
電話に出て見込み客を選別し、予約まで入れるAIエージェント:800.com の AI 受付
This AI Agent Answers the Phone, Qualifies, then Book's Calls

電話に出て見込み客を選別し、予約まで入れるAIエージェント:800.com の AI 受付
AI Marketing
要約
AI Marketing のホストが、800.com でマーケティングを統括する Tom English 氏に、同社のインバウンド向け AI 受付エージェントについて聞いた回。自社サイトを読み込んで知識ベースを作り、質問対応・リード選別・予約受付を行う。24時間365日の人間代替ではなく、不在時にボイスメールの代わりとして使うものだと位置づけている。主な対象は B2C の個人事業主や小規模事業者で、コールセンターは対象外としている。
この回のポイント
- ●800.com は事業用電話番号と転送サービスから出発し、不在着信をボイスメールに送るだけの状況を改善する AI 受付を提供している。
- ●設定は自社サイトの URL を入力するだけ。主要ページから知識ベースを作り、営業時間や FAQ を学習する。文書追加やプロンプト編集で追加学習もできる。
- ●全通話は録画・文字起こし・要約され、通話ごとのスコアを見てプロンプトや資料を改善できる。リード選別ではサイトの見積もりフォームと同じ項目を聞き取り、スコア付けする。
- ●人間への引き継ぎは、発信者の希望、緊急時、AI が機能不全に陥った場合のフェイルセーフの3通りで設定できる。
- ●主な対象は住宅サービス、不動産、コーチなどの B2C 事業者で、大規模コールセンターは対象外。電話での決済は現時点で不可だが、2027年に変わる可能性に言及した。
- ●7日間の無料トライアルがあり、ホリデー向けのサンタやエルフの声も追加された。
章立て
自己紹介と800.comの概要
Tom English 氏が経歴を語り、事業用電話番号と転送から AI エージェントへ進化した流れを説明する。
ボイスメールより良い AI 受付
不在着信が多い小規模事業者向けに、ボイスメールの代わりとなるインバウンドの AI 受付を作った背景を述べる。
サイトからの学習と通話レビュー
URL を入力するとサイトを読み込み知識ベースを構築する仕組みと、録音・要約・スコアによる改善方法を解説する。
回答・選別・予約とカレンダー連携
主な3機能と、複数エージェントの構成、リードのスコア付け、予約リンク送信について話す。
不在時の利用と人間への引き継ぎ
24時間代替ではなく、現場作業中や夜間・休日の代替として使う考え方と、3つの引き継ぎ条件を説明する。
導入事例と無料トライアル、音声
Zero Empty Spaces の事例、不在着信が機会損失になる点、7日間トライアルとホリデー音声を紹介する。
対象顧客とコールセンターの扱い
B2C の個人事業主に注力している理由と、大規模コールセンターや決済対応が現状では対象外である点を語る。
ホストの総括と連絡先
ホストが企業は同種の解決策を検討すべきだとまとめ、Tom 氏が試用方法と電話番号を案内する。
解説記事
AI Marketing の本エピソードでは、800.com でマーケティング部門を率いる Tom English 氏が、同社のインバウンド向け AI 受付エージェントについて紹介した。電話に出られない小規模事業者のために、ボイスメールよりも役に立つ受け皿を用意するという発想が中心だ。
ボイスメールの代わりという位置づけ
800.com はもともと事業用電話番号と通話転送を扱ってきた会社だ。English 氏によれば、小規模事業者や起業家が着信を逃し、発信者をボイスメールに回すしかない状況に課題を見て、AI 受付の開発に力を入れたという。
興味深いのは、他社の多くが打ち出す「24時間365日対応」とは違う見方を取っている点だ。同氏は、チームの完全な代替ではなく、現場作業中、夜間、週末、休暇シーズンなど手が離せないときにボイスメールの代わりに使うものだと述べている。
サイトを読み込んで学習する仕組み
設定は自社サイトの URL を入力するだけだ。主要ページをスクレイピングして知識ベースを作り、事業内容、営業時間、よくある質問を把握したエージェントが生成される。社内資料の追加やプロンプトの編集も可能で、学習は好きなだけ続けられるという。
リード選別(見込み客の絞り込み)では、サイト上の見積もりフォームも読み取り、同じ項目を電話で聞き取れるようにする。全通話は録音・文字起こし・要約され、通話ごとにスコアが付く。担当者はそれを見てプロンプトや資料を改善し、次の通話の品質を高められる。
できることと複数エージェント構成
主な機能は、質問への回答、リード選別、予約受付の3つで、伝言を預かることもできる。エージェントは複数作れる。たとえば、通話を適切な相手へ転送するだけのフロントオフィス役と、台本どおりに質問してリードをスコア付けする選別役を分けられる。翌日にスコアと回答内容を確認し、有望な見込み客かどうかを判断する運用だ。カレンダーへのリンク送信にも対応している。
人間への引き継ぎは3つの場合に設定できる。発信者が人間や特定の担当者を希望したとき、緊急時、そして AI が理解できず誤った対応を続けるなど機能不全のときのフェイルセーフだ。ホストが挙げた「リードスコアが70を超えたら即転送」といった条件も設定できる、と English 氏は同意した。
対象顧客と現時点の限界
注力するのは B2C で、住宅サービス、不動産業者、物件管理者、コンサルタント、プロのコーチ、個人事業主などだ。顧客対応中のため新規の電話を取れない人を想定している。事例として、フロリダで空き施設をアートスペースに変える Zero Empty Spaces が、週末や出張中の電話対応に使っていると紹介された。
