Chasing Elephants, Not Peanuts (3 of 3)
In the first two pieces, I wrote about elephants. Big problems that don’t announce themselves as urgent until the ground has already shifted beneath our feet. This final piece is about where one of those elephants most often lives.
It lives in the first mile of data.
- Part 1 of 3 (2-minute read): Chase Elephants, not Peanuts
- Part 2 of 3 (2-minute read): The Elephant in the Room: Why Manufacturing Data is Under Pressure
The Moment Before Platforms Matter
As AI, sustainability reporting, and cybersecurity requirements continue to develop, organizations increasingly use operational data to support a wider range of enterprise outcomes than it was initially intended for. Increasingly, that data is expected to be:
- Trustworthy, to support analytics and AI use cases
- Auditable, to withstand emerging ESG and Scope 3 assurance expectations
- Secure, in line with modern OT and cybersecurity guidance
The catch is where these expectations first meet reality.
It doesn’t start in the cloud.
It doesn’t start in dashboards or digital twins.
It starts at the device.
The Reality of Industrial Estates
Most industrial environments are not greenfield; they are brownfield sites with assets that are often:
- 20–30 years old
- Built for reliability, not connectivity
- Using serial, fieldbus, or proprietary protocols
- Physically connected, but not readily accessible by modern digital standards
This isn’t a failure of platforms. Instead, it reflects how industrial systems were originally designed and operated. In practice, that reality has raised a practical question many customers now face:
How can operational data be captured at the source in a way that is secure, credible, and repeatable – without increasing operational risk or burden on the plant?
Until teams answer that question, downstream platforms struggle to deliver reliable value.
When Risk Becomes a Price Tag
Outside the plant, one signal is getting louder: Cybersecurity insurance is becoming more difficult to obtain, renew, and justify cost-wise. Underwriters are asking stricter questions about industrial environments: segmentation, access pathways, monitoring, and how operational data moves beyond the plant boundary.
For many organizations, this isn’t just theory. Premiums are rising, exclusions are increasing, and coverage is tightening. In that environment, a trusted first-mile approach isn’t simply a preference for architecture. It becomes a crucial way for companies to show risk controls that are easier to explain and defend.
A Subtle but Important Distinction
Within Autodesk’s ecosystem, it’s useful to be clear about roles. Tandem plays an increasingly important role as a human-governed system of record for operations. It provides shared context, visibility, and accountability – qualities that are especially important in regulated and safety-critical environments.
However, first-mile acquistion is a different class of challenge. Teams that must securely acquire data from legacy OT environments – using push-only, provenance-preserving, device-level controls – constitutes a different category of challenge. In many plants, that work sits in a separate layer, focused on governed data ingress at the OT boundary rather than on visualization, analytics, or workflows.
The pattern, and the work ahead
The key point here isn’t about any specific software. It’s the architectural philosophy:
- Capture once.
- Trust early.
- Reuse responsibly.
This is where Autodesk’s approach is already strong. Tandem offers the human-governed system of record – shared context, accountability, and operational decision support. And Tandem Connect plays a key role in integrating data into that environment once it becomes accessible and suitable for integration. The industry is still catching up on the step before that – especially in brownfield estates.
How do you establish secure, governed, provenance-preserving data transfer at the OT boundary in ways that align with modern security guidance and are practical for plants to implement?
That “first mile” is not just a simple integration task. It often involves a separate layer of work – one that must be handled carefully and credibly so that downstream systems can trust it. Autodesk is actively addressing this challenge – both directly and through its broader ecosystem – because customers do not benefit from duplicating operational data across various tools. The long-term goal is a reliable operational data pathway that supports Tandem and enables other enterprise systems that customers already rely on.
Why the first mile is different
Solutions that live here tend to assume tough constraints:
- Devices that aren’t IP-native
- Inbound connectivity that’s restricted or discouraged
- A need to establish integrity before data moves
- Data that must serve multiple purposes over time
They operate at the OT boundary, not the IT boundary. Their job isn’t analytics or visualization. It’s to make data credible before it moves. That’s often where complexity – and risk – can be most effectively reduced.
Interoperability becomes evidence
As regulatory expectations mature, organisations are assessed not just on outcomes, but on how data is generated and governed.
That shifts conversations:
- from tools → trust
- from features → evidence
- from ownership → stewardship
In many environments, the most constrained point – the place where friction matters most – is still the first mile.
One bite at a time
Big structural problems don’t move all at once. A more durable path usually looks like this:
- Secure, governed capture at source
- Trust with operations and security teams
- Responsible reuse across platforms
- Human accountability for decisions
That’s not about expanding scope. It’s about executing well under real-world conditions – so everything downstream becomes simpler, safer, and easier to defend.
That’s often how elephants actually move.
One deliberate bite at a time
