Tantalum starts with evidence, not a capability catalogue. We diagnose the structural constraint, decide what should change first, install the operating mechanism, measure what happens in real use, and transfer recurring ownership back into the company.
Find the constraint that is actually worth solving.
We map how work, information, decisions and money move through the company. We separate symptoms from causes, locate founder and key-person dependency, and establish the evidence behind the primary constraint.
Choose the smallest operating change that deserves to happen next.
The roadmap is a decision, not a wish list. We define what should change, why this intervention outranks the alternatives, what the client must own, and what evidence would prove the decision right or wrong.
Make the operating change real.
Depending on the constraint, that may mean decision rights, workflows, reporting, integrations, custom software, automation, AI agents, documentation or a new operating cadence. Technology is used when it serves the operating outcome — not because Tantalum needs to sell technology.
Watch what happens in real operations, not in a workshop.
The installed capability goes live, gets used, fails in places, gets corrected and is measured against the baseline. We distinguish adoption from activity and expected impact from what actually happened.
Make the client stronger, not more dependent on us.
Recurring ownership moves into the company. We document what the team needs to run, monitor and recover the capability, then decide whether the next constraint warrants another intervention or the engagement should close.
Every expensive implementation failure we have seen started the same way: a tool was chosen before anyone agreed what the process was. We do not let a vendor decide your operating model.
A large share of the work inside a growing company exists only because two systems don’t talk. Automating it preserves it. The first pass is always deletion.
Most companies have three. Reconciling them is a permanent tax paid in meetings, spreadsheets and mistrust of the numbers. Choosing one is a design decision, not a preference.
Nothing is delivered as a document and left to a team that is already at capacity. Each phase goes live, gets used, and gets corrected before the next begins.
If the system depends on us the way the company depended on the founder, we’ve moved the bottleneck rather than removed it. Ownership transfer is part of the design, not the wrap-up.
We start with the operating problem, not with AI. Agents get defined work, access to the right context, bounded authority, measurable responsibilities and explicit escalation paths. If deterministic software is the better tool, we use deterministic software.
Being specific about the boundaries saves both sides a quarter. If what you need is on the right, we will say so early and point you somewhere better where we can.
The Business MRI is a short directional screen. If it shows a material structural signal — or the stakes are already high enough to warrant evidence — the paid Valuation Gap Scan validates the hypotheses and produces the 30/60/90 operating roadmap.
How much of the company still depends on specific people?
What you get back