English 氏は自身の家を購入した経験から、Google マップで検索して最初の業者に電話し、出なければ次へ移ると語り、不在着信は機会損失だと強調した。
一方、大規模コールセンターとは正面から競ってきておらず、転送やルーティング、同時通話数など別種の課題があるとしている。電話での決済も現時点では不可で、2027年には変わるかもしれないと述べた。ホストは、大企業の長い待ち時間を例に、十分に訓練されたボットなら歓迎だという見方を示した。
試用方法
7日間の無料トライアルが今月始まったばかりで、通話分数は少ないものの、設定して応答を試すには十分とのことだ。声の種類も豊富で、ホリデー向けにサンタやエルフの声が加わった。800.com/AI agents か、電話番号 844-TRY-800-AI で、同社が用意したエージェントと実際に話せる。
まとめ
本回の要点は、AI 受付を人員削減の道具ではなく、取りこぼしを減らす補完策として売り出している点にある。サイトから知識を作り、通話ログで改善を重ね、必要なら人へ渡す設計は、中小事業者にも導入しやすい構成だ。日本の個人事業主や小規模サービス業でも、不在時の機会損失は共通の課題だろう。ただし決済不可や B2C 中心という制約も明言されており、用途の見極めが必要だ。番組内で語られたのは提供側の説明が中心で、効果の定量的な裏付けは示されていない。
文字起こし(英語・自動生成)
The way that customers work, especially with myself, I just bought a house recently with home services, I'll just search it on Google Maps, call the first one. If he doesn't answer, I'm calling the next one until I get someone who picks up and is able to come over. This helps with that because people are actually getting to somewhere where they can book an appointment. Hello, everyone. Welcome to AI Marketing. Joining me today is Tom English of 800.com, and we're going to talk about his agents and what they do in terms of helping you sell. So, Tom, before we jump in, why don't you give me a little background on yourself? Yeah, you mentioned I'm Tom English. I work for 800.com. I run the marketing department. I've been there for about almost five years. In March, I'll be there for five years. Yeah, so it's been a fun ride so far going from, you know, our legacy product of business phone numbers and call forwarding to where we're at now with the agents.
Yeah, I was going to say 800.com sounds like, you know, the 800 calling service that I used to do a long time ago. So it sounds like you've evolved. Yeah, we've been around for a long time. That's wonderful. Okay, so what exactly at a high level is 800.com doing now? So, yeah, so obviously we've been like our legacy product has been the business phone, the phone number, call forwarding and doing things like that. We saw a need for, you know, businesses, especially small businesses, entrepreneurs who are missing phone calls, a way for them to, you know, do something more than just send their callers to voicemail. and that's why we started putting effort into creating an AI receptionist to have something better than voicemail to send callers to. So this is inbound? Inbound. And it's an AI receptionist. I assume it's been trained on that person's business?
