Making Job Search Easier by Finding the Great Companies First

Find a
JOB
Title/Keywords Company Name
City, state or zip (optional)
 

The Techcrunch Interviews: FuseMachines Artificial Intelligence Lead Machine

Fusemachines on Media Jobs

The arti­fi­cial Intel­li­gence lead machine has arrived!   The name of the lead mas­ter is Fusema­chines and Sameer Maskey is the Founder and CEO and his team have been devel­op­ing the soft­ware that applies Arti­fi­cial Intel­li­gence, Nat­ur­al Lan­guage Pro­cess­ing and Machine Learn­ing to the process of lead gen­er­a­tion for B2B com­pa­nies.

At Media Jobs we are always inter­est­ed in find­ing the most inno­v­a­tive com­pa­nies where you might want to work.  The big talk today is about Arti­fi­cial Intel­li­gence and today we’re talk­ing with a com­pa­ny that is using the tech­nol­o­gy to dri­ve more sales.

Its tak­en 3 PHD’s and 30 engi­neers to cre­ate SAM, the name they are call­ing their AI sys­tem.  But it gets bet­ter.  SAM inter­acts with humans through emails, mobile apps and sales soft­ware and learns from user behav­ior to per­form some of the same tasks auto­mat­i­cal­ly. SAM learns and improves itself every­day!

Today, with a staff of 80 the com­pa­ny is blaz­ing a trail to take the pain out of lead gen so that us humans can focus more on cash gen­er­a­tion then lead gen­er­a­tion.  If you like what you hear check out some of the jobs avail­able now at Fusema­chines.

In this pod­cast I speak with Alex Roeskestad from Fusema­chines recent­ly and he talked about how every­thing works and some of the cus­tomers that are cur­rent­ly ben­e­fit­ing from SAM.

Roy: This is Roy Weiss­man from Mediajobs.com and we’re talk­ing with Alex Roeskestad from Fuse Machines. What is Fuse Machines and how did this come about? I see it says arti­fi­cial intel­li­gence for sales. Does that mean we no longer need sales peo­ple? What’s the prob­lem we’re solv­ing here?

 

Alex: That’s a good ques­tion. I don’t think you take the sales per­son out of the equa­tion, but we try to do in so many words, is instead of hav­ing a sales depart­ment or a CEO or who­ev­er does sales, instead of hav­ing that depart­ment spend a lot of time doing research and kind of wast­ing their time being bogged down by these things, we try to opti­mize the sales depart­ment to such an extent that the sales per­son could be focus­ing much more on the sales aspect of sell­ing. So that being on the phone or writ­ing with emails and clos­ing deals ver­sus spend­ing a lot of time research­ing infor­ma­tion. To give you one exam­ple, say that you’re sell­ing a soft­ware solu­tion and you want to have the right indus­try, you’re sell­ing it to an indus­try. You’re try­ing to reach CEOs of small com­pa­nies, but you also want to fig­ure out your hir­ing needs. If these guys are hir­ing some­thing where your prod­uct could be com­pli­men­ta­ry or could even take the job. You also want to have your Face­book feed, if that’s impor­tant, those four ele­ments. You could cer­tain­ly have a per­son do that for you, whole­sale depart­ment. What we do is that we opti­mize it to such an extent that it takes us one sec­ond to find 1,000 leads that are pre-qual­i­fied leads, ver­sus spend­ing weeks and weeks gath­er­ing a list of 1,000 peo­ple. That’s kind of one way of doing sales using arti­fi­cial intel­li­gence.

 

Roy: How do you say it takes you a sec­ond to find 1,000 qual­i­fied leads? How does the com­put­er know the lead is qual­i­fied?

