instead of four to eight hours, and the team builds its own agents after the handover.

Time savings are the team's estimates from the first weeks, not a twelve-month measurement.Read the case study ↗

TrainingFinancePrivate markets

A nine-week cohort where non-technical investment professionals became builders, shipping custom AI apps on their own workflows without a line of code.

The objective: A finance association wanted its alumnae to operate as AI-augmented investment professionals, not AI-literate ones: working agents on real sourcing, diligence and reporting workflows, built without code.

Two weeks of foundations: LLMs, agent building, data confidentiality and firm-level governance
Seven weeks of live building: a shared chief-of-staff agent, then PE and asset-management tracks
A Demo Day where each pod shows one working agent, with before and after metrics, to a jury from the association and sponsor firms

9 weeks2 of foundations, 7 of live building

0 codecustom apps built in plain English, one per participant

A 2.5-hour masterclass for about 150 finance professionals became a 50-person cohort, with a second track per role and a playbook for the next ones.

Read the case study ↗

AI ImplementationMarketplaceSmall business

An AI content factory, a voice agent and automated re-engagement for a fishing guide marketplace: +300% revenue in a year.

The objective: Guides had no time to produce content, calls went to voicemail while they were on the water, and past customers were never followed up.

Built a content factory producing listings, fishing reports and destination pages from the guides' own notes
Deployed a voice agent that answers, qualifies and books when guides cannot pick up
Automated re-engagement of past customers in each guide's voice

+300%revenue, year on year

~2xconversion with the voice agent

Monthly traffic up 50% from the content factory and a 30% return rate from re-engagement. The guides no longer create content themselves.

Read the case study ↗

AI ImplementationFinanceFintech

Self-hosted LLM tooling for the sales and solutions team of a 30-person fintech, replacing shadow AI with an auditable system.

The objective: The team was pasting client data into public chatbots on personal accounts. A company that sells risk detection to banks could not leave that unanswered.

Deployed the models inside the company's own environment, with every prompt logged and reviewable
Built proposal, questionnaire and meeting-prep tools on the company's approved material
Rolled out with power users first, then two training sessions for the rest of the team

90%team adoption in four weeks

-30%admin tasks for the sales team

Shadow AI eliminated and knowledge kept in-house, self-hosted and auditable.

Read the case study ↗

They trusted us

Insights

What we read in the market

Read the journal

Your company could be the next reference.

Book a call. We start from the work, not from the tools.

Book a call charlotte@neon-slate.com