Your customers have more problems worth solving

A conversation starter — March 2026

The customers you already work with

The transport company hired you to put ChatGPT on their data. Then you found out their drivers were quitting before 90 days, costing millions. So you built an attrition model. Then they wanted to see it in an app. Then they wanted to flag at-risk employees in real time. Then they wanted a communication system to act on those flags.

Seven projects from one relationship. Not because anyone planned it that way — because once you're inside a company's operations, you see what's actually going on. And what's actually going on is always messier and more interesting than the original ask.

Every customer is like this. A company that hires you for data engineering has scheduling problems, onboarding problems, compliance problems, forecasting problems. Their operations run on email and spreadsheets and one person's memory. These problems were always there — but until recently, solving them with custom software was too expensive to justify. Now it's not.

What changed

AI makes a specific class of work economical for the first time: turning unstructured input into structured data, routing it through a business process, and putting a human-in-the-loop interface on top. Voicemails become tickets. Insurance claims get classified and scored. Driver interview data becomes a retention prediction.

This is what your engineers already do — data pipelines, ML models, extraction, classification. The new part is that the same capabilities now power operational systems that people use every day. Dashboards, queues, approval workflows, notification systems, CRM integrations. The data architecture and pipeline orchestration — the hard parts — are what your team already thinks in. The user-facing layer is the straightforward part.

And the data hygiene work that every one of these customers needs (consolidating sources, resolving entities, cleaning duplicates) is itself an AI pipeline problem. It's not a prerequisite you get through before the interesting work starts. It's the first engagement, and it makes everything that follows more valuable.

Classes of problems that show up everywhere

Once you're inside a mid-market company's operations, the same patterns repeat:

Intake and triage — Requests arrive through multiple channels (phone, email, text, web forms). Someone manually reads, classifies, routes, and tracks them. When that person is on holiday, things get dropped.

Workforce intelligence — Who's likely to leave? Who should be hired? Are people compliant with certifications, training, hours-of-service rules? Companies track this in spreadsheets or not at all.

Operational forecasting — Revenue by region, demand by product, resource allocation by season. The data exists across multiple systems. Nobody has time to pull it together, so decisions get made on gut feel.

Document processing — Contracts, claims, compliance filings, inspection reports. Someone reads them, extracts the relevant parts, and enters data into another system. Hundreds of times a day.

Customer communication — Confirmation emails, status updates, follow-ups. Written manually, inconsistently, often late. The information to automate them exists — it's just not connected.

Each of these is a real engagement. Each one reveals the next. And each one makes the customer's operations measurably better — not by replacing people, but by giving them systems that handle the routine so they can focus on the exceptions.

Why these problems compound

The transport company's attrition model is a good example. It didn't replace recruiters — it told them where to focus. And the pattern generalizes: any labour-intensive business with high turnover has the same problem. Logistics, staffing, field services, construction, hospitality.

The model already exists. The question is how to bring it to the next company — and whether it becomes a product. Does the attrition model plug into Workday? Can it be sold as a standalone subscription? Is there a vertical sales motion that uses the $1.5M savings case study to open doors across an entire industry?

And once you're inside one of those companies for attrition, what else is broken? Scheduling. Onboarding. Training. Compliance — DOT logbooks, hours of service, certification tracking. These companies have problems everywhere.

Why consultative sales matter here

These problems don't surface in an RFP. Nobody writes "we need a unified intake system with AI extraction and completeness scoring." They say "our dispatch person is overwhelmed" or "we keep losing drivers" or "our claims processing is too slow."

Getting from that conversation to a scoped engagement requires someone who can sit with the operations team, understand the actual workflow, and design a solution from the capabilities available. That's consultative sales — and it pairs well with a transactional product entry point.

The product sale (DataLoom, a data engineering project, an ML model) gets you in the building. The consultative layer looks around and asks: what else is going on here? What does this team do manually? Where does data move between systems by copy-paste? What process depends on one person?

The answers to those questions are the next 3-5 engagements. And each engagement compounds — the data foundation from engagement one makes engagement two faster. The attrition model from one customer becomes a product for an entire vertical. The patterns repeat across industries.

Commercial situations, for our consideration

Account expansion on existing customers — Juan or the existing relationships open the door. A consultative team runs discovery workshops, identifies high-value operational problems, designs engagements scoped at outcome-level pricing. Spatialedge engineers deliver. Engagements are larger and stickier because they're solving business problems, not fulfilling technical requests.

Vertical productization — The attrition model works for any high-turnover labour business. Package it. Maybe it's a Workday plugin. Maybe it's a standalone subscription. Maybe it's the entry point into an industry-specific sales motion where the case study opens doors at every similar company. Same logic applies to insurance claims processing, retail demand sensing, revenue forecasting.

European expansion — Mid-market European companies with operational complexity. Timezone-aligned with SA. Discovery and relationship management run from Europe, engineering delivered from Stellenbosch. Commission-based for sourced deals, paid advisory for pricing and positioning work.

Platform referrals — AWS, Anthropic, Databricks all have account teams sitting with customers who want to do something real with the platform. The default referral is Accenture — 6 months of scoping and a team of 30. A company that can show up and build a working system in weeks is what those account teams actually need.

GG hires Spatialedge MLEs for customer engagements — Goose Group has its own customers who need operational systems built (field services, construction, retail). Spatialedge engineers join those build teams directly. GG provides product leadership and customer relationships, Spatialedge provides the engineering. Real work, real revenue, and a way to see the collaboration in practice.

What we'd want to explore on the call

Which existing customers have the most untapped operational problems?

What does the engineering team's capacity look like for this kind of work alongside current projects?

How does product development work today — is there appetite to take engagement patterns and turn them into repeatable products?

What's the ideal commercial arrangement — agent, advisory, or both?