ServiceTitan AI: What Titan Intelligence Does and Who It Suits

Titan Intelligence is ServiceTitan's name for its AI: Atlas, a chat assistant; a Virtual Agent that answers overflow and after-hours calls; and Pro features for dispatch, pricing, review replies and invoices. For an HVAC or plumbing firm it suits about ten or more technicians with a dispatcher; below that, tools with published prices usually do.

ServiceTitan's pages describe the AI at length but publish no prices. Its pricing page is per technician, in three packages, each with a Request Pricing button. So the size rule here is my judgement, built from what the pages say about requirements and set-up. ServiceTitan itself says only that its pricing suits "any size".

Follow me on Instagram@sagnikteaches

Details were read from ServiceTitan's pages in October 2026. The AI Virtual Agent's parent product is labelled early access and some Atlas features are marked coming soon, so confirm current status on a demo before you plan around any of them. Where a page says nothing, the text says so.

Connect on LinkedInSagnik Bhattacharya

What does "Titan Intelligence" cover, item by item?

ServiceTitan describes Titan Intelligence as its AI for the trades, which analyses ServiceTitan data and other data to surface insights, recommendations and strategies. The Pro products are described as powered by it. Here is what the vendor's own pages list.

Subscribe on YouTube@sagnikteaches

Two things stand out in that list. Most of it is drafting and advice (replies, emails, summaries, benchmarks), which any assistant can do at a basic level; the parts that need your own job, call and dispatch data are Atlas, Dispatch Pro, Second Chance Leads and the virtual agent. And the page makes no statement about human approval of AI output, saying only that ServiceTitan is "fully transparent about your data's usage". Ask in the demo what an AI-drafted invoice email does before a person has seen it. One note on names: this Atlas is ServiceTitan's assistant and has nothing to do with the OpenAI browser of the same name, which stopped working in August 2026.

What does the AI Virtual Agent do, and how is it billed?

According to the Contact Center Pro page and ServiceTitan's help article, the agent answers overflow and after-hours calls, books jobs "based on your real capacity", reschedules and confirms appointments, and answers customers by call and text. It escalates to a live agent when the caller asks for one, when it detects frustration, or when the caller asks for something it does not support, and predefined keywords can route a call to a number you choose. If the customer speaks Spanish it switches automatically; Spanish is the only language the article names.

The requirements are the part a small firm should read twice. The help article says the agent needs an active Basic Phone integration or Phones Pro. It names the Essentials and The Works packages as the target, so a firm on Starter should ask whether it is eligible at all. Payment collection needs ST Payments switched on. Adaptive Capacity is optional but "highly recommended" for booking jobs. Setup covers a persona, skills, job types and dispatch fees, and escalation rules, so someone has to own that configuration.

Billing is where owners get caught. The help article says you are "billed based on usage" and gives no per-call rate. It adds that all calls, "regardless of their outcome or the caller's intent", are included in your bill. A wrong number or a sales call to your forwarding line costs the same as a booked boiler replacement. Also at launch the agent cannot add call notes, summaries or transcripts to the booked job record, and the forwarding number "should never be shared with customers" because direct calls will not be logged.

Is it worth it? A break-even for a 12-technician plumbing firm

The working example is a made-up firm with 12 technicians, 2 dispatchers, 3 customer service reps and 1,200 inbound calls a month. Of those, 240 arrive after hours or while every rep is busy. That is the pool the virtual agent covers. Every figure from here is an assumption you should replace with your own.

The last column divides gross profit by 240 calls, because every call is billed. If ServiceTitan quotes you a usage rate under about $20 a call and you believe a 15% booking share, the agent pays for itself on these assumptions. At $30 a call it does not, unless your jobs are worth more than $380. You cannot finish this sum without the usage rate from the quote and the booking share from your own phone log, so collect those first.

For the baseline, log every call that reaches voicemail or rings out for two weeks, and mark which ones became jobs when someone called back. That gives you the pool size and a floor on the booking share. The HVAC receptionist tutorial shows what to include in the routing rules, and this comparison of built-in and standalone receptionists sets the same question against tools with published prices.

An assistant such as ChatGPT or Claude can do the counting once the log exists. Export the two weeks from your phone system as a spreadsheet, cut caller numbers to the last four digits, and paste it in with an instruction like this:

Below is a two-week export of calls that reached voicemail, rang out or
arrived outside our opening hours ([opening hours]). Columns: date, time,
caller (last four digits), reason noted on the callback, callback outcome
(booked job, no job, no answer).
1. Count all calls, then count unique enquiries, treating repeat calls from
   the same caller within 24 hours as one enquiry.
2. Split the unique enquiries into out-of-hours and in-hours overflow.
3. Give the share of unique enquiries that became booked jobs, per group
   and overall.
4. List rows with a blank outcome instead of guessing them.

A reply for the 12-technician firm might read: 131 calls from 104 unique enquiries, 61 out of hours and 43 in-hours overflow, 19 booked jobs for an 18.3% booking share, and 7 rows with no outcome. Check two things before you use it. Tick the 19 bookings against your job list, because a caller who rang twice can be counted as booked twice. And chase the 7 blanks rather than reading them as "no job": if four of them booked, the share becomes 22.1%, above the top row of the break-even table. For the cost side use the raw call count, since the agent bills every call: 131 calls a fortnight is roughly 260 a month, close to the 240 the example assumed.

How do ServiceTitan's headline numbers read after the footnotes?

The vendor pages carry several figures, and the footnotes change how much weight they bear.

• "2x increase in capacity per dispatcher" for Dispatch Pro is footnoted as "illustrative based on average time savings from automated dispatching".
• "37% of un-converted calls recaptured as bookings" for Second Chance Leads carries "results are indicative and your results may differ materially".
• "75% reduction in accident risk" for Fleet Pro is attributed to research by two named companies, with sources on request.
• The headline results on its AI page (more revenue per technician, more service agreement renewals) are footnoted as based on data from commercial customers who had been on ServiceTitan for two or more years, with results "not guaranteed".
• One plumbing customer's results for the virtual agent appear on the Contact Center Pro page with no footnote: 60% fewer missed calls, with bookings up 11% in one place and booking rate up 17% in another.
• The Pro products page gives Contact Center Pro a "10% increase in raw booking rate", footnoted as the average from early-access customers whose "results may differ materially".
• The AI Virtual Agent's own page shows a 70% overall booking rate (90% "capacity adjusted"), footnoted "Pilot program represents early results".

None of that is dishonest, and none of it is evidence for a ten-technician residential firm. The commercial-customer footnote in particular matters if you do residential repairs. Turn each claim into a question for your own data: how many calls did we miss last month, how many dispatcher hours go on re-sorting the board, how many unbooked calls could have been recovered?

Fit by size: under five, five to nine, and ten-plus technicians

ServiceTitan says the software "allows even small businesses to scale field teams efficiently" and that its app can be learned after a day or two of training. It states no minimum. My rule rests on what its own documents require: a dispatcher whose time the dispatch claims are measured against, two packages above Starter for the virtual agent, a phone integration, usage billing, and a quote-only price that you can only compare after a demo.

Three made-up firms make the bands concrete. A four-technician heating firm where the owner books jobs from the van has no dispatcher whose capacity could double, so Dispatch Pro's headline claim has nothing to act on. A nine-technician plumbing firm with one dispatcher is the tipping case: published-price tools still fit, but the dispatcher's board is where ServiceTitan's tooling would show a difference, so ask for a trial on that. A 14-technician HVAC firm with two dispatchers and a CSR team has the call volume, the dispatch complexity and the headcount to justify the set-up. For the matchups, the Simpro versus ServiceTitan tutorial and the FieldPulse comparison go through the same decision in detail.

What an HVAC firm and a plumbing firm each get from it

The same feature list lands differently in the two trades. An HVAC firm lives on maintenance agreements, seasonal surges and equipment with model numbers. Atlas's field features suit that work: equipment information from manuals, spec sheets and installation guides, and troubleshooting help, both marked as available with Field Pro. ServiceTitan's AI page also leads with a 21% rise in service agreement renewals, footnoted to commercial customers of two or more years. The office half of Atlas, which the page says will surface opportunities managers overlook and give CSRs real-time answers, is marked "coming soon", so an HVAC owner hoping for help with renewals should check its status on the demo.