Yeah, so what it does is when you get set up, you put in your website, it'll go, it'll scrape the main pages of your website, build a knowledge base based off of your website so it knows what you do, knows your hours, knows frequently asked questions. You can add any additional documents that you have to. So it learns everything about your website, and then it creates the agent, you know, to be the knowledge focus of your phone call. Okay, and can I train it further, or is that the extent of it? No, you can train it further. You can train it as much as you want. You can put as many, like, docs that you might have, any prompts. You can edit all the prompts. like say if you were doing like lead qualification, any type of, you know, certain fields that you might want. Like the one thing we like to do is like if you have a form on your website, we also go in, we'll scrape that to make sure if you're setting up a qualifying agent that it's pulling exactly what you would pull if someone were to, you know, request a quote from your website. Got it.
And so you could, is the call being recorded so that you can go back and train the AI some more if they made some mistakes? Yeah yeah So all the calls are recorded transcribed and summarized through the entire AI suite that we have And then yeah with each call you can go in you can see the score of it and you can go in and improve the prompt, or you can create more documents to make the agent better at handling your phone calls. Okay, so let's quickly walk through kind of a use case. So I set this up, and I want my AI answering. In some cases, I want to talk to AI more than I want to talk to an untrained human anyway, especially from a foreign country. No offense to them, but a lot of them don't have a grasp of the English language or that business that they're doing it for. So the AI then answers, and then what happens? So it depends on what skills you want the AI to learn.
It could just be as simple as answering questions. It could be, like I said, qualifying leads. It could book appointments for you. Those are like the main three things that our agents do. So, like, there's a lot of things that it can do there. If you just wanted to, you know, take a message for you when you're not there, it can do that. Okay. And are they qualifying the lead potentially or is it customer service or is it both? It's both, yeah. You can really set it up how you want to set it up. You can create agents that do different things. so you just don't have one. You can create one that's, you know, purely just a front office that's able to, you know, just transfer calls to where you need to go. Or you can set one up to be, you know, purely just qualifying leads as they come in. It follows exactly your script, you know, asks for everything that you need. And on the back end, it'll score all those leads for you so that when you come in, say, the next day, you can come and look at the call, and it'll have all of, like, the things
scored in there for you, like what they answered, and gauge it as whether it's, you know, qualified or not qualified. Okay. If it's qualified, does it send out like something like a Calendly booking link? So it can do that, yes. So it has the ability to send links, and it has the ability to send links to your calendar. So it's able to do all that as well, yeah. Okay. And so after hours, I assume, you know, that this now becomes a potential profit-making solution as opposed to voicemail and hope somebody gets back to me the next day. Yeah, that's kind of like the biggest thing that we've seen with doing this and like our whole intention with doing this is not something – there's a lot of companies out there that are pushing their AI receptionist to be like 24-7 answering. We don't see it the same way. We see it as just like that thing instead of voicemail, like when you're on the job or, like you said, after hours on the weekends. We have the holidays coming up. You know, we set the age and stuff so that it can answer when you're not around. and then when you come in, you have, you know, all your calls are at least handled in some certain way
and maybe taking even a step further. Okay, and then when, so when does the AI hand off the call to a human? So the AI will hand the call off to the human in two cases, well, three really. So pretty much if you ask to speak to a human, you can set it up so that, you know, if it says the caller is like, hey, I want to speak to a human, or say, hey, I want to speak to Mark, it able to transfer directly to you or directly to whomever the human is that is supposed to be answering the call You know in emergencies you can set it up as like if there an emergency transfer it to me Say this is, like, after hours, home service emergency calls. And then fail-safe, if anything were to, like, the AI is not understanding or something's going wrong with the AI or it's just, like, completely miscommunicating and is failing, then there's a fail-safe number that will get transferred to, and that will go to a human as well. Okay, so you can program it depending on –
Yeah, for sure. Like if I say, hey, if this reaches a lead score above 70, you know, transfer to human right away or something like that. Okay. All right. So, all right. At the end of the day then, you know, what is it that you're hearing from businesses that are using this that is either saving them time, money? I mean the list would go on if it works well yeah so definitely saving them time we had a recent case study with a customer called Zero Empty Spaces in Florida who does like these like he takes abandoned or unused locations in Florida and turns them into art spaces he does a lot of traveling he does a lot of his work over the phone and he was able to set up the agent to answer like on the weekends and when he's traveling so that he doesn't have to be there and answer every call. So it's definitely saving people time. And I feel like it's gaining people money, actually,
because missed calls are missed opportunities. And if you're in, say, home services, like every call is a potential lead. And, like, what we've seen and what I'm sure everybody has been reading online is that, like, the way that customers work, especially with, like, myself, you know, I just bought a house recently with home services. I'll just search it on Google Maps, call the first one. If he doesn't answer, I'm calling the next one until I get someone who picks up and is able to come over. This helps with that because people are actually getting to somewhere where they can book an appointment. Okay. And you've got a seven-day free trial, right? Yeah. We've just launched a seven-day free trial earlier this month. So anybody can get in, you can put your website in, get started. It's not a robust pretrial. You get seven days, you know, a small amount of minutes, but it's really there to be able to set it up, see how it can answer, and, you know, test it out.
