How we “married” a cloud-based PBX, GSM and realtors (part 2)

    Continuation of the hub published on January 30th.

    Summary of the previous series: Russia, Moscow, 2014, the telephone provider decides to launch a cloud-based IP-PBX , integrate it with the modified FMC platform of one of the mobile operators and charge all this in commerce, so that the service is useful, customers are happy , and the bosses are satisfied.

    We resolved the mobile component of the service - SIM cards began to make calls as we wanted, and not as intended by the developers of the FMC platform: we see all the calls, we receive everything that is dialed on the mobile phones (we accept from 2 to 20 characters), we manage calls online, turn it on and off, and even created an interface, read the “Client’s Personal Account”, in which you can assign short numbers to SIM cards, block and unblock traffic and, in general, bring hitherto unprecedented beauty in communications.

    Here's a personal account:


    Now the client himself decides what short numbers he will write on the SIM cards of the secretary Natasha and the storekeeper Lena. But this is not a ready-made solution - it is necessary that the SIM cards do not just ring, but ring as it is now accepted - with IVR, with recording of conversations, with statistics and other hunt groups.

    So, you need to make friends with SIM cards with a cloud-based VoIP platform, so that GSM and VoIP merge in communication ecstasy and a new, super-convergent service appears that customers manage just like they used to manage regular virtual PBXs.

    For several years of hard work in the voice services market, our company has simply grown up with a variety of platforms - from Asterisk to tricky paid solutions of foreign vendors, and each platform has spent a lot of effort, time, sweat and blood from product experts, techies and sellers, but nothing sane -custom-oriented was never created, since each such platform solved highly specialized tasks or served a variety of callbacks, and the interfaces were drawn more for themselves, and not for sale. And then it became clear that five years were wasted, the market ran far ahead, winking colorful lights of animations and flat design, and we will have to do another “tricky crap”, write scripts, throw out the unnecessary, add the right one, draw the interfaces and get rid of hereditary skeuomorphism like a class.

    For several days spent in meditation and meetings, it was decided to buy or rent something commercial and beautiful, from those times. complete support, with polite and loving managers on the other side of the helpdesk, with the right functionality and the guarantee that the developer will not fall in love with us right after the wedding. It was at that moment that we recalled the cunning-eyed representatives of one platform-building company, who friendlyly waved flags at the last SvyazExpoCom, offering to test their platform for free and did not ask for anything in return.

    We decided to twist the paid happiness a bit and requested test access.

    Here's what you got:


    Honestly, despite the good functionality of the switch itself, the user interface was a little depressing and strongly resembled Rambler's mail during the heyday of the dial-up. We wanted to shock and delight, about which it was immediately written off where it should be. The received answer was intriguing - "expect, what you want is about to come to light, but for now, test and don’t think about anything bad." Yes, we don’t even think about the bad, we want to look modern and not spend much on painting the clouds. Ok, continue to test.

    A couple of weeks of testing revealed a critical bug - the core of the system didn’t want to understand the non-standard identifiers of the SIM cards that we sent to the platform every time we called - problems arose when trying to present the SIM card as an ordinary client sip device. The server did not understand where to call him. Correspondence with the vendor lasted several weeks. By regular means, the problem could not be fixed and a decision was made to crawl into the code. The rule of thumb: since the vendor has agreed to go into the code, why not be shy - ask for more and get how much is due. What was done: the inflated TK flew to the mail to fellow developers. A couple of weeks passed and a miracle happened: a couple of treasured buttons appeared in the admin panel of the platform, which made our service very similar to what we originally intended.

    Here they are, these buttons:


    Now we can work with FMC (SIM identifiers are processed correctly, we can enable or disable FMC support for those who need it), and we also get the opportunity to integrate with third-party CRMs. The evening ceases to be languid: FMC plus VoIP plus CRM integration - these are three in one, it’s convenient, this is necessary for the market.

    By that time, the promised new design arrived in time - the result of sleepless nights by marketers and interest writers. At our request, redundant functionality was removed (who need multilevel IVRs, complex schedules and three-stage redirects), and the vacant space was used to improve the usability of the interface. Ideology - the simpler, the better, and if someone wants "goodies" we will either dissuade or give access individually. In most cases, the client’s initial wishlist and objective reality are two non-correlating entities. We want a lot of buttons, but in fact we use two or none at all. I have never seen a seller who enthusiastically set up call forwarding according to five tree-like rules, not forgetting to include colleagues in the hunt group. Willingly believe that

    Now we began to look like this:


    What happened in the end? The result was a cloud platform , with screwed-in FMC SIM cards and good functionality. The SIM card appears to the system as an ordinary sip phone and can be included in any PBX scenario — assign it as an internal line, record conversations, see statistics, when calling from a SIM card, substitute a common company number and, in general, do with a SIM card what is usually done with iron cards telephones. Our corporate vulture has grown legs.

    Now you need to understand the needs of potential customers, finish the functionality, integrate with several CRMs and can be sold. Moreover, the cherished telephone number of the director of one of the real estate agencies has been an eyesore for several months, and the director just asked for a sip with legs, CRM and telephony in one bottle. So we will experiment together.

    Wait for us, Director. We have almost assembled the stand, we will soon call and ask for TK.

    To be continued…

    Also popular now: