Business process automation

Business process automation services: what BPA covers, what it costs, and where it fails

Updated 7 September 2026 10 min read

Short answer

Business process automation services map an operational workflow end to end, then replace the manual handoffs inside it with software that runs on your existing systems. A serious BPA engagement delivers four things: a process map with touch time and cycle time measured at every step, a costed exception log showing where work stalls, an automated straight-through path with named approval gates, and an SLA you can report against. For a 20-300 person company the work runs as a fixed-fee process audit followed by per-workflow builds, not a platform migration and not a license purchase. If a provider quotes before mapping your handoffs, they are pricing a guess.

Process mapping first: touch time, cycle time, and exception rate recorded per step
Built on the ERP, CRM, and helpdesk you already run — no migration, no license resale
Every workflow ships with a named owner, an exception path, and an SLA it is measured against
You keep it: source, credentials, process documentation, and the runbook

Book a process audit Model the payback first

Written for ops, finance ops, and supply chain leads in 20-300 person companies: distribution, light manufacturing, logistics, professional services, and healthcare admin.

How we decide which processes get automated

Published because "it depends" is the least useful answer a BPA consultancy can give an ops lead with a budget cycle to hit. These are the six tests a process has to pass before we will quote a build.

01Volume: enough instances per month that a percentage improvement is a real number. Below roughly 100 instances a month, most multi-system builds do not clear their own cost.
02Handoff count: how many times the work changes hands or systems. Handoffs are where cycle time hides — a three-minute task that waits nine hours for the next queue is a cycle-time problem first and a task problem second.
03Straight-through processing rate: what share of instances complete with no rework today. A process already at 95% has little left to win; one at 60% usually has three or four recurring exception causes worth fixing before any automation is built.
04Stability: the steps have not changed in the last quarter and no reorganisation or system replacement is scheduled. Automating a moving target produces rework, so we stabilise or wait.
05Integration surface: every system in the chain needs an API, a scheduled export, a database we can read, or a dedicated mailbox. A closed system with none of the four is a hard stop and we say so before quoting.
06Ownership: a named process owner who can approve a rule change. Workflows without an owner drift out of policy within two quarters, whoever built them.

Why BPA programs stall after the mapping phase

Rarely because the technology was hard. Usually because the map described the intended process, the exceptions were never costed, and no baseline existed to prove anything moved.

The map is of the process as designed, not as run

The documented flow has six steps. The real one has eleven, four of which are a person checking something in a second system because they stopped trusting the first. Automating the documented version produces a workflow that fails on contact with the real one, and the team quietly goes back to the spreadsheet.

Nobody costed the exceptions

A workflow that handles the clean path and drops everything else creates work rather than removing it, because now someone has to notice the failures. Exception handling is not the edge of an ops automation project; it is usually a large share of the cost and the part that decides whether the workflow survives contact with a real week.

The time is in the handoffs, not the tasks

Add up the touch time on an order-to-fulfillment chain and it might be twenty minutes. Measure the cycle time and it is two days. The difference is queues — work sitting in an inbox waiting for someone to look. Automating the twenty minutes and leaving the queues in place moves the cost number a little and the SLA not at all.

There is no baseline, so the result is unprovable

Without straight-through processing rate, median cycle time, and cost per instance measured before the build, nobody can show what changed. Budget for the second workflow then depends on a feeling rather than a number, which is how BPA programs end after one project.

How a BPA engagement runs

Audit first, then build one workflow, then decide. You can stop after the audit and take the findings to your own team, to a platform vendor, or to a different consultancy.

