Which videos and podcast episodes deserve your time?

1. How can you define the workflow before paying an automation agency?

262 — Stop Losing Money on HVAC Jobs: Automate Your Workflows with AI

• Who: Khalil Benalioulhaj and Martin Holland, The Cash Flow Contractor.
• Published: 3 July 2025; approximately 29 minutes.
• Format and evidence: Podcast; published episode notes and chapter markers.

Best for establishing what the agency is actually supposed to improve. The episode uses HVAC materials and billing problems to explore AI-assisted process mapping. It includes constructing a prompt, producing a Mermaid flowchart and walking through the resulting workflow. Useful entry points are 01:28 for process mapping, 05:07 for the flowchart and 11:30 for the detailed workflow. Episode notes

Our buyer takeaway: ask staff to correct the diagram before asking an agency to automate it. Mark who supplies information, which system holds the authoritative record, and what happens when something is missing.

Question for an agency: “Show us the current and proposed process, including exceptions and staff responsibilities—not just the automation diagram.”

Caveat: this is a planning method, not evidence of a deployed integration or audited savings. Its software references should not be treated as a current product comparison.

2. Could your existing job-management software do the work?

Introducing ServiceM8 15 / ServiceM8 2026 Launch

• Who: ServiceM8; company-produced launch presentation.
• Published: 18 September 2026, according to the launch announcement.
• Format and evidence: Embedded YouTube video; official release page. Runtime was not established.

Best for challenging unnecessary bespoke development. The launch page describes an Office Agent for conversational assistance and scheduled routine tasks, alongside a revised booking interface and greater visibility of scheduled customer messages. These are vendor-described capabilities, not independently tested outcomes. Launch and feature descriptions (servicem8.com)

For an AI receptionist for heating companies, also read the separate Phone Agent documentation. It describes taking messages, submitting online bookings, transferring calls and retaining call records. Do not confuse that description with proof that your particular booking rules are supported. Phone Agent documentation

Our buyer takeaway: give the incumbent vendor and the agency the same workflow specification. Ask each to identify configuration work, additional subscriptions and unsupported requirements.

Question for an agency: “Which requirement genuinely needs your integration, rather than an existing feature?”

Caveat: a launch presentation does not settle permissions, UK plan eligibility, implementation effort or total cost.

3. What should an AI receptionist be compared against?

Don’t Let AI Replace the Power of the Human Voice in HVAC Customer Service

• Who: Nicole Krawcke with Erica Leonor, identified as founder and CEO of Voxa.
• Published: 24 June 2026.
• Format and evidence: HVAC Chats podcast; publisher summary and chapter markers.

Best for an AI receptionist versus human answering-service comparison. Leonor discusses reception staff as part of the customer-acquisition process, including tone, communication, training and accountability. The AI discussion begins at 13:50; the programme also identifies a customer-introduction technique at 22:08. Publisher’s episode guide (contractingbusiness.com)

Our buyer takeaway: compare AI with a credible human alternative, not merely with today’s voicemail. Use the same test calls and assess whether the caller’s problem was understood, the booking was appropriate and a person could take over.

Question for an agency: “How will you demonstrate better overall handling than improved callbacks or trained overflow cover?”

Caveat: this is an expert interview, not a controlled comparison of answering services. The guest’s professional perspective should inform your testing rather than decide the result.

4. Where should voice AI operate—and where should it not?

AI Agents for HVAC Businesses: What Actually Works

• Who: Tersh Blissett with Jim Christian of Podium, Wrenches n Robots.
• Published: 15 September 2026.
• Format and evidence: YouTube/audio links, published transcript and chapters.

Best for defining a limited call-handling pilot. The discussion distinguishes agents from simpler automation, examines maintenance coordination, and considers when after-hours or overflow voice handling is preferable to routing every call through AI. Start at 14:50 for deployment boundaries and 17:02 for the discussion of abandoned calls and customer response. Transcript and chapters (homeservices.podium.com)

Our buyer takeaway: assess the complete call funnel, including callers who leave before a useful conversation. Do not judge performance only from conversations the AI successfully completed.

Question for an agency: “Which calls will remain human-first, and how will you distinguish unsuitable enquiries, abandoned calls, valid bookings and completed jobs?”

Commercial caveat: the show was introduced as a Podium partnership. Treat performance statements as speaker claims, not independent benchmarks. Partnership disclosure

5. What operational details appear in an owner’s account of using AI?

How AI Is Changing HVAC (Without Replacing Techs) | Chris Plunkett

• Who: Tersh Blissett with Tucson HVAC owner Chris Plunkett, Service Business Mastery.
• Published: 22 July 2026; approximately 50 minutes.
• Format and evidence: Podcast; syndicated programme notes and an available transcript excerpt.