A plumbing firm's pressure point is the phone. Burst pipes and blocked drains arrive at night and at weekends, and the caller wants a time, not a callback. That makes the virtual agent the feature to test first, with two cautions from the help article: it bills every call, including the wrong numbers a busy out-of-hours line attracts, and it cannot yet write call notes onto the booked job, so the plumber arriving in the morning may know less than the caller told the agent. Price Insights, which compares your average price per service with regional averages, is the other feature a plumbing owner will notice, because a flat-rate price book is easy to leave unrevised for years. Treat its comparison as one input to your prices, not as a price list.

Firms doing both trades should pick the feature for the side that brings in more of the revenue and judge the package on that, rather than paying for both halves of the list at once.

Eight questions to ask before the ServiceTitan demo

1. What is the per-technician price for each package, with the AI and Pro products listed separately?
2. Which package do I need to run the AI Virtual Agent, and does my phone system qualify?
3. What is the per-call usage rate, and what is the cap?
4. What is the one-off implementation fee, and how long until the first booked job?
5. What is the contract term, and what does a technician cut cost me at renewal?
6. Which Atlas features are live today, and which are coming soon?
7. Is there any human approval before an AI-drafted email, reply or booking goes out?
8. If I leave, in what format do I get my customer, job and call data?

Write the answers in a table next to the quotes from two published-price tools. If the demo cannot give a number for questions 1 and 3, you cannot run the break-even above, and that is information too. For the wider budgeting problem of quote-only pricing, the Simpro pricing tutorial works through the same questions.

A demo script for Atlas: six questions to run on your own jobs

Atlas answers "using the data already in ServiceTitan", so test it on jobs you know. Ask the presenter to run these against a customer record from your own data, or a copy of it, and compare each answer with what you already know.

1. "What equipment is installed at this address, and when was it last serviced?" Check against your own records.
2. "What does the manual say about this error code on this unit?" Check that the answer names the manual for that model, not a generic one.
3. "What did we quote this customer last time, and did they accept?" Check the figure and the date.
4. "Which service agreements expire next month?" This is an office question, which the page lists as coming soon; see whether it is live.
5. "Draft a payment reminder for this unpaid invoice." Check the amount, the due date and the tone, and that it promises nothing you have not agreed.
6. "What should I check if this unit short-cycles?" Check whether the answer comes from the unit's documentation or from general knowledge.

Pass marks: every answer names its source, matches your record, and says so when it cannot answer. An assistant that answers question 2 confidently from general knowledge is a risk for a technician standing on a roof with the wrong torque value. Because Atlas draws on the data already in ServiceTitan, its answers are only as good as your equipment and job history. A firm with years of untidy records should expect untidy answers, whichever AI it buys.

A 90-day order of work if you do sign

Sequence matters more than speed, because each AI feature depends on something underneath it.

• Weeks 1 to 4: the core. Dispatching, scheduling, call booking, invoicing and the pricebook, plus clean-up of customer and equipment records. Nothing AI yet. The packages' own feature lists start here.
• Weeks 5 to 8: calls, after hours only. Connect the phone integration and start the virtual agent on after-hours calls with a written keyword list for escalation. The help article says calls are classed as Booked, Unbooked or Excused, unbooked calls that need a callback are tagged "Voice Agent Follow Up Needed", and all of it shows in Call Search, the Calls Dashboard and Agent Scorecards, so read every call daily at first.
• Weeks 9 to 12: the rest, one at a time. Atlas for technicians, then Dispatch Pro if your two-week time log shows the dispatcher's board is the bottleneck, then drafted review replies with a person approving each one.

Switch on one new AI feature at a time, so you can tell which one moved a number. If two arrive in the same week, an improvement in bookings could belong to either.

A scorecard to know whether the AI paid for itself

Record the baseline before go-live. The figures are for the made-up 12-technician plumbing firm from the break-even above.

If the booked-jobs row misses its trigger after three months, the break-even says to stop paying for the agent or renegotiate the rate, whatever the vendor's average results show. A feature earns its place on your numbers, and ServiceTitan's footnotes already say as much.