 

Alex: Say that you’re the head of sales and you want to reach out to head of sales, head of mar­ket­ing, CEO. You want the com­pa­nies to be 1–10, 11–50 if you were to use the length of para­me­ters. You want them to have a cer­tain fund­ing struc­ture, series B if that’s impor­tant to you or cal­cu­la­tions of rev­enue. We will basi­cal­ly scrape that infor­ma­tion for you. We have all the infor­ma­tion in the world and we can also scrape it on a dai­ly basis and we’ll find all the infor­ma­tion out there. We’ve spent a lot to give you some back­ground. This has actu­al­ly been devel­oped at Colum­bia Uni­ver­si­ty by our CEO, Sameer Maskey, over many years. It’s not a sim­ple thing. It’s almost like an aca­d­e­m­ic endeav­or dis­guised as a com­pa­ny in a way. It’s a very com­plex machin­ery in the back­ground, but we will find it for you using a lot of arti­fi­cial intel­li­gence and some human beings. It does it extreme­ly quick­ly.

 

Roy: You said it scrapes web­sites. A lot of times when you scrape web­sites you get a lot of garbage infor­ma­tion and email address­es, info@, webmaster@, how does the soft­ware val­i­date so that you end up with a good clean data­base? The rea­son peo­ple end up hir­ing peo­ple is because they need some­body to look at the infor­ma­tion and say this com­pa­ny can’t be at 852 West 38rd Street, there’s no such address or info@ is worth­less or John Jones has­n’t worked at this com­pa­ny in five years. How does the soft­ware address all of these issues so that you actu­al­ly end up with a high­ly accu­rate data base?

 

Alex: Right, that’s a very good ques­tion. Let’s start with say­ing does Don John­son work here? We can dou­ble check that by shoot­ing out sig­nal to LinkedIn, so we scrape 24/7 so if Don John­son changed his job last week, we re-scrape this week and we know there’s been a change there. We don’t sell leads which are junk leads in the sense that the per­son isn’t work­ing there in that posi­tion.

 

Your point about does this street exist. Like 250th Street, that does­n’t exist in the New York City in Man­hat­tan. We will then check the post office and data­bas­es, Google maps, etc. So we will make sure that the infor­ma­tion is cor­rect then. What it does for us is that first it takes a snap­shot of every­thing out there. Then it goes through and starts to clean all the data. You will see that in the clean­ing process, it marks, for exam­ple in the col­or red, invalid emails. Emails that the machine knows are incor­rect. These emails can’t be cor­rect based on the pat­tern in that com­pa­ny. If a pat­tern is first and last name at domain and sud­den­ly an email in that com­pa­ny comes up with a com­plete­ly dif­fer­ent ver­sion, it will mark that per­son as most like­ly incor­rect. You can always over­ride it, but it will help you go through junk leads. If peo­ple aren’t work­ing there, the machine knows it. If the email is wrong then the machine knows it.

 

Roy: Does it val­i­date the emails or var­i­ous val­i­da­tion sys­tems where it will tell you if the email is valid? Does it do that?

 

Alex: Yes. We will send out what we call a hand­shake and it will come back in basi­cal­ly reds, orange, and green. Red are invalid emails. We know that they’re incor­rect. Yel­low is what they call accept all. It’s a prob­a­bil­i­ty that they’re cor­rect, but there’s some uncer­tain­ty. Green are valid emails. You basi­cal­ly move the emails through the tabs and the sys­tem helps you crack the emails and if you give it some time, it will give you all the cor­rect emails, all valid, all greens. Then you can take the peo­ple with the emails and you can cre­ate a tem­plate. It shoots out to all these peo­ple. You don’t have to use gmail or mail chimp or some oth­er sys­tem.

 

Roy: So this is also a CRM?

 

Alex: That’s a tough ques­tion. It does­n’t take over the job in sales for us or any­thing else. It tries to be as close to a CRM with­out step­ping on the CRM’s toes.

 

Roy: I assume it’s inte­grat­ed with com­pa­nies like Sales Force. What kind of inte­gra­tions do you have?

 

Alex: We are work­ing on that inte­gra­tion. It’s not yet there. We are work­ing on it to have a bridge from Fuse Machines to SAM, a soft­ware choose sales force. We’re also work­ing on oth­er inte­gra­tions with a big kind of CRM. We’re try­ing to inte­grate with every­one out there. They’re kind of being dis­cussed as we speak.