Okay. And different voices, or is it one voice? No, there's so many different voices. and what's actually funny, I don't know when this is going to air, but for the holidays we actually just added some fun holiday-themed voices as well. Yeah, so you've got Santa and you've got elves and stuff like that. Oh, wow. Just add something fun to it. And then to me it sounds like it's more B2C, but I suspect you could probably do B2B, but I think it's really complicated. Are you more focused on one or the other? we're definitely more focused on B2C. A lot of the businesses that we're looking after right now is like I said home services real estate agents property managers business consultants professional coaches things like that like solopreneurs things where there not a lot of people around to answer the call and you out there hustling You're making your money by being either in person on meetings or doing meetings on your
computer or stuff like that that you can't answer new client calls while also handling the clients that you already have. So we're really focusing on that person that needs someone there to answer the phone. Okay, and so why don't you do like mid to large size companies where you're just replacing call centers? And I know I'm going to take a lot of flack for saying that, but why couldn't you do that? I mean, you could definitely try to do that. It's a little bit out of, you know, and for me, like the realm of like thinking, because we've never really gone toe to toe with full call centers before. I would imagine there's a lot of transferring and routing and number of agents and number of calls being able to answer at the same time that we'd have to think about. It's not to say that you can't do it. It's just I think call centers become like a different type of animal once you come into that call volume and things like that. Well, I think you guys should take a look at that because, you know, when I get on the line with any big company, it's always a 40-minute wait, first of all. Sure. Especially airlines.
And then, you know, I'd much rather talk to a bot anyway as long as it's well-trained. And I suspect the issue is if the bot were to try and do anything financial, then humans need to be in that loop. But I still feel like that could easily be programmed in. For sure. Yeah, none of our AI agents currently, and that's not to say that won't change in 2027, able to take payment over the phone, not yet at least. Yeah. Okay. So, yeah, I mean, an interesting market to kind of take a look at. All right. So great summary. Glad you walked through everything. You know, for the marketers listening to this, to me, this is something that every company should be looking at, whether you're small or large. I know Mr. English here has positioned this as for small entrepreneurs and solopreneurs.
It's a no-brainer, but you might want to, whether it's his solution or somebody else's, I'd always be looking at solutions like this that they're just going to be better for your customers. I mean, sure, maybe you've got some really well-trained customer support people, but 24-7, and they're always on. They're never in a bad mood, and when they're insulted, they don't take it personally. I've managed some of these call centers before, so you've got to take a look at this. It's probably not in your department, but get somebody that's running customer support to kind of take a look at a solution like this or another one that's handling or focused on bigger organizations. So with that, Tom, how do people get a hold of you? Yes, you can go directly to the website, 800.com forward slash AI agents to check it out, get to the free trial. Also, if you're, say, on the go right now, you can call 844-TRY-800-AI. You can speak directly to the AI agents that my team has set up, and that's like our spokesperson is in that line right there, and you can get started there too.
Excellent. All right. Thanks, Tom. Really appreciate it. Thank you.
