What is Workflow Automation and What to Automate First

For most growing businesses, workflow automation is straightforward in concept: automated processes run in the background while manual ones require constant human input. Lead routing in a CRM, invoice generation tied to closed deals, Slack notifications triggered by form submissions - all can be automated. Yet teams consistently struggle with where to start, and most approach it backward.

The typical mistake is tackling the largest, most complex workflow first - the one spanning four systems with dozens of manual steps. These sprawling processes hit API rate limits, require enterprise plans across multiple platforms, and take weeks to configure before producing a single automated result. They stall momentum before teams understand how automation tools actually behave in production.

The better sequence: start with high-frequency, low-complexity processes contained within a single application. Build fluency with a few wins, then move to cross-platform workflows of moderate complexity. This approach avoids the common pitfalls that leave companies with half-built automations and frustrated teams.

This guide covers which processes to automate first, how to sequence implementation as a company scales from 10 to 200 employees, and why most automation strategies fail when headcount doubles. What works at 15 people creates bottlenecks at 75, and few companies plan for that transition.

How workflow automation works

Workflow automation connects a trigger event with actions across various systems without human intervention, using three layers: a trigger that initiates the process (webhook, scheduled task, or manual button), conditional logic that determines what happens based on specific criteria, and actions that execute tasks like updating CRM fields, sending notifications, or calling external APIs. As an example, a deal closes in the HubSpot CRM system.

A workflow typically has three different layers: the trigger, the logic, and the action.

Trigger: The trigger is the start point of the automation. This can be a webhook, a scheduled task or a manual task in an automation tool. Webhooks are typically set up to trigger when something new is added to a system. For example, a new record is added to a CRM or a payment is received. Scheduled tasks are automatically run at a set time and date. For example, every Monday at 9am. Manual tasks are started by a user clicking a button in an automation tool.

Logic: The logic or conditional layer outlines the conditions for subsequent actions. For example, "if deal value > $50k then route for approval by VP". Data transformations occur here, such as extracting the domain from an email address, converting from one format to another, or mapping between fields in two systems that are not previously integrated. Error handling in this layer is important as well. For example, "if API call fails then retry twice then alert someone".

Action: The action is where something is actually done as a result of the workflow. This can be anything from writing to a field in a CRM (e.g. HubSpot) to sending a message to a channel (e.g. a Slack channel) to creating a PDF to making an API call to an external service (e.g. a payment gateway).

In addition to the native functionality of workflow automation in applications, a number of no-code platforms have emerged that allow non-technical individuals to automate tasks and business processes. Platforms including Zapier and Make support integration with a wide variety of applications. These platforms typically enable simple workflows to be set up quickly, but as the number of actions and conditions increases, the limits of the platform typically become apparent, such as workflow timeout limits, API rate limits on the number of requests per minute/second, and cost considerations for large volumes of tasks.

For large processes and large amounts of data, automation via native tools (e.g. Salesforce Flow) or even via custom-built APIs that can be called from other systems (e.g. a CRM plus a project management tool), can be very powerful. These types of workflow automation support long sequences of actions with many branches and even parallel paths. They can also connect to legacy systems for which there is no Zapier integration. However, the process of setting up these automations can be very complex and may require the services of a developer. For simple automation, the effort required to set up such automation is not worth it.

Most teams will take a hybrid approach: using the native workflow functionality in their CRM for processes that live within the CRM (i.e. all Salesforce.com or all HubSpot deals) and custom automation integration for more complex processes that need to integrate multiple systems or include legacy systems that have no public API or Zapier integration.

What to automate first

Teams first look to automate the work that is most common and lowest value add for a human - the work that does not require judgment and is taking people away from revenue-generating activities. For example: data entry (such as adding a new customer to a CRM), synchronizing data between different systems, automated notifications (that would otherwise be forgotten by a human), and automated lead routing (so that new leads automatically go to the right sales person every time).

Here are four quick ways for small to medium size businesses to start automating workflows and see a fast return on their investment in automation software.

CRM data entry and enrichment

Workflow automation can transfer form fill data into the CRM and enrich each contact with firmographic data from Clearbit or ZoomInfo to ensure that the database is up to date and has been cleaned of any duplicate accounts that would have been added by web forms.

