Hey, Vlad here. I believe AI should make your business better without making your life more complicated. Let me be honest – the word “AI” can sound intimidating, like something only tech companies need. But think of it this way: AI in Exoserva is like having an incredibly fast, tireless assistant who handles the repetitive stuff so you can focus on the work that actually requires a human. It dispatches technicians, responds to customers, predicts equipment failures, and generates reports – all automatically. The AI Settings Hub is your control center for ALL of that. It is one single page where you can see exactly what AI is doing for your business, how much it costs, how much it is SAVING you, and you can turn any feature on or off with a single click. Every dollar is tracked, every feature has a toggle, and you are always in complete control. This guide walks you through the entire Hub from top to bottom.
Estimated time: 15 minutes
Before You Begin
- An active Exoserva account with Owner or Admin role – AI settings affect your entire business operation, so only Owners and Admins can change them. If you are not sure what role you have, click your profile icon in the top-right corner and look under your name
- A paid plan that includes AI features (Starter, Pro, or Enterprise) – different plans include different AI features. The Starter plan includes basic AI features, Pro includes advanced ones, and Enterprise includes everything. You can see which plan you are on in Settings > Billing
- At least a few days of operational data in your account (jobs, customers, messages) for meaningful statistics – AI features learn from your data, so the more data you have, the smarter they get. If you just signed up yesterday, the AI will work but the usage stats and ROI numbers will not be very meaningful yet
- A willingness to experiment – I recommend turning on a few features at a time, watching how they perform for a week, and then adjusting. AI is not an all-or-nothing decision
Step 1: Navigate to the AI Settings Hub
Look at the left sidebar menu and click on “Settings” (the gear icon). Inside Settings, you will see several option cards. Find and click the “AI Configuration” link (you can also navigate directly by typing /settings/ai in your browser address bar). The AI Settings Hub page opens.
At the top of the page, you will see a breadcrumb trail showing Settings > AI Configuration – breadcrumbs are those small text links that show you where you are in the app, like a trail of breadcrumbs in a forest. Below that, the page is organized into three main sections, accessible through tabs: Features (sparkles icon – where you turn individual AI features on and off), Settings (gear icon – where you set budgets and global preferences), and Usage (bar-chart icon – where you see how much AI is being used and what it is saving you).
There is also a clever feature for when you are scrolling down this page: when the hero section (the big overview area at the top) scrolls off screen, a compact sticky header pins itself to the top of your browser window. Think of it like a toolbar that follows you as you scroll. This sticky header shows the essential controls in a condensed form: a brain icon (green when AI is enabled, red when disabled), the page title, an enabled/disabled status badge, how many features are active (like “12/18 active”), your month-to-date cost, and the master AI toggle switch. This means you can always see the AI status and turn it on/off without scrolling all the way back to the top.
There is also a skip link at the very top of the page that is visible only when you use keyboard navigation (Tab key) – pressing it jumps you directly to the AI features list. This is an accessibility feature for users who navigate with keyboards instead of mice.
Tip: Use the keyboard shortcut Cmd+K (on Mac) or Ctrl+K (on Windows) to focus the feature search bar when you are on the Features tab. This is the fastest way to find and toggle a specific AI feature without scrolling through the entire list.
Step 2: Understand the Quick Stats Hero Section
The top of the page features a large section called the QuickStatsHero – “hero” is web design terminology for a big, prominent section at the top of a page. Think of it as the dashboard of your car – it gives you an at-a-glance overview of everything important about your AI deployment.
The most prominent element is the master AI switch – a large, clearly visible toggle that globally enables or disables ALL AI features across your entire Exoserva account. Next to it, you will see the current AI status displayed as a badge: a green “Enabled” badge when AI is on, or a red “Disabled” badge when AI is off.
Below the master switch, four quick stat metrics are displayed in a row:
- MTD Cost – “MTD” stands for “Month-To-Date,” which means “how much you have spent so far this month.” This shows your total AI spending in dollars for the current calendar month. Think of it like checking your phone bill midway through the month
- Active Features – shows how many AI features you currently have turned on out of the total available. For example, “12/18” means 12 features are active out of 18 total. You do not have to use all of them
- Success Rate – the percentage of AI operations that completed successfully. A healthy success rate is above 95%. If this number drops, it might mean a feature needs attention
- Avg Latency – how fast the AI responds, measured in milliseconds. Lower is better. Under 500ms is excellent. “Latency” just means “how long it takes” – like how long you wait after pressing a button before something happens
When AI is enabled, an additional AI Savings Summary section appears below showing three ROI (Return on Investment) metrics: Time Saved (estimated hours the AI saved you by handling tasks automatically), Labor Cost Saved (the dollar value of those saved hours – calculated using the hourly labor rate configured in your settings), and ROI (your return on investment as a multiplier, like “4.2x” meaning for every $1 you spend on AI, you get $4.20 back in saved labor). If some features are not yet enabled, a Quick Setup bar appears with an “Enable Popular” button – we will cover that in Step 4.