 

Roy: Are you inte­grat­ed with any pri­vate data­bas­es for gath­er­ing infor­ma­tion?

 

Alex: No, we scrape every­thing in house. We don’t pur­chase oth­er lists. We don’t …

 

Roy: So it’s all pub­licly avail­able.

 

Alex: Yes, peo­ple often ask us if we can hack infor­ma­tion, which we can, but we would­n’t be stand­ing here talk­ing, I guess. It’s pub­lic infor­ma­tion. We scrape North Amer­i­ca, West­ern Europe, and New Zealand and Aus­tralia. In the future we’re going to make in-roads into Chi­na and per­haps Rus­sia, but that’s kind of into the future. It’s good to be a lead gen­er­a­tor for the world, but for the time being we’re focused on Eng­lish.

 

Roy: You would antic­i­pate work­ing any deals for API’s with pri­vate data­bas­es?

 

Alex: I could be. I’m not sure. We’re still ear­ly. We’re a small com­pa­ny. We have 40 clients. So I think it’s a bit ear­ly to make those strate­gic deci­sions, but we’re open­ing up for part­ner­ships. We have a cou­ple of part­ners. We’re try­ing to kind of, I would­n’t say rev­o­lu­tion­ize the space, I think that’s too much of a big word, but we’re try­ing to re-think sales for peo­ple. That’s also why we’re here today is to tell peo­ple that sales can be done a bit dif­fer­ent­ly than hav­ing humon­gous sales teams doing research and wast­ing time in the offices.

 

Roy: Can you tell us any names of any of your clients? Notable clients. Can you men­tion any­one?

 

Alex: Sure, I hope so. We’ve signed with a com­pa­ny called Mon­do. Expand we signed last Fri­day. We have a lot of small­er com­pa­nies. Some are based in Eng­land. I also spoke to a rep­re­sen­ta­tive of the gov­ern­ment of Slove­nia which were inter­est­ed in pur­chas­ing this. I don’t know for what rea­son, but they’re inter­est­ed in lead gen­er­a­tion and kind of get­ting the start­up scene to become a lit­tle more vibrant in Slove­nia. There’s kind of dif­fer­ent strokes for dif­fer­ent folks.

 

Roy: How would you define the mar­ket both in size … what kind of num­bers would you define the mar­ket for this as in dol­lars?

 

Alex: That’s a very good ques­tion. As I said, we’re a small com­pa­ny and it depends on where we want to aim in the end. The mar­ket is humon­gous because arti­fi­cial intel­li­gence and sales could be sold to any­one. I’m also talk­ing B to C. The mar­ket is humon­gous. For us as a com­pa­ny I think we’re going to hone in on the B to B side. Be more spe­cial­ized in tech soft­ware, basi­cal­ly the com­pa­nies that you see here today, and TechCrunch Dis­rupt. We can do oth­er types of com­pa­nies, but for us it’s also impor­tant for us to be very good at what we’re doing and not spread­ing our­selves too thin. It’s bil­lions of dol­lars, for sure.

 

Roy: When you do research for com­pa­ny A, are you sav­ing that in a data­base, so when com­pa­ny B needs the same infor­ma­tion you already have it or are you just pro­duc­ing it one off thing.

 

Alex: Very good ques­tion. We try not to resell infor­ma­tion, so say that we have five media agen­cies with us, we try not resell the infor­ma­tion so that when com­pa­ny B reach­es out and A has already done it, then they kind of meet the same answers and we’ve already spo­ken to com­pa­ny A. It’s impor­tant to us that the leads are fresh and they’re not resold. This is a very big space so we try to build sets of infor­ma­tion. We try to build data­bas­es for dif­fer­ent types of sec­tors. We could very eas­i­ly tap into these sec­tors quick­ly. It’s kind of a fine bal­ance between not reselling infor­ma­tion which could hurt com­pa­ny B, but also hav­ing an info cen­ter inter­nal­ly at Fuse Machines so that we could sell this at mass.

 

Roy: You said the mar­ket was humon­gous. For your mar­ket, where you guys are look­ing, do you have a dol­lar val­ue for that mar­ket?