番組の概要欄(原文)
I sat down with Tom English, who runs the marketing department at 800.com, about the company's inbound AI receptionist. We talk about how it trains on a business's website, qualifies leads, books appointments, and steps in when a call would otherwise go to voicemail. DESCRIPTION Tom English runs the marketing department at 800.com. He has been there almost five years, and in March it will be five. The company started with business phone numbers and call forwarding. It now offers an inbound AI receptionist for small businesses and entrepreneurs who were sending missed calls to voicemail. You set it up by entering your website. It scrapes the main pages, builds a knowledge base, and learns what you do, your hours, and frequent questions. You can add documents, edit prompts, and, for lead qualification, have it pull the same fields as a quote form on your site. Every call is recorded, transcribed, and summarized. You can see a score on each call and improve the prompt or add documents so the agent handles the next one better. The main skills Tom walked through are answering questions, qualifying leads, and booking appointments. You can also have it take a message. You can run more than one agent, including a front office that transfers calls and a qualifying agent that follows your script and scores leads for you to review the next day. It can send calendar links as well. Tom's view is that this is not mainly a 24/7 replacement for a team. It is what you use instead of voicemail when you are on the job, after hours, on weekends, or over the holidays. The AI can hand a call to a human if the caller asks for one, in an emergency, or if the agent is failing. 800.com is focused on B2C work: home services, real estate agents, property managers, business consultants, professional coaches, and solopreneurs who cannot take a new-client call while they are already with a client. Tom said they have not gone toe to toe with full call centers, and that their agents cannot take payment over the phone yet, though that could change in 2027. He pointed to a customer, Zero Empty Spaces in Florida, who uses the agent on weekends and while traveling. His own example: when he bought a house, he searched home services on Google Maps, called the first one, and if no one picked up, he called the next. Missed calls, he said, are missed opportunities. Tom said there are many different voices, and that they had just added holiday-themed ones, including Santa and elves. 800.com has a 7-day free trial, and Tom said you can try it at 800.com/AI agents or by calling 844-TRY-800-AI. KEY TAKEAWAYS - This is an inbound AI receptionist, built as something better than voicemail for missed calls. - The agent is trained on your website, including hours and frequent questions, and you can keep training it with documents and prompts. - For lead qualification, it can use the same fields as a form on your site, follow your script, and score the leads. - Every call is recorded, transcribed, and summarized, so you can review the score and improve the prompt. - It can answer questions, qualify leads, book appointments, send calendar links, take a message, and transfer calls. - It hands off to a human if the caller asks, in an emergency, or if the AI is failing. - 800.com is focused on B2C operators, such as home services and solopreneurs, not full call centers. The agents cannot take payment over the phone yet. CHAPTERS 00:00 Teaser 00:27 Welcome and introductions 01:26 An AI receptionist instead of voicemail 02:08 Training the agent on your website 03:08 Recorded calls, transcripts, and scores 04:08 Answering, qualifying, and booking 05:35 After hours, instead of voicemail 06:21 When the AI hands off to a human 07:32 Missed calls and a customer example 08:58 Free trial and voices 09:47 Who it's for, and who it's not 13:13 How to reach Tom NOTABLE QUOTE Tom English: "Because missed calls are missed opportunities, and if you're in, say, home services, like, every call's a potential lead." LINKS AGENT SHOCK: Surviving and Profiting from the Rise of AI Agents and Robots, the book I wrote with Jas Dhillon, is out in early December. You can reserve a copy at https://www.agentized.com/book Agentized: https://www.agentized.com My LinkedIn: https://www.linkedin.com/in/fidelman/ X: https://x.com/markfidelman Instagram: https://www.instagram.com/markfidelman/ YouTube: https://www.youtube.com/@AIMarketingMark
出典・この記事について
- 番組
- AI Marketing(Mark Fidelman)
- 元の回
- This AI Agent Answers the Phone, Qualifies, then Book's Calls
- 配信日
- 要約作成日
- 公式サイト
- www.fanaticsmedia.com
この要約・解説記事は、元の音声を文字起こしした内容をもとに AI(Claude)が日本語で作成したものです。 固有名詞や数値に誤りを含む場合があるため、正確な内容は元の音声でご確認ください。制作方法について
関連エピソード

地下室から年商1,000万ドルへ:父娘ブランド Lilac Blonde のAI活用術
AI Marketing

マーケティングダッシュボードは死んだ:AIエージェントが主役になる
AI Marketing

AIに賃貸物件の鍵を開けさせますか? MyRemotely創業者Ran Fridmanが語る無人内見と賃貸管理の自動化
AI Marketing

AIでEコマース企業をゼロから立ち上げる方法
AI Marketing

AIはマーケティング分析をどう変えるか:データサイエンスの役割と既存データ活用
AI Marketing

ボイスエージェントはどう会話のリズムを学ぶのか — Shawn Wen(PolyAI)
Machine Learning Street Talk (MLST)