From Vlad: Here is why I put the ROI front and center on this page: I want you to see, in real numbers, that AI is not a cost – it is an investment that pays for itself many times over. Every dollar you spend on AI should generate multiple dollars in labor savings. If the ROI ever drops below 1x (meaning you are spending more on AI than it is saving you), something needs adjusting – and the Hub gives you all the controls to fix it. But honestly, I have never seen a user with an ROI below 2x after the first month. The AI features pay for themselves quickly.
Step 3: Enable AI with the Master Switch
The master AI switch in the QuickStatsHero is the most important toggle on this entire page. Think of it like the main circuit breaker in your house – when it is ON, electricity flows to all the rooms. When it is OFF, everything goes dark. The master switch controls whether ANY AI functionality is active in your Exoserva account.
When the master switch is ON (the default for paid plans): all individually enabled AI features are active and running. The brain icon in the hero section and the sticky header turns green. The status badge says “Enabled.” AI dispatch suggestions appear on your schedule board, AI auto-respond sends immediate replies to customer messages, predictive maintenance alerts pop up on your dashboard, and all other enabled features work as expected.
When the master switch is OFF: EVERY AI feature is suspended immediately. It does not matter if you had 15 features individually toggled on – the master switch overrides them all. The brain icon turns red, the status badge says “Disabled,” and no AI processing occurs anywhere in your account. No AI dispatch, no auto-respond, no predictive analytics, nothing. Think of it as a “kill switch” for all AI activity.
When you toggle the master switch, it sends an immediate API call to update your settings on our servers. While the update is processing (usually less than a second), the switch is briefly disabled to prevent you from clicking it twice rapidly. Once the update completes, a green toast notification confirms “Settings updated.” An important detail: when you turn AI back ON, all the features that were individually enabled before you turned it off will be restored to their previous state. The master switch does not reset your individual feature preferences.
Warning: Turning OFF the master switch stops ALL AI features IMMEDIATELY. This includes active automations, scheduled AI dispatch, and AI responses that are currently being generated. Only use this global switch if you have a specific reason to disable everything at once. For turning off a single feature, use the individual feature toggles instead (covered in Step 5). Think of the master switch as an emergency stop button – powerful but broad.
Step 4: Quick-Start with Popular Features
If you are new to AI features and feeling overwhelmed by the number of options, this step is for you. I designed the Quick Setup feature specifically for users who want to get started fast without having to research every individual feature. Think of it like a “recommended settings” option in a video game – it picks the best options for beginners so you can start playing immediately and customize later.
The Quick Setup bar appears at the top of the Features section whenever you have AI features available to enable. It shows a count of how many features are still available (e.g., “6 features available to enable”). The star of the show is the “Enable Popular” button with a star icon. Click it, and Exoserva automatically activates the five most commonly used AI features across ALL Exoserva customers. These are the features that have been proven to deliver the most value for field service businesses, and they typically include AI dispatch suggestions (helping you assign the right technician to each job), automated customer responses (responding to inquiries within seconds), smart scheduling (optimizing your calendar for efficiency), and predictive maintenance alerts (warning you before equipment fails).
When you click the button, each feature activates one by one, and you will see a green success toast notification for each one. The feature cards in the grid below update their toggle state in real time – you can watch them switch from “off” to “on” as each feature activates. The active feature count in the hero stats section at the top also increments accordingly. Once all features are enabled, the Quick Setup bar disappears because there are no more features to recommend.
An important safety note: the “Enable Popular” action only enables features that are available on your current plan tier. If a feature requires a Pro or Enterprise plan and you are on a Starter plan, it will NOT be activated. Features that require a higher tier are shown with tier badges (like “Pro” or “Enterprise”) on their cards, so you can see what is available to upgrade to.
Step 5: Explore and Toggle Individual Features
Now let us dive into the meat of the AI Settings Hub – the Features tab (selected by default when you open the page). This is where you see every single AI feature available in Exoserva, organized neatly by category. Think of it like a menu at a restaurant – everything is grouped by type (appetizers, entrees, desserts), and you can order whatever you want.