01
Process mapping
A structured session per process with the people who actually run it, not only the manager who owns it. We record every step, every system boundary, every handoff, and every place work waits — including the informal checks that never made it into the SOP.
You get A swimlane map per process: steps, owners, systems, and handoffs marked.
02
Baseline the numbers
Instance volume from the system of record rather than from memory, touch time per step, queue time between steps, and the straight-through processing rate. Then the exception log: what went wrong, how often, and how long each category took to resolve.
You get Cost-of-manual-work table, median cycle time, STP rate, and a categorized exception log.
03
Rank the candidates and quote them
Each process gets a build estimate, an integration risk note, an expected effect on cycle time and STP rate, and a payback period. Ranked by payback, not by how well the workflow demos. Where automation is the wrong answer — a deleted step, a changed form, a renegotiated supplier deadline — that goes in the same document.
You get Ranked shortlist with fixed build prices, payback, and explicit do-not-automate flags.
04
Build the straight-through path
One workflow, end to end: triggers, data extraction, validation rules, the approval gates that stay human, exception routing to a named person, logging, and SLA reporting. It runs in parallel with the manual process until the two agree on live data.
You get A live workflow with exception routing, an audit log, and an SLA dashboard.
05
Measure it, then extend or hand it over
Same three numbers as the baseline, measured again after four weeks of real volume. Then a decision: we build the next process, your team takes over ops automation in-house with the runbook, or the program stops because the remaining candidates do not pay back.
You get Before-and-after on STP rate, cycle time, and cost per instance, plus the runbook and admin access.

What the arithmetic looks like on an order intake chain

A model, not a case study. Every figure below is an input you should replace with your own — this is the same table our audit fills in per process, and the one that decides what we recommend building first. The point is that you can check it.

A distributor receiving purchase orders as email PDFs. Each order is keyed into the ERP, credit-checked against terms, allocated against stock, handed to the warehouse as a pick note, and confirmed back to the customer. Five manual handoffs, three systems, one team of order coordinators.

Purchase orders received per month: 640
Manual handoffs per order (intake, credit, allocation, pick note, confirmation): 5
Manual handoffs per month (640 × 5): 3,200
Average touch time per handoff on the clean path: 3.5 min
Clean-path hours per month (3,200 × 3.5 ÷ 60): 186.7 h
Loaded hourly cost of an order coordinator: $32
Clean-path cost per month (186.7 × $32): $5,973
Straight-through processing rate today (orders needing no rework): 68%
Exception orders per month (640 × 32%): 205
Extra handling per exception: chasing, re-keying, re-confirming: 25 min
Exception cost per month (205 × 25 min ÷ 60 × $32): $2,733
Total monthly cost of running order intake by hand: $8,706
Same figure over twelve months: $104,472

Now run the same table against a target state: four of the five handoffs automated, the credit decision left as a human approval gate, and the STP rate lifted to 90% by fixing the three recurring exception causes. Clean-path cost falls to about $1,195 a month (640 × 1 × 3.5 min × $32) and exception cost to about $853 — roughly $2,050 against $8,706. Two caveats worth stating. Most of that saving arrives as redeployed capacity and shorter cycle time rather than a smaller payroll, and the business case should say which one you are buying. And the inputs decide everything: at 60 orders a month the same table totals about $813, a five-system ops automation build never pays back, and the right answer is a better intake form.

What we take on, and what we refuse

The fastest way to tell a BPA consultancy from a platform reseller is to ask what they will not do. Here is ours, in writing, before the first call.

In scope

Process mapping and baselining: touch time, cycle time, STP rate, exception categories
Cost-of-manual-work analysis per process, with a ranked shortlist and payback periods
Ops automation design: triggers, validation rules, approval gates, exception routing, escalation
Implementation on n8n, Make, Zapier, or direct API integration against your ERP, CRM, and helpdesk
Document and email processing: PDF and invoice extraction, classification, matching, enrichment
SLA instrumentation and reporting, so the workflow is measured rather than assumed
Written SOPs and a runbook, so the process survives staff turnover and our departure

Not in scope

Replacing your ERP, WMS, CRM, or accounting system — automation runs on top of what you have
Selling or reselling BPM and iPaaS licenses with a margin attached
Multi-quarter "operational transformation" programs that end in a deck rather than a running workflow
Automating a process that is broken or mid-reorganisation — we flag it and stop, rather than building on it
Attended RPA screen-scraping bots against systems that have a real API, because they break on the vendor’s next release
Model training projects that need a data science team you do not have and we are not
Headcount-reduction plans. We produce the cost and cycle-time numbers; what you do with the recovered capacity is your decision, not our deliverable.

