Best Automation Workflows for Virtual Assistants: Complete Guide from Beginner to Advanced (2026)

Layer 1 — The Foundation (Start Here)

One automation platform + one project management tool + one form tool.

This three-tool foundation covers 80% of the best automation workflows for virtual assistants at the beginner and intermediate level.

Automation platform: Zapier for simple linear workflows (under 5 steps, mainstream apps, fast setup). Make for complex workflows (branching logic, data transformation, bulk operations). Start with Zapier if you are new to automation, migrate specific workflows to Make when Zapier’s logic depth becomes the constraint.

Project management: ClickUp for operations-heavy VAs (tasks, automations, templates, dashboards). Notion for documentation-heavy VAs (SOPs, knowledge bases, client wikis with AI integration).

Form tool: Tally (free, clean, Notion-style interface) or Typeform (more polished client experience). Either integrates with Zapier and Make as the trigger for most onboarding and lead capture workflows.

👉 Zapier vs Make for Virtual Assistants — complete comparison with pricing and workflow examples.

Layer 2 — Client Operations

(Add at Intermediate Level)

Contract tool + file storage + communication tools.

Contract tool: PandaDoc (strong Zapier/Make integration, template system) or Bonsai (all-in-one for solo VAs, contract, invoice, CRM). Add when you are ready to automate the contract-to-onboarding sequence.

File storage: Google Drive with a Master Client Folder Template is the standard for most VA operations. Make handles folder duplication and renaming natively.

Communication: Gmail or Outlook as the email automation layer; Slack for internal notifications. Both integrate natively with Zapier and Make.

Layer 3 — Intelligence

(Add at Advanced Level)

AI tools — for content generation, data interpretation, and personalization at scale.

Claude / ChatGPT: email drafting, client summaries, SOP generation, intake form analysis. Accessible via Make‘s HTTP module for API calls embedded directly in automation scenarios.

Notion AI: documentation generation, SOP drafting, knowledge base management.

ClickUp AI: task generation from briefs, project setup, workspace configuration.

Add AI tools to workflows only after the mechanical automation layer is stable, AI enhances a working system, it cannot substitute for one that is not yet configured.

👉 ChatGPT for Virtual Assistants: Complete Guide — for the complete guide to prompt architecture, API configuration, and the system prompt structures that produce reliable automated outputs.

The automation workflows in this guide are built around a combined Notion + ClickUp stack, Notion as the knowledge and documentation layer, ClickUp as the task execution and automation layer. If you have not yet decided which tool to use as your primary workspace, Notion vs ClickUp for Virtual Assistants covers the complete comparison and the decision framework before you build the automation layer on top.

Workflow 1 — Email-to-Task Automation

Time to build: 10-15 minutes in Zapier.

Time saved: 5-10 minutes per actionable email.

Best for: VAs managing client inboxes or operations-heavy accounts.

How it works: When an email arrives in Gmail containing specific keywords (configurable per client), Zapier creates a ClickUp task automatically with the email subject as the task name, the email body as the task description, and a due date set to today + 1 day.

Zapier configuration:

TRIGGER: Gmail — New email matching search
Search filter: subject: (invoice OR approval OR urgent OR action required)
AND from:[client email domain]

ACTION 1: ClickUp — Create task
Name: [Gmail subject]
Description: [Gmail body plain text]
List: [relevant client list]
Due date: today + 1 day
Priority: High

The one configuration decision that matters: The Gmail search filter. Too broad (every email creates a task) produces noise. Too narrow (misses relevant emails) defeats the purpose. Start with 3-4 keywords specific to the client’s communication style, then refine after one week of testing.

Workflow 2 — Auto-Send Intake Form to New Leads

Time to build: 15-20 minutes in Zapier.

Time saved: 10-15 minutes per new lead.

Best for: VAs with active lead generation or referral-based business development.

How it works: When a new lead is captured, via website contact form, CRM, or manual addition, Zapier sends the intake form automatically with a confirmation email, and creates a follow-up task in ClickUp due in 3 days.

Zapier configuration:

TRIGGER: Tally / Typeform / CRM — New lead

ACTION 1: Gmail — Send email
To: [lead email]
Subject: "Next step — tell me about your project"
Body: [intake form link + brief intro]

ACTION 2: ClickUp — Create task
Name: "Follow up — [Lead Name]"
List: Admin & Operations
Due date: today + 3 days
Priority: Medium

Workflow 3 — Auto-Organize Files in Google Drive

Time to build: 20-25 minutes in Make.