At the top of the Features tab, you will find a search bar (magnifying glass icon) with a hint showing the Cmd+K shortcut. Type any word to instantly filter features by name or description – for example, typing “dispatch” will show only features related to dispatching technicians. Next to the search bar, category filter buttons let you view features from one specific category at a time. Seven categories are available:
- Dispatch – features that help assign the right technician to each job
- Communication – features that handle customer messaging and responses
- Financial – features related to invoicing, payments, and cost estimation
- Routing – features that optimize driving routes for your technicians
- Predictive – features that predict future events (equipment failures, demand spikes)
- Analytics – features that analyze your business data and generate insights
- General – features that do not fit neatly into the other categories
Each filter button shows the category icon, an abbreviated name, and a small badge showing how many features in that category are enabled versus the total (like “3/5”). The badge is green when ALL features in the category are enabled, amber/yellow when some are enabled, and grey when none are enabled. Click “All” to clear any active filter and see every feature.
Features are displayed in a two-column grid (which becomes a single column on smaller screens like phones or tablets), grouped under CategoryHeader sections. Each category header shows the category name, its icon, the enabled count, and bulk toggle buttons: “Enable All” (turns on every feature in the category with one click), “Disable All” (turns them all off), or “Enable Remaining” (if some are already on, this enables the rest). These bulk toggles are massive time savers when setting up or adjusting an entire category.
Each individual FeatureCard in the grid displays a wealth of information: the feature name, a description of what it does, a tier badge (Free, Starter, Pro, or Enterprise – shows which plan is required, plus a “beta” badge if the feature is still in testing), a toggle switch (the on/off control), impact level (High, Medium, or Low – how much this feature affects your operations), estimated time saved per week, ROI estimate, usage count (how many times it has been used), error count (how many times it failed), and last-used timestamp. To configure a feature in detail, click the settings button (gear icon) on its card to open the configuration modal (covered in Step 8).
Step 6: Configure Global AI Settings
Click the Settings tab (the second tab, with a gear icon) to open the GlobalSettingsPanel. If individual feature toggles are like light switches in each room of your house, global settings are like the thermostat and circuit breaker panel – they control the overall environment and set limits that apply everywhere.
At the top of the panel, you will see security compliance badges – these are small icons that confirm your data is protected: Encrypted (all AI data is encrypted in transit and at rest – meaning no one can read your data as it travels over the internet or while it is stored on our servers), SOC2 (our infrastructure meets SOC 2 compliance standards – a rigorous security audit), GDPR (we follow GDPR data handling rules – important for European customers), and 99.9% Uptime (our AI services are available 99.9% of the time – that is less than 9 hours of downtime per year). These badges are informational – they cannot be changed, they are just here to give you peace of mind about security.
Below the badges, you will find the configurable settings. Let me explain each one:
Model Preference – This lets you choose which AI model powers your features. Think of AI models like car engines – some are powerful but expensive, others are economical but slightly less capable. The options typically range from faster/cheaper models to more accurate/premium models. For most businesses, the default model is the best balance of quality, speed, and cost.
Monthly Budget – This is a spending cap for your AI usage. Think of it like a spending limit on a credit card. When your AI spending for the month reaches this amount, features will automatically pause until the next month. This prevents unexpected charges. For example, if you set a $100 budget and your AI spends $100 by the 20th of the month, all AI features will pause for the remaining 10 days.
Rate Limits – This controls how many AI requests can be made per minute. Think of it like a speed limit – it prevents any single feature from making too many requests too quickly, which could run up costs unexpectedly. The default rate limit is appropriate for most businesses.
Data Retention – How long your AI interaction logs are stored. Longer retention means more data for analytics and insights, but takes up more storage. Shorter retention means less historical data but leaner storage.
Notification Email – Where budget alerts and AI error notifications are sent. Make sure this is an email you check regularly so you do not miss important alerts like “you are at 90% of your monthly budget.”
After making any changes, click “Save Settings” at the bottom of the panel. A green toast notification will confirm “Settings updated.” These global settings serve as guardrails – they ensure AI usage stays within your budget and performance expectations without you having to monitor it constantly.
Step 7: Monitor Your AI Usage and ROI
Click the Usage tab (the third tab, with a bar-chart icon) to open the UsageStatsCard dashboard. This is where you become a data-driven business owner. Think of this tab as your AI report card – it shows you exactly how AI is performing, what it is costing, and what it is saving you. Every number here comes from real data, not estimates.