 

Alex: No, to be hon­est we don’t have a dol­lar val­ue because it’s very hard to assess what’s going to hap­pen with the tech mar­kets where we’re going. We have a lot of cus­tomers. A lot of peo­ple today are kind of sur­prised that we’ve pushed arti­fi­cial intel­li­gence so far. It does­n’t mean they con­vert into cus­tomers because there’s some ten­sion between the old way of doing sales and using arti­fi­cial intel­li­gence. Some peo­ple think it’s a buzz word and they don’t buy it. It’s some­what hard to see where the mar­ket is going. I would say it’s in the bil­lions of dol­lars, but I don’t dare to put a straight num­ber and say this is the val­ue. I would rec­om­mend all com­pa­nies to get in on this before it’s too late.

 

Roy: Why would it be too late?

 

Alex: Because we give peo­ple very good deals in the begin­ning. To be hon­est with you this arti­fi­cial intel­li­gence was avail­able 10 years ago. It put hun­dreds and hun­dreds and hun­dreds of dol­lars per leads. It was only done by the big boys. We’re basi­cal­ly sell­ing you leads for $1. For the infor­ma­tion that we give you, the price is unheard of, I would say. It’s very dif­fer­ent from the old lists you could pur­chase from India or oth­er coun­tries, where you got copi­ous amounts of leads with emails, but half of them were pure crap, to be hon­est with you. It did­n’t con­vert to per­son who did­n’t work any­more, etc. We’re sell­ing you infor­ma­tion which is guar­an­teed 100% no bounce back. They’re real emails. The per­son­’s real­ly work­ing there. All the infor­ma­tion is fresh. That’s why I’m say­ing that it would­n’t be too late to get into it lat­er, but with Fuse Machines I think this is a com­pa­ny which is real­ly going to be val­ued a bit high­er than we are right now.

 

Roy: When was the com­pa­ny found­ed?

 

Alex: It was found­ed in 2013. If you live in New York City, if you ever called 311 to com­plain about some­thing or if you ever called the bank or chat­ted with some­one in the bank, we did all the back­grounds, the soft­ware and com­put­ing in the back­ground. We devel­oped this ISR, inter­nal sales soft­ware, to sell the oth­er soft­ware. The oth­er one was a very long sales cycle. It was very expen­sive so we had to come up with some sort of pro­gram where we could guar­an­tee sales instead of wast­ing two years in the sales cycle. We need­ed to hone in on the right banks, the right prin­ci­pal­i­ties, etc. Then we fig­ured out that this soft­ware is actu­al­ly sal­able and peo­ple who real­ly need­ed it so com­plex that you get a lot of data out of it. We tweaked it a bit and now we sell it B to B, busi­ness to busi­ness. We have done some B to C, busi­ness to cus­tomer, but I think it’s most­ly busi­ness to busi­ness, to be hon­est with you.

 

Roy: You said it’s about a dol­lar a lead for some­one to buy it?

 

Alex: Yes.

 

Roy: Is there a month­ly fee?

 

Alex: It’s a month­ly fee. We do a do-it-your­self soft­ware ver­sion, where there is a $499 fee per month. We don’t ask for annu­al sub­scrip­tions. There’s a full ser­vice, lead gen­er­a­tion and con­sul­tan­cy where you’re the account man­ag­er and the whole she­bang for $1,000 (US). Then we can build pack­ages on top of that. Basi­cal­ly …

 

Roy: It’s $499 per month plus the cost of the lead?

 

Alex: No, its $499 then you do the lead gen your­self. You basi­cal­ly plug in the ICP, you help set­up the soft­ware to clear the leads, and you shoot out emails. You kind of do it your­self. Per­haps a bit more for small­er teams. As a $499 there’s a cap of 1,000 leads per month. But it’s $500.

 

Roy: What hap­pens if you go over 1,000 leads?