Time saved: 5-10 minutes per file uploaded to the wrong location.

Best for: VAs managing high document volume across multiple clients.

How it works: When a new file is uploaded to a designated “Inbox” folder in Google Drive, Make reads the filename, identifies the client name or file type based on naming conventions, and moves the file to the correct client subfolder automatically.

Requires: a consistent file naming convention before the automation can work, files must include the client name or project code in the filename for the router logic to function.

Make scenario:

TRIGGER: Google Drive — Watch files in folder
Folder: /Inbox/

MODULE 1: Router — Branch by filename pattern
IF filename contains [Client A name] → Move to /Clients/Client A/02 — Deliverables/
IF filename contains [Client B name] → Move to /Clients/Client B/02 — Deliverables/
IF filename contains "invoice" OR "contract" → Move to /Admin/Finance/
DEFAULT → Move to /Inbox/Review/
  AND notify VA via Slack

Workflow 4 — Automated Email and Slack Notifications

Time to build: 8-10 minutes in Zapier.

Time saved: eliminates manual monitoring of multiple tools simultaneously.

Best for: all VA types, the highest frequency manual check in any VA operation.

How it works: When a new task is created in ClickUp (by any method, automation, client, team member), Zapier sends a Slack notification with the task name, assignee, due date, and a direct link to the task. No more opening ClickUp to check for new tasks.

Zapier configuration:

TRIGGER: ClickUp — New task in list
List: [all client lists, Space-level trigger]

ACTION: Slack — Send channel message
Channel: #task-alerts
Message: "🆕 New task: [task name]
Due: [due date] | Priority: [priority]
Client: [list name] | Link: [task URL]"

Workflow 5 — Client Onboarding System

Time to build: 3-4 hours in Make.

Time saved: 2-3 hours per new client.

Best for: all VA types managing retainer or project-based clients.

This is the highest-ROI workflow in the entire virtual assistant automation system, the one that produces the most visible operational improvement per hour of setup investment.

Trigger: intake form submitted.

What the scenario does automatically: creates the Google Drive folder structure from the Master Template, generates the contract from PandaDoc template populated with form data, waits for signature confirmation, creates the ClickUp client list from the Client Template, sends the welcome email with workspace link and access form, notifies the VA via Slack, and logs the new client in the CRM.

Full implementation detail, tool configuration, and Make scenario structure: 👉 How to Automate Client Onboarding for Virtual Assistants

Workflow 6 — Content Scheduling Automation

Time to build: 45-60 minutes in Zapier.

Time saved: 3-5 hours per week for content-focused VAs.

Best for: Social media VAs, content managers, VAs managing editorial workflows for clients.

Trigger: content status changes to “Approved” in ClickUp or Airtable content calendar.

Automation sequence:

TRIGGER: ClickUp — Task status changes to "Approved"
In list: [Client Name] Content Calendar

ACTION 1: Buffer / Later — Create scheduled post
Platform: [from custom field]
Content: [from task description]
Schedule: [from due date custom field]

ACTION 2: ClickUp — Update task status
New status: "Scheduled"
Apply tag: "Ready to Publish"

ACTION 3: Gmail — Send notification to client
Subject: "[Platform] post scheduled — [Date]"
Body: post title + scheduled date + platform

ACTION 4: ClickUp — Create follow-up task
Name: "Check analytics — [post title]"
Due date: scheduled date + 7 days

The configuration that saves the most time: the “Approved” status trigger, it means the VA never manually initiates scheduling. The approval action from the client (status change or checkbox) triggers the entire sequence.

The social media automation workflow is one of the highest-return automation builds for VAs whose clients include social media management in their service scope. The complete four-layer system (content generation, scheduling, client approval, and automated reporting) with the full Make scenario sequence is in 👉 How to Automate Social Media as a Virtual Assistant.

Workflow 7 — CRM Automation

Time to build: 30-45 minutes in Zapier.

Time saved: 1-2 hours per day for VAs managing lead pipelines.

Best for: executive VAs, sales support VAs, VAs managing client acquisition.

Trigger: new lead added to CRM or lead status changes.

