We are pret­ty posi­ti­ve about Natu­ral Lan­guage Gene­ra­ti­on (NLG). It would be a sur­pri­se, frank­ly, if we were not. We honest­ly belie­ve that NLG solu­ti­ons can be vital to ans­we­ring ques­ti­ons facing com­pa­nies and orga­ni­sa­ti­ons today.

The­re are nume­rous bene­fits to com­pa­nies that deci­de to imple­ment NLG pro­ducts as a sup­port to their ever­y­day ope­ra­ti­ons. In e-com­mer­ce, an online retailer can take a data­set and use it to pro­du­ce an unli­mi­ted num­ber of pro­duct descrip­ti­ons, hund­reds per second, that can be updated auto­ma­ti­cal­ly. Tho­se texts pro­du­ced with one data­set can be trans­for­med auto­ma­ti­cal­ly by adding ano­t­her data­set, saving the time, effort, and resour­ces nee­ded to make such chan­ges manu­al­ly.

In the news­room, brea­king deve­lop­ments can be repor­ted imme­dia­te­ly while flesh-and-blood repor­ters race to gather facts, con­duct inter­views, and give events their pro­per con­text. The­re are nume­rous, real pos­si­bi­li­ties, many of which we have alre­ady been—or are—working on. Think about all the strai­ght­for­ward texts that you read in your day-to-day life that could be auto­ma­ted from a data­set: wea­ther reports, real esta­te lis­tings, finan­ci­al reports, school reports, medi­cal reports, etc… If some­thing can be mea­su­red, it can be ren­de­red as data. And if it is ren­de­red as data, then it can be auto­ma­ted.

That is not to say that the sai­ling has always been smooth. The­re are issu­es and chal­len­ges that com­pa­nies and orga­ni­sa­ti­ons find when they try to imple­ment an NLG solu­ti­on.

I will address that in a second. Howe­ver, at this point, it is inst­ruc­tive to take a step back and out­line the NLG pro­ducts that Ret­res­co offers.

The first is known as tex­ten­gi­ne pro, which is Retresco’s mana­ged-ser­vice pro­duct. Custo­mers opting for this solu­ti­on work with Retresco’s team of data archi­tects, pro­gramm­ers, deve­l­o­pers, and lin­gu­is­tic data archi­tects to deve­lop and deli­ver bes­po­ke, tailo­red solu­ti­ons.

The second solu­ti­on is textengine.io, which comes out of beta this week. This revo­lu­tio­na­ry self-ser­vice pro­duct puts the power and exper­ti­se to crea­te NLG con­tent into the hands of our cli­ents.

Both ser­vices are built using the rtr tex­ten­gi­ne, one of the world’s most-power­ful pie­ces of NLG soft­ware. The rtr tex­ten­gi­ne is built to gene­ra­te text in up to five lan­guages (English, Ger­man, French, Dutch, and Ita­li­an).

Textengine.io cur­r­ent­ly sup­ports NLG in English and Ger­man. We will add Dutch soon and aim, over the com­ing mon­ths, to expand the offe­ring to inclu­de Ita­li­an, Spa­nish, and French.

We deve­lo­ped textengine.io for a num­ber of rea­sons, but main­ly to faci­li­ta­te orga­ni­sa­ti­ons that want total con­trol in deve­lo­ping an NLG solu­ti­on. This is becau­se we have noti­ced issu­es with the imple­men­ta­ti­on of NLG that we think can be easi­ly over­co­me with textengine.io.

First­ly, many orga­ni­sa­ti­ons find that the texts they wish to auto­ma­te are often intri­ca­te and com­pli­ca­ted. A wea­ther pro­ject, for examp­le, reli­es on mas­ses of com­pli­ca­ted data, inclu­ding tem­pe­ra­tu­re, humi­di­ty, time, air pres­su­re, pre­ci­pi­ta­ti­on, cloud cover, time of day, and dura­ti­on. Tho­se things all work in com­bi­na­ti­on, and any chan­ge in one neces­si­ta­tes a who­le new set of text, plus its many varia­ti­ons. The same goes for finan­ci­al ser­vices, whe­re data such as EBIT, EBITDA, mul­ti­ple rates of taxa­ti­on, pro­fit, and loss often has to be cal­cu­la­ted and put into con­text. While any robust sys­tem desi­gned to do such work can fire out hund­reds of texts per second, it is important from the out­set that they are set up with the know­ledge of how to inter­pret this data cor­rect­ly. Even though our experts are experts at what they do, they are not experts on every busi­ness sec­tor. So, many com­pa­nies may feel more com­for­ta­ble if they have more con­trol of the NLG pro­ject.

This moves us to the second issue that com­pa­nies have in imple­men­ting an NLG solu­ti­on. Occa­sio­nal­ly, with tex­ten­gi­ne pro, some com­pa­nies find them­sel­ves off-balan­ce as they feel that they have lost ‘con­trol’ of their pro­ject as it is put toge­ther intern­al­ly by Ret­res­co staff. It is a dis­con­cer­ting fee­ling for many when they see their pro­ject come along in spa­tes and spats, and not having over­all manage­ment of its deve­lop­ment. With textengine.io, howe­ver, com­pa­nies are able to deve­lop their own solu­ti­ons in-hou­se.

Ret­res­co is proud of textengine.io. It is a pro­duct that we have been deve­lo­ping for over two years, and is built on the ground­brea­king work that Ret­res­co has done in Natu­ral Lan­guage Pro­ces­sing (NLP) sin­ce the company’s incep­ti­on in 2008. It comes out of beta with a host of fea­tures. The­se inclu­de:

  1. An intui­ti­ve inter­face that makes crea­ting your own NLG pro­jects strai­ght­for­ward and quick.
  2. Auto­ma­tic text adap­tati­on through intel­li­gent lin­gu­is­tic ana­ly­sis takes the data pro­vi­ded and adjusts your text auto­ma­ti­cal­ly.
  3. Adap­ta­ble and easi­ly-chan­ge­ab­le pro­ject manage­ment mea­ning that you can con­ti­nu­al­ly hone your NLG work, main­tai­ning con­trol at every step of the pro­cess.
  4. Com­pre­hen­si­ve onboar­ding and sup­port from Flo­ri­an, our custo­mer suc­cess mana­ger, to gua­ran­tee a suc­cess­ful pro­ject.

We built textengine.io to address the needs of custo­mers loo­king to imple­ment NLG solu­ti­ons at a rea­son­ab­le cost and with litt­le fric­tion. It is an aim that we think we have suc­cee­ded at. We hope you think so, too.

 

 
For more infor­ma­ti­on, plea­se con­tact:

Pete Car­vill (@pete_carvill)
Com­mu­ni­ca­ti­ons Mana­ger
+49 (0)30 555 781 999
peter.carvill@retresco.de

 

 

About Ret­res­co

Foun­ded in Ber­lin in 2008, Ret­res­co has beco­me one of the lea­ding com­pa­nies in the field of natu­ral lan­guage pro­ces­sing (NLP) and machi­ne learning. Ret­res­co deve­lops seman­tic app­li­ca­ti­ons in the are­as of con­tent clas­si­fi­ca­ti­on, recom­men­da­ti­on, as well as high­ly inno­va­ti­ve tech­no­lo­gy for natu­ral lan­guage gene­ra­ti­on (NLG). Through near­ly a deca­de of deep indus­try expe­ri­ence, Ret­res­co helps its cli­ents acce­le­ra­te digi­tal trans­for­ma­ti­on, increa­se ope­ra­tio­nal effi­ci­en­ci­es, and enhan­ce custo­mer enga­ge­ment.