What it costs, next to what the same work costs elsewhere

Published in full because the first thing an ops lead needs to know is whether this is a four-figure engagement or a six-figure one. We sit deliberately at the bottom of the market for process automation consulting services, and the reasons are listed under the table rather than left for you to guess at.

Process audit

$490

Typical market rate: $2,500–8,000

You know the ops overhead is real but cannot say which process is the expensive one.

Up to three processes or departments mapped
Swimlane map with every handoff and system boundary marked
Baseline table: touch time, cycle time, STP rate, cost per instance
Ranked shortlist with fixed build prices and payback periods
Credited in full against your first build

First workflow build

from $1,900

Typical market rate: $6,000–15,000

The process is chosen and you want it running, instrumented, and documented.

One workflow end to end, built on your existing systems
Approval gates, exception routing, and escalation to a named owner
Audit logging and SLA reporting
SOP and runbook written for whoever inherits it
Parallel run against the manual process, then two weeks of tuning

Ops automation retainer

$690 / month

Typical market rate: $3,000–8,000 / month

Workflows are live and you want them watched, maintained, and extended.

Monitoring, incident response, and exception-queue review
One new or reworked workflow per month
Integration maintenance as your vendors ship changes
Quarterly re-measurement of STP rate and cycle time
Cancel with 30 days notice

The build price is a floor, not an average. A workflow crossing two systems with clean structured input lands near $1,900; one crossing five systems with scanned PDFs, per-customer pricing rules, and a compliance log lands well above it. System count and input quality move a quote more than anything else you will be asked about. You get a fixed number in writing after the mapping session and before any build work starts, and the $490 audit comes off it in full. No free pilots — a pilot with no baseline proves nothing anyway.

Why this is a fraction of the BPA consultancy price

A price this far below market usually means one of two things: junior people learning on your process, or a loss-leader with the real money in license commission. Neither is what is happening here, so it is worth being specific about where the difference comes from.

There is no sales layer to fund

The person on your first call is the person who maps the process and builds the workflow. No account director, no pre-sales solution architect, no bid team. On a traditional BPA engagement those roles are a real share of the invoice, and they add nothing to the workflow.

The audit is productised, not bespoke

Same intake, same baseline table, same exception taxonomy, same deliverable format every time. Consultancies bill discovery as custom work because for them each one genuinely is. The template already exists here, so you are paying to have it filled in rather than invented.

The plumbing repeats even when your process does not

Document intake, field validation, an approval gate, an exception queue, retry logic, a log — that skeleton is the same on the tenth order-intake build as the first, and we price the tenth. Your business rules are specific to you; the mechanism underneath them almost never is.

No license margin, so no reason to oversell

We take no vendor commissions and resell no platform. Revenue comes only from work delivered, which is exactly why the audit is willing to conclude that a process should be simplified rather than automated — that outcome costs us a build and saves you the budget.

BPA consultancy, platform vendor, or an internal ops automation team

All three deliver working automation. They fail in different places, and the honest answer depends on how much recurring process work you have and whether it is permanent.

Criterion	BPA consultancy (us)	Enterprise BPA / iPaaS vendor	Internal ops automation team
Time to first workflow in production	3-6 weeks	3-6 months including procurement	4-6 months including hiring
Cost shape	Fixed audit, then fixed price per workflow	Annual license plus professional services days	Salaries, tooling, and management overhead
Where the process knowledge ends up	In your runbook and SOPs, handed over	In the vendor’s platform and their consultants	In your team, until they leave
Best when	You have a handful of expensive, well-defined ops processes	You need governed, audited workflow at hundreds of processes	Process change is continuous and central to operations
Main risk	Engagement ends before your team can maintain it — the runbook and parallel run exist to cover this	Paying for platform capacity you never use, and being unable to leave	One person becomes a single point of failure
Sensible company size	20-300 people	500+ people or regulated at scale	150+ people with a real process backlog

Which of these you actually need

Read down to the line that matches your situation. Three of the six answers do not involve hiring us, and they are the honest recommendation in those cases.