Automation sequence:

TRIGGER: HubSpot / Airtable / Pipedrive —
New contact created OR status changes

ACTION 1: ClickUp — Create task
Name: "Follow up — [Lead Name] — [Company]"
List: Admin & Operations > Business Development
Due date: today + 1 day (new lead)
         OR today + 3 days (existing lead — status change)

ACTION 2: Gmail — Send lead response email
Template: [appropriate template for lead source]
Personalization: first name + company from CRM

ACTION 3: CRM — Update fields
Last contacted: today
Next follow-up: due date from Action 1
Status: In Progress

ACTION 4: Slack — Notify VA
Message: "New lead: [Name] — [Company]
Source: [lead source] | Status: [CRM status]"

Workflow 8 — Client Reporting Automation

Time to build: 2-3 hours in Make.

Time saved: 4-8 hours per month across all clients.

Best for: all VA types managing clients with reporting deliverables.

Trigger: scheduled (weekly or monthly, per client reporting cadence).

Make scenario:

TRIGGER: Make Scheduler — Every Friday 4:00 PM
(or first Monday of month for monthly reports)

MODULE 1: ClickUp — Search tasks
Filter: Status = Done
Date completed = this week (or this month)
List = [Client Name]

MODULE 2: Iterator — Loop through completed tasks
Extract: task name, completion date, time tracked

MODULE 3: Aggregator — Compile task list

MODULE 4: Google Analytics / Social Platform API — Pull performance metrics
(if applicable to client's service type)

MODULE 5: Claude API — Generate insights
Prompt: "Summarize these completed tasks and metrics into a professional weekly/monthly client report.
Include: key accomplishments, metrics summary, observations, and next week/month priorities.
Tone: professional and concise. Under 300 words."

MODULE 6: Gmail — Send report
To: [client email]
Subject: "Weekly Update — [Client Name] — [Date]"
Body: AI-generated report + task list + metrics

MODULE 7: Google Sheets — Log metrics
Row: client, date, tasks completed, time tracked, key metrics

Workflow 9 — Automated Client Management System

Time to build: 6-8 hours across multiple sessions.

Time saved: 10-15 hours per month across a full client portfolio.

This is the backbone of a scalable VA business, a single connected system that manages the complete client lifecycle from onboarding to renewal without manual coordination between stages.

System architecture:

STAGE 1 — ONBOARDING (triggered by form submission)
→ Folder creation → Contract → Workspace setup → Welcome email → Access collection → Kickoff

STAGE 2 — ACTIVE DELIVERY (recurring triggers)
→ Weekly recurring tasks generate automatically
→ Weekly report sent every Friday
→ Monthly invoice on last day of month
→ Client dashboard updated in real time

STAGE 3 — RENEWAL (triggered by contract end date - 30 days)
→ ClickUp task: "Prepare renewal proposal — [Client]"
→ Gmail: send renewal discussion email to client
→ Calendly link for renewal call included

STAGE 4 — OFFBOARDING (triggered by status change)
→ ClickUp: create offboarding checklist
→ Google Drive: archive client folder
→ CRM: update status to "Past Client"
→ Gmail: send final delivery + feedback request

The critical configuration: the contract end date must be stored as a custom field in ClickUp from the onboarding stage. The renewal trigger is a Make scenario scheduled daily that checks all active client records and activates the renewal sequence when the date condition is met.

Workflow 10 — Automated Lead Qualification System

Time to build: 3-4 hours in Make + Zapier.

Time saved: 30-45 minutes per new lead.

Best for: VAs building their own client acquisition system or supporting clients with lead management.

System architecture:

TRIGGER: New lead via website form / CRM / referral

MODULE 1: Make — Score the lead
Scoring criteria (assign points):
+ Budget mentioned → +3
+ Timeline under 30 days → +2
+ Service matches VA specialty → +2
+ Referral source (vs cold) → +1
+ Complete form submission → +1

MODULE 2: Router — Branch by score
Score ≥ 7 → HIGH PRIORITY branch
Score 4-6 → MEDIUM PRIORITY branch
Score ≤ 3 → LOW PRIORITY branch

HIGH PRIORITY branch:
→ Gmail: personalized response within 1 hour
  (AI-generated from form data)
→ Calendly: include discovery call link
→ ClickUp: task due today, Priority High
→ CRM: status = Hot Lead

MEDIUM PRIORITY branch:
→ Gmail: intake form sent
→ ClickUp: task due in 2 days, Priority Medium
→ CRM: status = Warm Lead

LOW PRIORITY branch:
→ Gmail: standard information email
→ ClickUp: task due in 5 days, Priority Low
→ CRM: status = Cold Lead

Workflow 11 — Automated Content Production Pipeline

Time to build: 2-3 hours in Make + ClickUp.

