We trace why things happened
Your systems record what happened. We reconstruct the chain of decisions, information, and timing that caused it — and show you what would have changed the outcome.
Temporal forensics for business
Every business failure has a traceable cause. Most are never traced.
Post-mortems rely on memory and narrative. Dashboards show correlations. Neither can tell you which decision in March caused the outcome in September, what information was available at the moment of that decision, or what would have happened if someone had acted differently.
We reconstruct the actual sequence: who knew what, when they knew it, what they did with it, and where the chain broke. You get a causal narrative backed by evidence you can follow and defend.
- Which decision in Q2 actually caused the Q4 loss?
- Was the warning signal visible before the approval went through?
- What would have happened if we'd acted on the first alert?
- Who had access to the information and when?
- Can we prove what we knew at the time — not just what we know now?
A forensic investigation, not a dashboard installation.
How we work
Scoping
You bring the question. A decision gone wrong, a failure to explain, an outcome nobody can account for. We scope the investigation around what you need to know.
Collection
We ingest the relevant data — system logs, documents, communications, workflow records, approval chains. The more complete the record, the more precise the reconstruction.
Reconstruction
Raw data becomes a temporal model. We resolve entities, establish sequences, and build the causal graph connecting actions to outcomes through time.
Analysis
We trace the causal chains and test counterfactuals. Where the evidence is ambiguous, we say so and show what additional data would resolve it.
Findings
A written report, an inspectable causal graph, and a data appendix showing the evidence behind each link. You get a defensible account, not a narrative.
Data & trust
Forensic analysis requires access to real data — often sensitive data. We've designed our approach around that reality.
Runs in your environment
Our tools can run on your infrastructure, in your cloud tenancy, or in whatever jurisdiction your data requires. Nothing needs to leave your perimeter.
Your data stays yours
Each engagement is standalone. Your data is never pooled, shared, or used to train models. The causal graphs we build belong to you.
We work with what you have
Server logs, CRM exports, event streams, email archives, approval workflows. The richer the record, the deeper we can trace. We'll be honest about what's possible with what's available.
Whatever agreements you need
NDAs, data processing agreements, custom security requirements — we expect these conversations and welcome them.
One decision. One incident. One question you can't answer.
You don't need to commit to a full engagement to see what temporal forensics can do. Start with a single decision you need to explain, or one incident you need to trace to its root.
Typical first engagement: 2–4 weeks, centred on one question. You'll see exactly how we work — and whether it's worth going deeper.
Temporal forensics applies wherever decisions ripple through systems and time.
Where this pays off
We don't claim deep domain expertise in every vertical — we partner with people who have it. What we bring is the ability to reconstruct what happened and trace why.
Finance & Investment
Reconstruct how a position evolved, trace which signals were visible before a decision, and stress-test what alternative actions would have produced.
Legal & Compliance
Establish what was known and when. Produce evidence chains for regulatory inquiries, litigation support, and defensible decision records.
Healthcare & Clinical
Trace patient journeys, treatment outcomes, and clinical decision chains with complete temporal audit trails.
Operations & Supply Chain
Trace how disruptions cascaded through your network. Identify which upstream decision caused the downstream failure. Test mitigations retroactively.
AI & LLM Pipelines
Trace exactly why a pipeline produced a particular output. Compare model configurations. Replay failures with full provenance.
HR & Organisational
Map how your organisation actually operates versus the org chart. Trace hiring decisions, training effectiveness, and where real authority sits.
Marketing & Attribution
Causal attribution, not statistical correlation. Trace actual conversion paths, test counterfactuals, and find what's genuinely driving outcomes.
Work with us
We work with advisors who want to offer temporal forensics to their clients without building the capability in-house.
Accountants & Advisors
Help clients understand not just what happened in their numbers, but the decision chains that produced them.
Management Consultants
Back your recommendations with traceable causal evidence rather than plausible narratives.
Law Firms
Establish what was known when, reconstruct decision sequences for litigation support, and challenge hindsight-biased timelines.
Fractional Executives
Bring forensic analytical capability to your portfolio companies without building an internal team.
M&A Advisors
See how acquisition targets actually operate. Map real decision flows, not org chart fictions.
About TNT Research
TNT Research Limited is a UK company providing temporal forensics — the practice of tracing decisions through time to establish what happened, why, and what would have changed it.
We reconstruct causal chains from your existing data, producing evidence-backed accounts that hold up to scrutiny. No correlation-based guesswork. No narrative hand-waving. Traceable chains from decision to outcome.
Chris Abbott
14 months building financial systems at American Express. Over a decade shipping DetectRight, a production distributed system competing globally in device detection. Five books on audio technology including one that reconstructed a composer's entire development timeline from source code forensics.
If you've got one decision you can't explain, start there.
Let's talk
Email us with a two-sentence description of the decision, incident, or question. We'll reply with whether it's a fit and what data we'd need.
chris@tntresearch.co.uk