You can name the process, it has three or more handoffs, and it runs at least a hundred times a month
Start with the audit, then one build. This is the standard path for business process automation consulting and the one with the shortest payback.
You know ops overhead is high but cannot say which process is worst
Buy the audit alone. Take the shortlist and decide afterwards whether we build it, your team does, or nobody does — roughly speaking, that is what the audit is for.
The process is one trigger and one action, like a form filling a spreadsheet row
Do not hire anyone. A Zapier or Make subscription and an afternoon will do it, and paying consulting rates for that is a waste of your budget.
The process changes every few weeks, or a system replacement is scheduled
Do not automate yet. Document and stabilise it first — the free SOP generator is enough for that step — and revisit once the steps have held for a quarter.
You already own a BPM or iPaaS platform with unused capacity and an internal admin
Use what you have paid for. If you want help choosing candidates and writing the specs, the audit is useful on its own; the build is not something you need to buy from us.
You are over 300 people with dozens of workflows a year and a governance requirement
Build an internal ops automation function, and evaluate a governed platform alongside it. Consultants are the wrong structure when the work is permanent, continuous, and audited.

A one-week baseline you can run before paying anyone

This is the first half of our audit, done by hand on one process. It costs nothing, it is yours to keep, and if it produces nothing worth acting on you have saved yourself a fee — which is a legitimate outcome, not a failed test.

Day 1Pick one workflow with at least three handoffs — order intake, purchase approval, onboarding, claims, invoice matching. Write down every step and who touches it, sitting with the people who run it rather than reading the SOP.
Day 2Get last month’s instance count from the system of record, not from anyone’s estimate. Count the emails, the tickets, the ERP records. Volume is the input everything else multiplies by, so guessing it invalidates the rest.
Day 3Time the touch time of each handoff, start to finish, including the cost of switching attention to it. Separately, record the queue time between handoffs — the gap between one person finishing and the next starting.
Day 4Add the two up per instance. Touch time is what automation removes; touch time plus queue time is your median cycle time, and that is the number your internal customers actually feel.
Day 5Count how many of last month’s instances completed with no rework, no chase, and no correction. Divide by total instances: that is your straight-through processing rate. Count it rather than estimate it — the counted figure is almost always worse than the remembered one.
Day 6Sort the exceptions into causes. Nearly every ops process has three or four recurring causes covering most of the volume — missing data, a late upstream handoff, a pricing exception, an approval nobody chased. Fixing the top cause is often cheaper than any automation.
Day 7Cost it: instances × handoffs × minutes × loaded hourly rate, plus exceptions × extra minutes. Then check the integration surface of every system in the chain, and name the one person who can approve a rule change.

If one process costs more than about $1,500 a month, has three or more handoffs, sits below a 90% straight-through processing rate, and every system in the chain has an API or an export, it is worth a conversation with someone — us or anyone else. If your top process clears none of those, ops automation is not this quarter’s bottleneck and the honest recommendation is to spend the budget elsewhere. If it fails only the integration test, the blocker is procurement, not automation.

Looking for something more specific?

Frequently asked questions

What are business process automation services?

Mapping an operational process end to end, measuring what each step costs in touch time and cycle time, then replacing the manual handoffs with software that runs on your existing systems. The deliverables are a process map, a baseline of cost and straight-through processing rate, a running workflow with approval gates and exception routing, and documentation. Anything that stops at the map is analysis, not a service.

How much do business process automation services cost?

Our process audit is $490 and is credited in full against the first build. A workflow build starts at $1,900 as a fixed price, and ongoing ops automation support is $690 a month with 30 days notice. For comparison, BPA consultancies typically charge $2,500-8,000 for discovery alone and $6,000-15,000 per workflow, and enterprise platform routes add an annual license on top. We quote a fixed number in writing after the mapping session — open-ended day rates are the arrangement most likely to go wrong for both sides.