Time saved: 45-60 minutes per client per week for content-focused VAs.

Best for: social media VAs, content managers, VAs managing end-to-end content production.

System architecture:

STAGE 1 — IDEATION (Monthly, first Monday)
Make Scheduler triggers:
→ Claude API: generate 4-week content plan
  Prompt: "[Client niche + tone + pillars] — generate 20 content ideas with platform, format, and hook for each"
→ ClickUp: create one task per content idea
  Custom fields: Platform, Format, Content Pillar, Publish Date, Status = "Idea"

STAGE 2 — PRODUCTION (Weekly)
→ VA selects ideas → sets status to "In Production"
→ ClickUp automation: assign subtasks (Brief, Caption, Design brief, CTA)
→ Claude API (via Make): generate caption draft from task name + content pillar field
→ Caption added to task description automatically

STAGE 3 — REVIEW (Client approval)
→ Status = "Client Review" triggers:
  Gmail: send content for approval
  ClickUp: assign to client (guest access)

STAGE 4 — PUBLICATION
→ Status = "Approved" triggers (Workflow 6):
  Buffer: schedule post
  Status update → "Scheduled"

STAGE 5 — ANALYTICS (Weekly, Monday)
→ Platform API: pull performance data
→ Google Sheets: log metrics per post
→ Monthly: Claude generates performance summary

Workflow 12 — AI Client Summary from Intake Form

When it runs: after every new client intake form submission.

What it replaces: 20-30 minutes of manual review and note-taking per new client.

Implementation via Make HTTP module:

MODULE: HTTP — POST to Claude API
URL: https://api.anthropic.com/v1/messages
Headers: Content-Type: application/json

Prompt:
"Summarize this client intake form into a structured internal brief. Output exactly:

CLIENT OVERVIEW
Name, business, service requested — 2 sentences

PRIMARY GOALS
3 bullet points maximum

CURRENT TOOLS
List all mentioned

COMMUNICATION PREFERENCES
Channel, frequency, time zone

ACCESS REQUIREMENTS
List all platforms mentioned

FLAGS
Any unusual requests, constraints, or red flags
(write NONE if none exist)

Tone: factual. No padding. No introductory sentence.

Intake form data: [ALL FORM FIELDS FROM MAKE DATA STORE]"

Output saved automatically to ClickUp Doc “Client Brief — [Client Name]” in the client’s workspace.

Workflow 13 — AI-Drafted Client Emails

When it runs: triggered by specific ClickUp task status changes or on a weekly schedule.

What it replaces: 10-15 minutes of writing per client communication email.

The three highest-value email types to automate:

Weekly update email (Friday trigger):

"Write a professional weekly update email for [Client Name]. Use the task list below as the source of accomplishments.

Format:
- Subject line (include client name and week dates)
- Opening: 1 sentence referencing their business goal
- Accomplished this week: bullet list from tasks (max 5)
- In progress: bullet list (max 3)
- Coming up next week: bullet list (max 3)
- Closing: 1 sentence + next touchpoint reference

Tone: professional and direct. Under 200 words total.
No filler phrases.

Completed tasks this week: [TASK LIST FROM CLICKUP]"

Post-meeting follow-up (Calendly trigger):

"Write a post-meeting follow-up email based on these meeting notes. Include: meeting summary (2-3 sentences), agreed action items (numbered list), owner and deadline for each item, next meeting confirmation if applicable.

Tone: professional and clear. Under 150 words.

Meeting notes: [PASTE OR PASS FROM MEETING NOTES TASK]"

Workflow 14 — AI-Enhanced Reporting

When it runs: weekly or monthly, on schedule.

What it replaces: 30-60 minutes of analysis and report writing per client per cycle.

This is the AI workflow with the highest perceived value differential, clients receive narrative insights instead of raw numbers, which elevates the VA’s service from data delivery to strategic interpretation.

