Automation Savings Targets: Build a Plan Around a Real Deadline

When a client has a savings target and a deadline, start by defining the benefit and the date it must be realized. Agree the current baseline, identify opportunities, validate their readiness, and schedule discovery, delivery, adoption and benefit measurement. Compare the resulting portfolio with the target and report any gap explicitly.

An executive target is a planning constraint, not evidence that enough automatable work exists. The assessment must be able to recommend a different scope, sequence, timeframe or target when the facts require it.

What does the savings target actually mean?

Ask the sponsor and finance owner to distinguish capacity released, avoided overtime, avoided hiring, reduced external spend and a reduction in an existing budget. Also separate quality, service and risk improvements that may be valuable without producing immediate cash savings.

The UK Government Efficiency Framework distinguishes cash-releasing benefits that reduce spending from non-cash-releasing improvements. That distinction is useful for this planning exercise, although the organization’s own finance policy governs its business case. Government Efficiency Framework.

A useful target statement specifies:

• The outcome and measure, such as a reduction in paid overtime within a defined operation.
• The baseline period and included business units or process types.
• Whether the target is gross benefit or net of relevant costs.
• Whether the deadline refers to deployment, annual run rate or benefits actually achieved by that date.
• The business and finance owners who will validate realization.
• Service, quality and control requirements that must continue to be met.

Do not describe released staff hours as a budget reduction without an agreed action that changes spending. Capacity may instead support additional demand or reduce a backlog.

How do you work backwards from the deadline?

Build the schedule from benefit realization, not only from a development end date. The workflow may need business acceptance, training, phased adoption and a stabilization period before it performs at the planned level.

Identify the latest feasible dates for approved access, agreed rules, representative test data, review capacity and operational handover. Include dependencies on application releases, procurement, business availability or another program. Confirm which activities can overlap and which cannot.

If little time remains, focus discovery on questions that could materially change a near-term decision. A broad mining rollout with unresolved data access may not fit that window, even if mining could be useful later. The workshop versus mining guide explains how to choose a proportionate evidence plan.

How do you find and assess a portfolio of opportunities?

1. Establish a common baseline

Break the target into process areas, case types, demand and effort. Use consistent periods and units. Record the source and confidence of each number. Check whether annualized volumes hide seasonal peaks or a temporary backlog.

2. Identify work that drives the target

Look for manual touch time, repeated corrections, avoidable demand and costly handoffs. Include simplification, better source data and existing system capability in the option set. Use the opportunity identification guide to create briefs that can be compared.

3. Test the important assumptions

For each candidate, validate the baseline, proposed change, affected population, residual work, access and required decisions. Investigate the assumptions that would most change its value or readiness first. A large estimate based entirely on interviews needs a different confidence rating from a sampled, reconciled baseline.

4. Sequence according to value and readiness

Create an early set of ready changes, a later set dependent on validation or access, and a deferred set whose prerequisites do not fit the current window. Do not count later or deferred items as benefits already secured for the deadline.

5. Assign realization ownership

The delivery lead owns building and validating the agreed solution. A business benefit owner also needs to arrange adoption, changed work allocation and outcome measurement. Finance confirms how any financial effect is recognized. These responsibilities may sit with different people.

6. Review scenarios and the remaining gap

Build conservative, central and upside scenarios from explicit differences in coverage, adoption, residual work and timing. Scenarios are planning cases, not statistical confidence intervals. Compare the conservative and central cases with the target and identify what must be proven to close a shortfall.

Worked example: annual run rate versus benefit within six months

Suppose an illustrative workflow could release 100 staff hours per month once fully adopted. It goes live at the start of month four, reaches 50% adoption in month four and full adoption in months five and six. Assume the 100 hours already accounts for ongoing review and support work.

Period	Assumed benefit level	Capacity released
Months 1–3	Discovery and delivery; no operational benefit assumed	0 hours
Month 4	Half of the steady monthly benefit	50 hours
Month 5	Full monthly benefit	100 hours
Month 6	Full monthly benefit	100 hours
Total achieved within six months	Sum of the monthly effects	250 hours
Annual run rate at the end of month six	100 hours × 12 months	1,200 hours per year

This invented example explains timing, not expected delivery performance. The six-month benefit is 250 hours, not 1,200 hours. Neither figure is automatically a cash saving. A delayed launch, lower adoption or increased review effort changes the result.

For a portfolio, make the same calculation by month for each opportunity and remove overlaps. Do not add a full-year benefit to a partial-year benefit and call the result an in-period total.

How do you avoid double-counting savings?

Assign a shared benefit identifier when multiple projects affect the same work. If an ERP change removes manual invoice entry, a proposed RPA workflow cannot also claim all of that original entry effort. Assess the residual work after the planned change.

Separate recurring effort reduction from one-time backlog clearance. Record additional work for review, exception handling, monitoring and support. Keep a growth-capacity scenario separate from observed savings on current demand.

Use a benefits register that records the baseline, counterfactual or comparison basis, owner, measurement method, realization action and timing. For public-sector guidance on establishing a baseline, agreeing benefit metrics and comparing actual results with forecasts, see the GOV.UK Service Manual: measuring the benefits of your service. The portfolio approach here is our practical application, not a sector-specific compliance requirement.

What role should the calculator and technology stack play?

Use the automation savings and capacity calculator to explore a candidate once its inputs are credible. Its fixed assumptions and gross-value scenario do not determine whether a target will be achieved by a particular date. Use a separate phased plan for that decision.

Keep the portfolio in an existing spreadsheet or work management system with a traceable evidence register. A reporting dashboard may help when several owners update it. Do not add a new planning tool before establishing the benefit definitions and review responsibilities.

Choose execution technology by the work: native application features, APIs, workflow automation, RPA or evaluated AI assistance. This guide does not assign license prices or implementation fees. Relevant costs still need an agreed treatment in a full investment case; unknown costs should remain visible as unresolved assumptions.

What should the planning exercise deliver?

Expect an agreed target definition, evidence-backed opportunity briefs, a dependency map, a phased delivery and adoption plan, benefit owners and a monthly realization forecast with uncertainty. Download the benefit portfolio worksheet (CSV) to record the key fields.

At each review, compare actual evidence with the plan. Retain options to improve the source process, narrow scope, change the sequence or revise the target. Our GATE stage supports investment decisions, while BUILD & RUN connects acceptance and handover to operational measurement.

Common questions about savings-led automation discovery

Should we start by dividing the savings target by an average bot saving?

No. That assumes comparable work and proven benefit before opportunities have been assessed. Start with the process baseline and validated changes; then aggregate the effects with timing and overlap accounted for.

What if the identified opportunities do not meet the target?

Report the gap and its causes. Compare changes to scope, readiness, adoption, sequencing and timeframe. Do not increase assumed savings percentages simply to make the portfolio total match the target.

Can a short deadline justify skipping discovery?

A short deadline makes a focused assessment more important. Reuse validated evidence and investigate the assumptions that could invalidate the decision. The scope of discovery can be proportionate without treating unknowns as facts.