At the top, three ROI summary cards show your AI investment returns in big, easy-to-read numbers. These tell you at a glance whether AI is paying for itself (hint: it almost always is). Below the summary cards, you will see four sparkline charts – these are small, simple graphs that show trends over the past seven days. Each chart has a color-coded line: green means the trend is positive (good), red means the trend is concerning (needs attention).
The four sparkline charts show:
- Requests – the total number of AI operations performed each day. This tells you how actively AI is working in your business. A healthy upward trend means your features are being used
- Success Rate – the percentage of AI operations that completed successfully each day. You want this line to stay high (above 95%). If it dips, something might need attention
- Latency – how fast the AI responded each day, measured in milliseconds. Lower is better. You want this line to stay flat or trend downward. A spike upward could mean the AI is overloaded or a feature is misconfigured
- Cost – how much you spent on AI each day in dollars. This helps you understand your spending pattern. Some variation is normal (busy days cost more than quiet days)
Below the sparklines, a Feature Breakdown section lists the top five features by usage. For each feature, you can see its name, how many times it was invoked (used), its success rate, and how much it contributed to your total cost. This breakdown is incredibly valuable because it shows you WHICH features are delivering the most value and WHICH might need attention. If a feature has a low success rate or a high cost relative to its usage, it might be worth adjusting its settings or reviewing its configuration in the Features tab.
Step 8: Fine-Tune Individual Feature Settings
Once you have been running AI features for a few days or a week, you might want to fine-tune how specific features behave. From the Features tab, find the feature you want to adjust and click the settings button (small gear icon) on its card. This opens the FeatureSettingsModal – a popup window with configuration options specific to that feature.
Each feature has different settings because each feature does different things. However, some common settings you might see include:
- Confidence Threshold – how confident the AI needs to be before taking action. Think of it like a student needing to score at least 80% on a test to pass. A higher threshold means the AI only acts when it is very sure, which reduces errors but might mean some actions do not get taken. A lower threshold means the AI acts more often but might make more mistakes. For most features, the default threshold is a good starting point
- Response Templates – pre-written text that the AI uses as a starting point for customer communications. You can customize these to match your business tone and style
- Scheduling Windows – when the AI is allowed to perform certain actions. For example, you might want the AI to only send customer texts between 8 AM and 8 PM
- Notification Preferences – who gets notified when the AI takes an action, and how (email, in-app notification, etc.)
The form fields in the modal are generated dynamically from the feature’s settings schema – this is a technical detail, but what it means for you is that the settings you see are always current and complete. As we add new configuration options to a feature on our end, they automatically appear in the modal without you needing to update anything.
After adjusting the settings, click “Save” at the bottom of the modal. It closes, a green toast notification confirms “Feature settings updated,” and the feature card on the grid reflects your changes. If you notice a feature has errors (shown as an error count on the feature card), reviewing and adjusting its settings here is the first thing to try. Most errors are caused by misconfigured parameters (like a confidence threshold set too low, causing the AI to try to act on ambiguous data) rather than actual system problems.
Common Mistakes to Avoid
- Using the master AI switch to turn off a single feature – this is like turning off the main circuit breaker in your house because you want to turn off one light. The master switch kills ALL AI functionality. To control individual features, use their individual toggle switches on the Features tab.
- Setting the monthly budget cap without checking your current usage first – too low a cap causes features to pause unexpectedly in the middle of your busiest week when you need AI the most. Always check the Usage tab first, understand your typical spending, and set the budget 20-30% above your average.
- Skipping the “Enable Popular” quick start button and instead trying to manually evaluate every single feature on day one – this analysis paralysis can take an hour, and you end up enabling nothing. Click “Enable Popular,” get value immediately, and fine-tune later.
- Not checking the Usage tab after enabling new features – a misconfigured feature can burn through your budget surprisingly fast if its confidence threshold is set too low, causing it to trigger on almost every event. Check Usage after a day or two of running new features.
- Ignoring low success rates on the Feature Breakdown – a feature with a success rate below 80% is not working well and might be wasting money on failed attempts. Review its settings in the FeatureSettingsModal and adjust the confidence threshold or check its configuration.
- Changing multiple global settings at the same time and not knowing which one caused an issue – if you change the AI model, the budget, AND the rate limits all at once, and something goes wrong, you will not know what caused it. Change one setting at a time, save, wait a day, and then change the next one.
What’s Next?
Now that you’ve completed this guide, check out:
Need help? Post in the Tech Support category or contact support@exoserva.com.