Best for turning “after-hours automation” into operational questions. The published chapters cover improving processes before automation, overflow call handling, trigger words, on-call routing and late-night service policies. Start at 02:37, then 18:04 and 27:07. The available transcript also describes using AI to support procedures and customer-facing explanations. Notes and transcript excerpt

Our buyer takeaway: write the escalation rota and exception policy before configuring the voice agent. Include what happens when the nominated person does not answer.

Question for an agency: “Demonstrate a complaint, a caller requesting a person, an unavailable engineer and a failed transfer.”

Commercial caveat: the programme explicitly identifies a partnership with Breezy, the voice product discussed. This is sponsored operator testimony, not a neutral product test. Programme disclosure

6. Is maintenance administration a better starting point than answering every call?

Too Busy to Grow? Unlock Revenue with AI Automation Systems — Part 2

• Who: Gary McCreadie with Jim Christian of Podium, HVAC Know It All.
• Published: 25 June 2026.
• Format and evidence: Podcast and YouTube links; published transcript and chapters.

Best for maintenance reminders, scheduling and supplier due diligence. The conversation examines the administrative work behind maintenance agreements, then turns to evaluating AI products and suppliers. Relevant chapters begin at 02:43 for maintenance workload, 10:20 for evaluating tools and 12:34 for supplier checks. Transcript and chapter guide

Our buyer takeaway: separate appointment scheduling, contract renewal, payment chasing and marketing. Build a clean exception list for disputed records, incomplete details and customers needing personal contact.

Question for an agency: “Can the system explain why each customer was contacted, what happened and which records require staff attention?”

Commercial caveat: the guest represents the product supplier, and examples are not independently audited. The title’s revenue language is not a forecast for your business.

7. What should an operations director ask before buying AI dispatch?

HVAC Contractor Turns Dispatch Data Into Revenue Growth

• Who: Nicole Krawcke with Brady Jolly, identified as owner and CEO of Jolly Heating, Air, Plumbing, Electric.
• Published: 9 September 2026.
• Format and evidence: HVAC Chats podcast; publisher summary and embedded player.

Best for understanding the questions around AI-assisted dispatch. The episode describes matching technicians with jobs using performance information, alongside call answering, follow-up and review requests. Its published summary also identifies implementation challenges, technician development and human oversight. Episode source

Our buyer takeaway: require operational constraints before optimisation. Your specification should identify competencies, customer commitments, travel assumptions and circumstances in which a dispatcher must override a suggestion.

Question for an agency: “Can staff see why a job was assigned, change the assignment and review the consequences?”

Caveat: a named contractor’s experience is more useful than an anonymous sales claim, but it remains one account. This source does not establish a transferable UK revenue uplift or prove that the same approach fits your workforce.

8. Which episode works as a management-team introduction?

We Ask an AI Expert How HVAC Contractors Should Use AI

• Who: Mary Carter and Stephen Ross with Chris Smith of All Contractor, HVAC Full Blast.
• Published: 28 September 2026; 45 minutes 3 seconds.
• Format and evidence: Podcast with embedded YouTube video; publisher description.

Best for giving managers a shared starting point. The published description covers administrative efficiency, reporting, customer service and business decisions rather than a single receptionist product. Use it to identify candidate workflows, not to approve a technology stack. Episode and speaker information

Our buyer takeaway: ask each manager to nominate one task, its current owner, the information it needs and the consequences of a mistake. Reject proposals described only as “transforming the business”.

Question for an agency: “Which single workflow would you pilot first, and what evidence would make you recommend stopping?”

Commercial caveat: the guest is a consultant and author. The publisher expressly states that Trane Technologies does not independently verify guest statements or endorse referenced third parties. Publisher disclaimer

9. What can heat-pump installers learn about AI knowledge systems?

Heat Pumps, AI & Donkeys

• Who: Nathan Gambling with Graham Hendra, BetaTalk.
• Published: 7 September 2026; 64 minutes 46 seconds.
• Format and evidence: Podcast with public transcript.

Best for separating knowledge retrieval from technical authority. Hendra discusses Energy Donkey, developed with Planet Devices, including curated information, homeowner questions, equipment information and proposed connected functions. Crucially, around 40:02, the conversation distinguishes the initial launch from capabilities intended for later stages. Episode transcript (buzzsprout.com)

Our buyer takeaway: for heat-pump enquiry qualification, specify approved information sources and require the system to acknowledge missing evidence. Keep collecting survey information separate from deciding system suitability or making technical commitments.

Question for an agency: “Which answers come from approved material, how are revisions managed, and when must the system defer to an engineer?”

Commercial caveat: the guest is discussing his own project. The recording is not proof that every described feature was available at publication—or is available now—and it is not a safety validation.