Lead routing and assignment

Workflow automation can route leads to sales representatives in a round robin fashion, automatically create a new task for the representative to follow up with the lead in x amount of time, and automatically send a notification to the representative in Slack instantly (i.e. within seconds) of a new lead being created.

Cross-system syncing

Customer data must automatically stay up to date across all tools in a system (e.g. a CRM, support tool and billing/accounting system). Workflow automation eliminates time-consuming CSV imports for updating customer data in deals, in support tickets (e.g. in Zendesk) and other systems.

Status updates and notifications

Deals update stages. Contracts get signed. Support tickets get escalated. Automated workflows can keep teams in the loop for tasks that currently get lost in the din of daily work.

Automating lead routing

Automating lead routing directs new leads from online forms to sales representatives using either basic round-robin or territory-based distribution built into CRM systems like Salesforce or HubSpot, with more complex lead scoring and territory rules requiring custom code through Apex in Salesforce or custom workflow actions in HubSpot. Many online forms on a website get created as leads in a CRM (customer relationship manager) such as Salesforce or HubSpot.

For basic lead routing no custom code is required. Most CRM systems such as Salesforce or HubSpot support round-robin as well as territory-based routing right out of the box.

As the process becomes more complex with additional data points for scoring leads and more intricate rules around territory boundaries (such as where lines of ownership over accounts fall), custom code is generally needed in order to set up the leads workflow to function as required (Apex in Salesforce, custom workflow actions in HubSpot).

Automating data entry and enrichment

Automating data entry and enrichment ensures that customer information updates in a CRM immediately sync to support software like Zendesk or Intercom, billing systems like Stripe or QuickBooks, and marketing software like Mailchimp or Braze, typically through native platform integrations rather than custom middleware, though teams must test edge cases like merging or deleting contacts where silent sync failures often go unnoticed for weeks. Data sync for support teams. As customers interact with support teams on the phone, support agents should be viewing the very latest information about customer accounts in a CRM. To this end, any update to information in a CRM (e.g. where a customer is, what plan they're on, etc.) should immediately sync with support software (e.g. Zendesk, Intercom), billing systems (e.g. Stripe, QuickBooks), and marketing software (e.g. Mailchimp, Braze). While it's possible to build this sort of data sync out with a custom API (or "middleware") layer, there are typically native integrations between platforms that can get the job done faster and with less chance for human error. The mistake that most people make with the built-in sync tools in their platforms is assuming that they handle edge cases just fine. They don't. Test the sync with a few scenarios (e.g. merging contacts, deleting contacts, marking contacts do-not-contact) and see where it breaks. Often this kind of silent failure can go on for weeks without anyone even realizing it's happening.

Automating customer onboarding

Workflow automation around customer onboarding eliminates manual handovers between teams by automatically triggering actions such as provisioning accounts, sending welcome emails, creating project tasks, and notifying implementation teams when a deal closes in the CRM, streamlining operations when processes are correctly configured and maintained as business needs evolve after six months. The following example illustrates this: A deal has been closed in a CRM.

These are typical business processes that can streamline operations if done correctly and then kept up to date after 6 months when reality usually has changed.

When workflow automation isn't worth it

Workflow automation isn't worth the investment when processes change monthly, haven't been standardized for at least three months, run infrequently with low volume, require high human judgment like enterprise lead routing, or are still experimental and being tested, because light process documentation through templates and checklists delivers most of the value for a fraction of the setup cost. Most teams fail with workflow automation, because they do not even understand what it's good for.

When automation is premature or wrong:

• Process is not yet standardized - every deal closes differently, sales reps negotiate different terms and conditions with customers and the finance department manually assesses the odd case here and there. Wait until a process has been executed consistently for 3 months and then start automating the bits that get repeated over and over.

• Low volume. A process that runs infrequently and takes a short time each time to complete, but requires significant time to set up workflow automation, would need to run many times just to break even on the setup time.

• High judgment required. For example, routing out enterprise leads by a program automatically. It makes sense to have workflow automation generate leads and then have an experienced person such as a sales person manually assign the lead to the best source (e.g. human sales person) for the deal. The automated workflow will make mistakes and it will be more expensive to fix in the automated workflow than for the sales person to assign manually.