What the published-price tools offer for AI, by their own pages

A small firm does not have to go without. Housecall Pro's AI Team page lists Marketing AI, CSR AI (chat), Analyst AI, Coach AI and Help AI as included on all plans, with phone answering a paid extra, though the same page also says "CSR AI sold separately" and the pricing page lists CSR AI as an add-on, so ask which applies to you. Jobber's Receptionist is a $29 a month add-on that answers calls and texts and puts urgent callers through by keyword. Service Fusion sells ServiceCall.ai as an add-on on every plan with no separate price shown. FieldPulse names an office assistant, Operator AI for leads and bookings, a chat tool and 24/7 AI dispatching; the office assistant comes on all plans, while Operator AI and Chat AI are optional premium products.

Most of these are narrower than Titan Intelligence: only FieldPulse names AI dispatching, and none of the pages I read names a benchmark report or a price-comparison feature. They are wider in another way: most have a price you can read. For how the three best-known ones compare feature by feature, the three-way AI comparison is the place to start. If you would rather work the numbers for your own firm with someone, that is what my AI implementation consultation covers.

ServiceTitan AI questions from small HVAC and plumbing owners

Does ServiceTitan publish its prices?

No. Its pricing page describes per-technician pricing in three packages called Starter, Essentials and The Works, each with a Request Pricing button and no figures. The Pro products, including the virtual agent, show no prices either. Plan on getting a written quote that lists the per-technician rate for each package, any set-up fee and the usage rate for AI calls.

Can a five-technician firm use ServiceTitan?

ServiceTitan says its pricing suits any size and that even small businesses can scale field teams with it, but it states no minimum. Whether it pays depends on your dispatcher time, call volume and appetite for set-up. At five technicians, test a published-price tool first and treat ServiceTitan as the step up when you outgrow it.

Is Atlas available to every customer?

Not fully, going by its page. The field features, such as equipment information and troubleshooting help, are marked as available with Field Pro. The office features, such as launching campaigns, helping dispatchers and reviewing invoices, are marked coming soon. Ask for a live demonstration of exactly the Atlas features you would use on day one.

Will the AI Virtual Agent replace my customer service reps?

ServiceTitan positions it for overflow and after-hours calls, and it escalates to a live agent when asked, when it detects frustration or on set keywords. At launch it cannot add call notes, summaries or transcripts to the booked job record, so a person still tidies the file. Treat it as extra cover rather than a replacement.

Further reads

• HVAC Dispatch Software Compared: Which Has AI Routing for Small Fleets? — Dispatch software with AI routing, compared for small fleets.
• Best AI for Plumbers: What a Solo, Two-Van and Ten-Van Firm Should Buy — What a solo, two-van and ten-van plumbing firm should buy.
• How Much Do Missed Calls Cost a Heating and Cooling Business? — Put a price on missed calls before buying any call AI.
• AI Job Scheduling and Dispatch for Trades and Field Service Teams — AI scheduling and dispatch for trades and field service teams.
• AI Tools for HVAC Business Owners: Calls, Quotes and Maintenance Plans — AI tools for HVAC owners across calls, quotes and plans.
• AI Software for Plumbers: What a Small Firm Pays Each Month, by Van Count — Monthly AI software bills for one, three and eight vans, tallied line by line from vendor pricing pages: platform, AI receptionist and ChatGPT seats.
• Six AI Automations for HVAC Companies, Ranked by Effort — Six HVAC automations ranked from an hour of set-up to weeks of habit-building, with real vendor prices, sample messages and a break-even sheet.
• Workiz Genius AI: Which Add-Ons Are Worth It for HVAC or Plumbing — Break-even sums, test calls and rep questions for each Workiz Genius add-on, written for HVAC and plumbing firms deciding what to pay for.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: ServiceTitan Titan Intelligence, AI, Atlas, Contact Center Pro, AI Virtual Agent, Pro Products, HVAC and pricing pages, and its help-centre FAQ on the AI Voice Agent for Basic Phones and Phones Pro; Housecall Pro AI Team and pricing pages; Jobber, Service Fusion and FieldPulse pricing pages (all read October 2026).