 

Alex: We would set a cap. If you go over 1,000 leads you basi­cal­ly pay 50 cents per lead or some­thing. It’s not the full price or you could just buy more licens­es. If you have a big team and you want to imple­ment this and you have five guys sell­ing, you could buy five licens­es. If you need 5,000 leads per month. Or you could go with the full ser­vice and we will do every­thing for you and you basi­cal­ly lean back and we will send you once a week any­thing from 150 to 250 leads and you do as you please with the list.

 

Roy: Your com­pa­ny was found­ed in 2013. Who was it found­ed by?

 

Alex: Our CEO is Sameer Maskey orig­i­nal­ly from Nepal, but he also teach­es at Colum­bia Uni­ver­si­ty. He teach­es machine learn­ing, arti­fi­cial intel­li­gence. He’s been in the coun­try for over 15 years.

 

Roy: What uni­ver­si­ty?

 

Alex: Colum­bia. Yeah, here in the city. He basi­cal­ly took 15 years of research from under grads, grads, PhD and kind of his pro­fes­so­r­i­al life and he want­ed to see if he could push sales and arti­fi­cial intel­li­gence togeth­er. If you look around it’s kind of the boy on sexy part of arti­fi­cial intel­li­gence, but it’s per­haps the more use­ful part. You can gath­er so much infor­ma­tion which is so use­ful for peo­ple.

 

Roy: Is there any­thing else you want to add? I think we’ve cov­ered every­thing. What else would you like? Any­thing you want to add about Fuse Machines?

 

Alex: I also do busi­ness devel­op­ment. I think that Fuse Machines is one of those com­pa­nies which is going to be some­thing to look out for. I think it’s for me, I would rec­om­mend that every­one take a look at our web­site which is Fusemachines.com. I think it’s going to be a very inter­est­ing tool. We’re also adding dif­fer­ent fea­tures. We’re adding futures where it has a com­po­nent where it drafts emails for you. Yours from your style. If you’re a good father or a good moth­er and you teach the machine, you will have a very good child then which could do much more than any oth­er soft­ware in the world can do.

 

Roy: So the soft­ware even learns your writ­ing style?

 

Alex: Yes.

 

Roy: It can write the emails for you?

 

Alex: Right. If you help it a bit in the begin­ning like a good par­ent, at the end in a cou­ple of months you can basi­cal­ly lean back and hit send.

 

Roy: Where are you guys based?

 

Alex: We’re based at 1 State Street in the city. It’s basi­cal­ly the last build­ing in Man­hat­tan. We have a beau­ti­ful view of Stat­en Island if that’s inter­est­ing.

 

Roy: How many employ­ees do you guys have?

 

Alex: We have 100+. The major­i­ty is in Kath­man­du in Nepal. We have about 50 com­put­er sci­en­tists, 20 peo­ple in New York City. We’re a grow­ing com­pa­ny. A new per­son start­ed today. We’re try­ing to do some­thing inter­est­ing with sales stuff and arti­fi­cial intel­li­gence.

 

Roy: Of the dif­fer­ent dis­ci­plines, what types of peo­ple are you hir­ing for?

 

Alex: Every­thing, to be hon­est with you. We’re grow­ing quick­ly. On the com­put­er side we’re hir­ing dif­fer­ent types of coders for the dif­fer­ent lan­guages, sales, cus­tomer ser­vice, mar­ket­ing … it’s basi­cal­ly an expand­ing team. It’s hard to know for the future what’s going to hap­pen, but I think when peo­ple start to see how ben­e­fi­cial this is going to be for them and how they can make mon­ey, because the con­ver­sion rate is bet­ter than most oth­er things I’ve seen, I think it’s going to be one of those com­pa­nies that you wish you knew about before it became famous.

 

 

Print Friendly, PDF & Email
Comments
One Response to “The Techcrunch Interviews: FuseMachines Artificial Intelligence Lead Machine”
Trackbacks
Check out what others are saying...
  1. […] we’re used to read­ing from sports writer sum­maries, using a tech­nol­o­gy known as NLG, or nat­ur­al lan­guage gen­er­a­tion. Not just sports games, though – any sto­ry that can be told from col­lect­ed […]



Leave A Comment

You must be logged in to post a comment.