Skip to content
Start a conversation
Menu

Technology

Know who, or what, made every decision, and why

We help organisations put governance around automation and AI so that every decision can be traced, explained and defended. That gives leaders, regulators and employees confidence, and gives you a plan for when things go wrong.

Who this is for

  • Organisations in regulated sectors that must be able to explain how decisions affecting customers or employees were made.
  • Risk, compliance and audit leaders asked to sign off automation or AI they cannot see inside.
  • Technology and operations leaders scaling automation who need to keep control as the number of bots and AI tools grows.
  • HR leaders using automation or AI in recruitment, screening or other people decisions.

The questions boards are starting to ask

Work used to have one kind of actor: people. Now a single process may pass through employees, rule-based bots and AI models. That brings speed, but raises questions many organisations cannot yet answer:

  • If an automated decision is challenged, can we show which step made it and on what basis?
  • Could an AI tool be making decisions that discriminate, and would we know?
  • Who is accountable when an automated step gets it wrong?
  • If an AI service becomes unavailable, does the process stop?

We call the gap the AI black hole: work goes in, results come out, and the steps in between are not visible. In regulated sectors that is a liability. In smaller organisations it is still a risk, especially where automated decisions affect people’s jobs or pay.

What the work covers

  • Inventory: where automation and AI already make or shape decisions, including tools adopted informally.
  • Decision mapping: which decisions are made by people, bots or AI at each step, and who is accountable.
  • Governance framework: policies, approval rules, risk tiers and oversight proportionate to each use.
  • Step-level monitoring, so each action can be traced to its inputs and outcome.
  • Resilience planning: fallback actors and procedures if an automated step or AI service fails.

How we approach it

We start with the cost and value question before the risk question. Running AI at every step may not be viable as compute costs evolve, and it is often unnecessary. Choosing the right actor for each task tends to produce systems that are cheaper, more resilient and easier to explain.

Then we make each step visible. Governance works best when it is designed into the process rather than added afterwards. If each step records who or what acted, on what information and with what result, explanation becomes a matter of looking it up.

Because our Human Performance practice works on decision rights and accountability, we treat governance as an organisational question as well as a technical one. A clear audit trail is of little use if nobody owns the decision it records.

Built for regulated sectors

NasIQA, 10k’s platform, is built for organisations in regulated sectors that need explainable automated decisions. It is live and supports more than 100,000 employees across hire-to-retire processes.

Questions

We only use a few AI tools. Do we need governance?

Proportionate governance, yes. A short inventory and clear rules on what tools may be used for, especially in decisions about people, removes much of the risk at little cost.

What is explainability in practice?

Being able to show, for any individual decision, what happened at each step, which actor took it, what information it used and why the outcome followed. It is a record, not a theory.

Does this slow down our automation programme?

It should not. Designing governance in from the start is usually faster than retrofitting it after a problem.

What happens if the AI goes down?

That should be planned for. We help define fallback actors and procedures for each critical step, so the process degrades gracefully rather than stopping.

Do we have to use NasIQA?

No. Governance advice applies to whatever technology you use. NasIQA is an option where you want explainability built into the platform.

Your organisation is the system. Let’s talk about how it could work as one.

Start with a conversation about what you’re seeing. No pitch, and no obligation.

Start a conversation