Make scenario with Claude integration:

MODULE 1-3: Pull metrics from connected platforms (same as Workflow 8, Module 4)

MODULE 4: HTTP — Claude API
Prompt:
"Analyze these performance metrics and write a professional [weekly/monthly] report section.

Include:
- 2-sentence executive summary
- What worked this period (specific, data-referenced)
- What underperformed and why (honest, not defensive)
- One strategic recommendation for next period
- Key metric callouts (bold the numbers)

Tone: professional analyst. Not promotional.
Use specific numbers. Under 250 words.

Metrics: [ALL METRIC DATA FROM MODULES 1-3]
Previous period metrics: [IF AVAILABLE]
Client context: [CLIENT NICHE + GOALS]"

Zapier Examples (Beginner-Intermediate)

Example 1 — Lead Capture System Form submission → Gmail confirmation + intake form → ClickUp task “Follow up — [Lead Name]” due today + 1 → Slack alert to VA → CRM record created. Build time: 20 min. Tools: Tally + Zapier + Gmail + ClickUp + Slack.

Example 2 — Email-to-Task Converter Gmail incoming email matching keywords → ClickUp task created with email subject + body → Priority set to High → Due date today + 1. Build time: 10 min. Tools: Gmail + Zapier + ClickUp.

Example 3 — Content Approval to Scheduling ClickUp status changes to “Approved” → Buffer schedules post → Gmail notifies client → ClickUp analytics follow-up task created for +7 days. Build time: 25 min. Tools: ClickUp + Zapier + Buffer + Gmail.

Zapier Workflow Example:
Form submission → Email notification → CRM update → Task creation

Make Examples (Intermediate-Advanced)

Example 4 — Full Client Onboarding Pipeline Typeform submission → Google Drive folder from template → PandaDoc contract generated + sent → Wait for signature → ClickUp list from template → Welcome email → Access form → CRM updated → Slack notification to VA. Build time: 3-4 hours. Tools: Typeform + Make + Drive + PandaDoc + ClickUp + Gmail.

Example 5 — Weekly Reporting Pipeline Make scheduler Friday 4PM → ClickUp done tasks pulled → Iterator aggregates task list → Claude generates insights → Gmail sends formatted report → Google Sheets logs metrics. Build time: 2-3 hours. Tools: Make + ClickUp + Claude API + Gmail + Sheets.

Example 6 — Lead Scoring and Routing New CRM lead → Make scores based on criteria → Router branches High / Medium / Low → Gmail sends appropriate response → ClickUp task priority set → CRM status updated. Build time: 2-3 hours. Tools: CRM + Make + Gmail + ClickUp.

Make Scenario Example:
Onboarding automation with contract, folder creation, ClickUp setup, and welcome email.

ClickUp Example (Native Automations)

Example 7 — Recurring Client Delivery System Recurring task activates weekly → ClickUp automation generates subtasks with due dates → Status change to “In Progress” triggers Slack notification → Status change to “Done” triggers Gmail client update → Archive after 7 days. Build time: 30-45 min in ClickUp native automations. Tools: ClickUp (native) + Slack + Gmail via Zapier.

ClickUp Recurring Delivery System:
Automate weekly or monthly client deliverables with recurring tasks and automated status updates.

The Four-Week Implementation Sequence

The four-week sequence below is the implementation path for the best automation workflows for virtual assistants, one layer at a time.

Week 1 — Foundation (Beginner workflows): Build all four beginner workflows in order: email-to-task, intake form automation, file organization, Slack notifications. Test each with real data before moving to the next. At the end of week 1: 4 automations running, approximately 3-4 hours per week saved.

Week 2 — Client Operations (Intermediate workflows): Choose the intermediate workflow with the highest time cost in your current operation. For most VAs: the reporting automation (Workflow 8) or the content scheduling automation (Workflow 6). Build one this week, test thoroughly, activate. At the end of week 2: 5-6 automations running.

Week 3 — Onboarding System: Build the client onboarding workflow (Workflow 5) as a standalone Make scenario. Test with a real new client or a complete simulation. This is the highest-impact single workflow in the best automation workflows for virtual assistants — it deserves its own dedicated week.

Week 4 — AI Layer: Add the AI client summary (Workflow 12) to the onboarding scenario. Add the AI weekly email draft (Workflow 13) to the reporting or communication workflow. Test the API calls. Verify outputs meet quality standard before removing manual writing from the cycle.

