Demo How It Works Cheat Sheet Stories Objections Downloads Flag an opportunity
How It Works

Invite us into your next qualified meeting.

Bring the Nexus Black team into your customer conversations. We'll demo, handle the hard questions, and close together. This page is everything you need to set that up.

Nexus Black AI engineers on site at customer facilities
This is a product in production. William Grant & Sons. £8.4M estimated annual savings. Not a proof of concept.
We ship in weeks, not months. 4–12 weeks to value vs. 12–24 months for traditional enterprise software.
We go deep. We read P&IDs, diagnose faults from voice and video, and find savings hidden in engineering drawings. This isn't surface-level AI.

What's Different

"We were very keen to avoid the whole proof of concept trap. We said, do something at scale. And that's what we've been able to do with Nexus Black. The pace at which we've gone from idea to something real has been fantastic."

Badri Narasimhan, CTO, William Grant & Sons

"You should never lose a deal because of a product gap. If there's a wider market opportunity, we'll build it — and we'll build it fast. Days and weeks, not quarters. That's the whole point of how this team is set up."

Kriti Sharma, CEO, Nexus Black

Products, not custom solutions

Every capability is a repeatable, scalable product. But the products are built with customers. The team embeds on-site, shadows workers, shapes the product to hit the mark on first delivery.

12–24 mo
Traditional time to value
4–12 wks
Nexus Black time to value

Resolve

AI field service and maintenance. In production with William Grant & Sons. Active pilots with Eight Group, Ecolab, BGIS.

£8.4M
Estimated annual savings at WGS

Look at what it actually does

An engineer speaks into the app: "Look into the pressure drop we're seeing across the heat exchanger after about 10 minutes of runtime, it recovers briefly before dropping again." Here's what happens next:

Resolve fault tracing interface — AI analysing a pump with voice input

The system cross-references equipment manuals, evaluates past failure patterns, reviews the P&ID for that section of the plant, maps every connected relationship, and starts reviewing work order history. The engineer gets a diagnosis with evidence — not a chatbot guess. All while wearing gloves, not typing.

"I wanted to test Resolve on a leak we'd already raised a fault for. It came back with the same diagnosis I'd made from just two images."

Operator, Girvan Distillery — second-generation employee, William Grant & Sons
How deep does Resolve go? Here's an example.

We taught it to read the most important document in process manufacturing.

Every process manufacturer has P&IDs — Piping and Instrumentation Diagrams — that map every pipe, valve, sensor, pump, and safety device. It's the single drawing that tells you how the whole process works. Most engineers need years to read one fluently. The retiring ones take that knowledge with them.

Resolve reads them — with reasoning, not template matching. Multiple AI models cross-check every extraction: one reads, a different one challenges, and the first corrects. By the time it's done, it has the full picture.

P&ID diagram with AI-identified failure modes — bearing wear, fouling, seal failure

WGS Mash Column P&ID — GIBD-GWA-0484. Resolve identified bearing wear, fouling, and seal failure risks automatically.

When it's not sure, it leans in closer.

If a tag is ambiguous — does TT-014 connect to pump P-26 or P-28? — the system zooms in, crops that area, and runs a separate AI agent to verify. The same instinct a veteran engineer has with a magnifying glass, except it never gets tired.

It doesn't just read the drawing — it thinks about what could break.

For every asset, Resolve maps connections, likely failure modes, and detection methods. Problem with pump P-26? Instant answer — 3 connected instruments, feeds into heat exchanger HX-04, most likely failure is bearing degradation, detectable via vibration sensor VT-014. Pre-computed. No waiting.

It catches things the humans missed.

Pressure relief valves flagged with testing requirements. Emergency shutdown paths traced end to end. Operational insights surfaced with real numbers. And yes — it has found genuine errors on existing P&IDs that passed manual engineering review.

Proof point: We pointed Resolve at a WGS distillation column P&ID. It identified every safety-critical element — pressure relief valves, emergency protection systems, overpressure setpoints. It found cascade loops that tighten quality variation by 50–70%, and traced an energy recovery system worth £1–2M per year in savings. All from one drawing. Minutes, not weeks.
Backstage

The stuff behind the stuff

When the CTO leans in after the demo and asks about hosting, security, or integrations — here's what you need.

Hosting
IFS-managed Azure environment. Single-tenant — each customer gets an independent deployment with isolated data stores. No shared database.
Data residency
Same residency rules as IFS Cloud. No customer data leaves the IFS Azure perimeter for AI processing. No data sent to external AI providers without governance controls.
Security
AES-256 at rest, TLS 1.2+ in transit. SOC 2 Type II, ISO 27001, GDPR. SSO, RBAC, tenant isolation aligned with IFS enterprise policies.
Integrations
IFS Cloud, IFS Apps 10, SAP S/4HANA, Oracle, Salesforce, Snowflake, Databricks. Connecting a new data source means configuring a new connector, not re-engineering the product.
Accuracy
90–95%+ on first pass, tested against real customer data. Every code change and model update runs through evaluation sets. If accuracy regresses, it doesn't ship.
Auditability
Every step logged with full lineage: what data was used, what reasoning was applied, what result was produced. All responses grounded in customer documents — the model must cite its sources.
Positioning

When to lead with what

Lead with Nexus Black

When the customer has multiple pain points across operations — maintenance + scheduling + compliance. Position Nexus Black as the AI capability across their operations, then go deep on whichever product fits.

Lead with Resolve

When the conversation is specifically about field service, maintenance, or technician productivity. Lead with the outcome, land with the proof points.