This is the newsletter of Pluris, a network of pre-vetted applied AI experts and AI-native service organizations that specialize in helping businesses design and implement AI solutions that drive value.

Two free ways to use us. Browse the Pluris AI Use Case Library — 163 published cases now, with about 171 more in the queue (334 in the dataset and growing) — for real AI implementations from our network: what got built, what it cost, and what changed once it shipped, organized around the problem you're solving. Or submit an AI need through an Expert Review: we'll anonymize it, match it to experts with relevant experience, and return 3–4 perspectives with pricing.

A lot of AI programs generate real activity and never move the P&L. This one is about a company that fixed that. In today’s case, A PE-backed SaaS company had rolled out developer tools, stood up a support chatbot, and told its board AI would move margin this year, and yet, not a dollar of it had. They brought in Jimmy Bilgani and AI Momentum Partners, who started from the financials, shipped one workflow in under four weeks, and underwrote $700K from it. We also have a healthcare case study that spedrun the same approach: Bonsai Labs found the single step slowing every hire and put a voice agent on it, cutting time-to-hire 5X.

Key takeaways from this week's cases:

If you've named AI as a value-creation lever, told the board it will move margin this year, and can't point to a single dollar it has produced, this a case for you.

A PE-backed vertical SaaS company, $100–150M in revenue at roughly 40% EBITDA margins, was exactly there. Developer tooling was rolled out, the CTO was engaged, a support chatbot was live, and everyone believed the hard part was done. The hypothesis, which AI Momentum Partners says it hears constantly, was "we're already doing AI, we just need to organize it and scale it." But none of it touched the P&L. There were explicit cost-takeout targets and a finite hold clock, yet no financial model behind the AI work and no operator accountable for capturing value.

When AMP dug in, the internal story fell apart. Developers had 90% adoption of Cursor and Copilot, sitting on top of a codebase that was about 60% a legacy language and proprietary IDE those tools couldn't touch, so engineers were copy-pasting code by hand. The live support chatbot ran on a knowledge base only about 20% usable. The AI policy was one most employees had never seen. Plenty of activity, but almost none of it moving numbers the board cared about.

So what did they do? AMP started from the financials.  It reviewed roughly 75 enterprise artifacts (financials, SOPs, system data, historical project data), interviewed people from the exec suite to the frontline, and mapped the workflows until it was clear where automation would actually pay, then cut about 18 candidate initiatives down to a 7-initiative roadmap scored on near-term margin impact. 

The first pilot went at one high-materiality workflow: professional-services document drafting, the BRDs, SOWs, and requirements analyses that ate senior-consultant weeks. AMP built it on the client's existing Azure and Microsoft stack (Microsoft 365, Power Automate, SharePoint) with a model-agnostic agent layer, currently Azure OpenAI's GPT-4.1 and routing individual tasks to Claude via Amazon Bedrock where it does better. It shipped in under four weeks with no procurement cycle, and senior-consultant drafting fell from one to two weeks to about two hours, roughly 95%.

Be precise about the numbers, because AMP is. The 95% is realized and measured on live documents. The $700K of NPV over three years and the projected 5–6% Year-1 margin improvement are underwritten, modeled against a real workflow and not yet banked. The hardest part was political: getting function leaders to cooperate once cost-takeout put headcount on the table. The drafting workflow runs in daily production today, with SOWs the next one in line.

The playbook: diagnose from the P&L, ship in four weeks. The method is more portable than the workflow it produced. Four steps:

An idea. In your next ops or portfolio review, ask one question: which AI initiative on our roadmap has a dollar figure attached to it? If the honest answer is "none of them," you've found the problem before spending another dollar, and that's exactly where AMP starts.

Best fit for: PE-backed services and SaaS businesses (roughly $15M–$500M) under pressure to show AI on the P&L inside the hold period, especially those with heavy, repetitive knowledge work and an enterprise software stack they already pay for but underuse.

→ Want to talk to Jimmy? Reply to this email and I'll make the intro. Or request one directly on the case study page.

A leading healthcare scale-up, north of $300M in revenue with a 40-plus-person HR team, was still taking 30 days to fill a role. The bottleneck was mundane: the lag between when someone applied and when a recruiter first called them back. Good candidates went cold inside that gap and accepted other offers.

Bonsai Labs put a voice AI agent on that first contact. It calls applicants within minutes of an application, runs a structured screening conversation, scores the candidate, and books the ones who pass straight onto a recruiter's calendar, with transcripts, summaries, and sentiment analysis flowing back into the ATS so recruiters walk in already briefed. The idea is lifted from sales, where speed-to-lead drives conversion, and applied to hiring: reaching candidates in minutes instead of days lifted how many converted to a recruiter call, and that compressed overall time-to-hire roughly 5X against the prior 30-day process.

Best fit for: High-volume hiring teams where speed-to-contact drives whether candidates convert: healthcare systems, staffing firms, and field-service or hourly employers with large recruiting funnels and an ATS to integrate against.

→ Want to talk to Csongor? Reply to this email and I'll make the intro. Or request one directly on the case study page.

You can learn a lot about what companies are actually trying to do with AI just by watching which case studies people open.

These are the most-read use cases in our AI Use Case Library this week. Real problems, real numbers, and every one has a "connect with the expert" button that goes straight to the person who built it.

Interest in the library keeps climbing. Here's what people read most this week. A couple of these have shown up before, which, I think, tells you something.

There's plenty more where those came from. Every one is a real company with a real problem, one click from the person who already solved it.

If one of these is your problem too, the library is open. Browse the full AI Use Case Library →

Working on something where one of these patterns might fit? Reply to this email — I'll either make the intro to the right expert directly, or open an Expert Review and come back with 4-5 expert perspectives and pricing in a few days. Both are free, and I read every reply.

— Stu

Proof of Work: Applied AI Case Studies. The newsletter of Pluris.