Weeks 5+ — Advanced systems: Build the advanced workflows (9, 10, 11) one at a time, with 2-3 weeks between each to allow for stabilization and edge case identification.

The One Rule That Prevents System Failure

Build, test, stabilize, then expand. Every automation failure that causes a client experience problem traces back to the same root cause: a new automation was activated before the previous layer was tested with real data in real conditions. The four-week sequence prevents this by building one layer at a time with a stabilization period between each.

A virtual assistant automation system built over four weeks and stabilized over eight is more reliable than one built over two weeks without testing. The speed of implementation is not the constraint, the thoroughness of testing is.

Mistake 1 — Automating an Undefined Process

Automation executes a process exactly as configured, including every inconsistency, exception, and gap in the underlying process. A VA who automates an onboarding sequence that was previously executed differently for each client produces an automated system that is inconsistently configured and requires constant manual correction.

The consequence: the automation runs but produces unreliable outputs, the wrong folder structure for some clients, missing tasks for others, welcome emails that reference the wrong service type.

The fix: map every step of the process in writing before opening any automation tool. Define the trigger, the action for every step, the expected output, and the handling for exceptions (what happens if the form is incomplete, if the contract is not signed within 48 hours, if the payment fails). Automation is the execution layer, the process design must precede it.

Mistake 2 — Building Without Testing

An automation tested only in theory will fail at the worst possible moment, when a real client is waiting for a contract that did not generate or a welcome email that did not arrive.

The consequence: client experience damage that cannot be undone by explaining that “the automation failed.”

The fix: before any workflow goes live, run a complete end-to-end test with your own email address as the test client. Submit the form. Follow every action through to completion. Verify every output: the folder was created with the correct structure, the email arrived with the correct content, the ClickUp task was generated with the correct due date and custom fields. Fix every failure before the first real client uses the system.

Mistake 3 — Over-Automating Communication

The highest risk in client-facing automation is mechanizing touchpoints that carry relationship weight. An automated welcome email that arrives in 90 seconds, followed by an access form 30 minutes later, followed by a reminder 48 hours after that, creates a conveyor-belt experience that signals the absence of human attention, exactly the opposite of what premium VA services are supposed to deliver.

The fix: audit every automated communication for relationship weight. Automate transactional communications (confirmations, form links, reminders, status updates). Keep manual the touchpoints that carry personal significance: the post-kickoff check-in, the feedback request, the renewal conversation, the message when something goes wrong.

Mistake 4 — Ignoring Edge Cases

Most automation failures in production are not bugs, they are edge cases that were not considered during build. A field left blank on the intake form. A client who signs the contract but does not pay within the expected window. A file uploaded with a naming convention that does not match the router filter.

The fix: for every automation, ask “what happens if…” for each input the automation depends on. Missing field: does the automation fail silently, fail visibly, or route to a fallback? Unexpected value: does the router have a default branch? Delayed trigger: does the automation re-trigger correctly or create a duplicate? Map these scenarios before build and configure handling for each one.

Mistake 5 — Not Documenting the System

An automation system that lives only in the automation platform, with no written documentation of what triggers what, what each workflow does, and where the manual steps are, cannot be maintained, audited, or handed to anyone else. When something breaks (and it will), diagnosing the failure requires reconstructing the logic from the platform interface instead of consulting a document.

The fix: maintain a single ClickUp Doc or Notion page titled “Automation System Documentation” with: a plain-language description of every active automation, the trigger, the tools involved, the expected output, the manual steps, and the last test date. Update it every time the system changes. 30 minutes of documentation per workflow prevents hours of diagnosis when something breaks.

Mistake 6 — Failing to Maintain Workflows

Automation is not “set it and forget it.” Tools update APIs, apps change authentication requirements, field names shift between platform versions, and connected services deprecate integrations. A workflow that ran correctly in February may fail silently in June because a connected app updated its API without notice.

The fix: schedule a monthly 30-minute automation audit, open each active scenario in Make or Zap in Zapier, check the execution log for recent failures, verify every connected app is still authenticated, and test any workflow that has not run recently with a manual trigger. Quarterly: review whether the workflow still reflects the current version of your process or whether the underlying process has changed and the automation has not kept up.