• The process is in experimental phase and will change while the team is testing different parts of the process (as an example, an onboarding process where every 2 weeks the team is changing up different aspects of the process and testing out different pieces). Manual process execution will be more effective in the short-term than automating a process out and then having to immediately rebuild and redeploy the process in the following sprint.

Light Process Documentation Is The Answer To Most Problems

The vast majority of problems can be solved by implementing process documentation (templates, checklists, etc.) that are then supported by a few simple tools and deliver much of the value of having automated a process for a fraction of the cost.

Process documentation (even something as simple as templates and checklists) can be set up extremely quickly and deliver the majority of value for a small fraction of the cost to set up. Additionally, process documentation via video (Loom, etc.) is extremely easy to update for processes that change on a quarterly basis.

Choosing an implementation approach

Start by leveraging automation within a CRM's native functionality for processes that reside completely within that system, then use no-code platforms like Zapier or Make for straightforward multi-tool workflows with moderate volumes, and finally consider custom API integrations for complex workflows, high-volume situations, legacy system connections, or when no-code platforms hit timeout or cost constraints. For workflows that span across multiple tools, no-code platforms will generally be the best bet.

But here's where they break down.

As a rule of thumb, no-code platforms have an upper limit of workflows that they can handle and often this is outside of the normal scope of what they have been built to do. There are certain limits on workflow automation that can be overcome by purchasing more expensive plans, but often the cost for high volume situations can become expensive. In addition, currently there are no good no-code integrations that handle complex error recovery in applications such as automation where if a step fails in a workflow the user often has to go to each of the affected applications and fix the data.

Custom API integrations may be needed for more complex workflows or when volume or timeout constraints are encountered with no-code integrations. Integrations can then be built to synchronize for instance data between a CRM and an ERP system but also with legacy systems like 15 year old SQL databases for which no connectors (Zapier, Make etc.) have been developed. High volumes of data can then be processed with required functionality such as retries, idempotency checks and proper error logging. Expensive up front, but with virtually unlimited flexibility once it has been built.

Frequently Asked Questions

What is workflow automation in simple terms?

Workflow automation automates business processes that are usually performed manually by a human. In order to automate a workflow, rules are set up that describe the process step by step. Workflow automation handles recurring tasks where human intelligence is not required.

What are some examples of workflow automation?

The first example of workflow automation is to automatically add people who fill out a web form as new contacts in a Salesforce CRM. The second example of workflow automation is to automatically notify team members in Slack every time a deal reaches the 'verbal commitment' stage. The third example of workflow automation is to automatically add customer information into a CRM and have it automatically synchronize with billing software so manual updates to two systems of record are not needed.

How do you create a workflow automation?

There are three different ways to build workflow automation. The most common is to build directly within a platform's native workflow automation tools (i.e. Salesforce Flow Builder, HubSpot workflow tool, etc). Another way is to use an integration platform like Zapier or Make to connect between different systems to create the automation outside of the native platform. Lastly, there is custom code development for the most complex automation situations.

What are the costs of workflow automation for small businesses?

Workflow automation pricing can vary depending on if a business uses the native workflow functionality of current systems and/or additional systems outside of current tools. Many automation tasks are included in mid-tier plans for systems like Salesforce, HubSpot, etc. For more complex workflows that span across multiple systems, such as customer, billing and payment information, then an integration platform would be used. According to Zapier's pricing page (2026), Zapier's Professional plan starts at $19.99/month, and according to Make's pricing page (2026), Make's Core plan is $12/month.

What should you automate first in your business?

Automate high volume / high impact work that could have large errors or negative impacts if done incorrectly. Another idea is to automate lead capture and transfer to a CRM. This is typically one of the best places to start because losing inbound leads due to manual entry mistakes can be a huge problem many companies face.

---

Gable Innovation can help assess a business's needs for workflow automation, and select the most suitable CRM, AI and workflow automation software for a business to implement. Gable can also assess the integrations required between proposed software tools and the business's current systems. Book a 30 minute free discovery call at gableinnovation.com.