What does business process automation consulting include, as opposed to implementation?

Consulting is the mapping, baselining, and ranking: which processes are expensive, what each would cost to automate, what the payback is, and which ones to leave alone. Implementation is the build. We do both, and we sell them separately on purpose, because the honest output of consulting is sometimes that no build should follow. Ask any provider whether their consulting fee is contingent on a build being sold afterwards.

How do I choose a business process automation company?

Four questions separate them quickly. Will they map your handoffs before quoting, or does a price arrive first? Do they take license commission from any platform they recommend? Who owns the workflows and credentials at the end? And can they show you the arithmetic behind a payback claim rather than a percentage from a vendor case study? A provider who cannot answer the second question directly is a reseller with a consulting page.

What is the difference between BPA consulting and RPA?

RPA drives the user interface of an application the way a person would — clicking, typing, screen-scraping. BPA is the broader discipline of redesigning the process and integrating systems properly, usually through APIs. We use interface-level automation only when a system genuinely has no API, export, or database access, because UI bots break on the vendor’s next release and their maintenance cost is routinely underestimated.

Do we need a BPM or iPaaS platform license for process automation?

Usually not at 20-300 people. Governed BPM platforms earn their license cost when you are running hundreds of processes with audit and compliance requirements. Below that, orchestration on n8n, Make, or direct integrations does the same job for a fraction of the total cost of ownership. If you already own a platform with unused capacity, the right answer is to use it, and we will say so.

How long does a process automation consulting engagement take?

Mapping and baselining takes two to three weeks for up to three processes. A first workflow typically goes live three to six weeks after that. The usual delay is not build time but system access — credentials, API keys, sandbox environments, and whoever has to approve them. Starting those requests during the mapping phase is the single most effective thing a client can do to shorten the timeline.

Which ops automation projects pay back fastest?

High-volume, multi-handoff, rule-based chains where the input arrives as structured or semi-structured documents: order intake, invoice matching, purchase approvals, claims triage, onboarding checklists. What they share is countable volume and a rule a person can state out loud. Processes that are genuinely judgement-based every time pay back as an approval queue with the context assembled for the approver, which is a smaller but still real win.

What company size are business process automation services worth it for?

Roughly 20 to 300 people is where this shape of engagement fits. Below that there is rarely enough instance volume to pay back a multi-system build, and the small-business route is a better fit. Above it you generally need an internal ops automation function with governance, and our useful role becomes helping you stand that up rather than running it indefinitely.

Who owns the workflows and the process documentation afterwards?

You do. Workflows run in your accounts wherever the platform allows it, and you receive the source, the credentials, the swimlane maps, the SOPs, and a runbook written for whoever inherits the process. The retainer is optional because the business model does not depend on you being unable to leave.

What happens when an automated workflow breaks?

Every build ships with logging and a failure path that escalates to a named owner rather than failing silently — silent failure is the one outcome worse than the manual process. Without a retainer, the runbook covers diagnosis of the common cases. With one, we monitor the exception queue and fix it. Integrations change on the vendor’s schedule rather than yours, so something eventually will break.

Next step

Start with the process audit

Name the process that costs you most and tell us roughly how many times a month it runs. We will come back with what it likely costs you today, whether the handoffs are worth automating, and what a build would involve.

Process audit request

We will tell you which handoffs to automate first

Work email and one sentence about the process. No account needed.

The $490 audit is credited in full against your first build, so it costs nothing if you go ahead — and the shortlist is yours either way.

Where is the manual handoff cost concentrated?

Work emailWhich process should we look at? optional

What happens next

We ask for one or two real examples of the process, map the handoffs, and come back with a baseline table — cost per month, median cycle time, straight-through processing rate — plus a ranked shortlist with fixed build prices. If nothing on it clears the payback bar, we say so and stop there.