The right actor for every task: person, automation or AI

Who this is for

• Leadership teams under pressure to do something with AI who want a grounded view of where it genuinely helps.
• Chief operating officers and HR leaders planning how roles will change as automation spreads.
• Organisations that have automated piecemeal and now have a patchwork of bots, tools and manual workarounds.
• Smaller organisations wanting practical help to free up their people’s time without a large technology programme.

Work used to have one actor. Now it has many.

Until recently, nearly every task was done by a person. Today the same process might involve employees, rule-based automation, AI models and external partners. Most organisations have not redesigned work for this.

• AI tools adopted team by team, with no view of the whole process.
• Automation that speeds up one step and creates a queue at the next.
• People doing tasks a machine could do, and machines attempting tasks that need human judgement.
• Unclear accountability when something goes wrong, and no plan if an automated step stops working.

The underlying issue is that nobody owns the question of how work is done. HR owns the people, IT owns the systems, the function owns the outcome, and the join between them is left to chance.

What the work covers

• Task-level analysis of priority processes.
• Actor allocation: whether a person, automation or AI is the best fit for each task, and at which point.
• Hand-off design: clean transitions between actors, with clear triggers and information passed on.
• Cost and viability testing: whether the proposed mix makes economic sense, now and as costs change.
• Role and capability implications: how jobs change and how to support people through it.
• Governance and fallbacks: who is accountable for each step, and what happens if an actor is unavailable.

How we approach it

The idea borrows from just-in-time manufacturing. On a well-run production line, the right part arrives at the right station at the right moment. We apply the same thinking to work: the right actor, for the right task, at the right time.

That leads to some deliberately unfashionable conclusions. Whether running AI at every step is viable depends on the future cost of compute, and on whether AI is actually better than a person or a simple rule for that task. Often the best design uses AI sparingly, where it adds most.

It also leads to visibility. Every hand-off between actors is a point where work can be delayed or a decision can go unexplained. Designing those hand-offs well avoids the AI black hole, where results emerge but nobody can say how.

Evidence

At a large outsourcing organisation, offer-to-deployment for overseas employees was cut from 24 days to 5 across an end-to-end workflow that ran through several functions.

NasIQA, 10k’s platform, is designed around this principle and is live, supporting more than 100,000 employees across hire-to-retire processes.

Bringing the two practices together

From Human Performance come role design, capability and decision rights. From Technology come automation, AI and the means to govern it. Right actor for the task is where they meet, and the design can run on NasIQA or on the systems you already have.