As of now, vacation has ended and I am back at work. I am starting a new role at work this year – more on that later. The last couple of weeks gave me some time to think about what is in store for our industry in 2018 . Despite my own misgivings on making predictions in general, I thought I will write these down any way in this blog. As always, these are strictly my personal musings .
1. Data becomes sexy, again, thanks to AI
Customers who have started on the AI journey all realize the same truth – this only works as good as the data that AI has access to. And most companies have less than stellar capabilities when it comes to data management. I totally expect 2018 to be the year of data ….again ! Of course tooling will change from the last time this “data is sexy” thing happened . Rejoice , my friends in data modeling, ETL and so on! 🙂
2. Data security and privacy becomes mainstream – thanks to GDPR and AI
All major companies always had to deal with security and privacy. Now with GDPR, this will become a mainstream topic both for SW and services – with cost and revenue impact . Its not just a back office problem like it was historically treated. Now front office functions need to be redesigned to make sure no regulations are broken. Europe started the trend, but obviously everyone else is going to have their version too soon. If history is any indication, we will end up with even more disparate rules and guidelines across the world. I have this feeling that most international tech companies will spend significantly in 2018 to lobby governments across the world.
GDPR is only one reason – the other is Artificial intelligence becoming a reality pretty quickly all around us. There is a lot of fear about privacy and security – some misguided and some very valid – and this will only amplify in 2018 and beyond.
I am tempted to say something about standards too – but the reality in this industry is that if there are two competing standards, people will come together to create a unifying standard, only to see that now we have three standards instead of two we started with. So – while much needed – I am not holding my breath 🙂
3. Chatbots will get a redo
Everyone seems to have a chat bot these days – but most are useless. I tested at least a dozen over the holidays as a consumer and it was a horrible experience. I think this will start seeing a big change in 2018. To begin with – I think more and more companies who jumped in and created the first generation of rules based chatbots will now start moving fully or partly to more of an AI driven chatbot. Instead of answering just short tail questions, I expect chatbots to answer more and more context sensitive long tail questions, and start to learn more from each interaction. This is another reason for data management to get a big boost. 2018 might also be the breakthrough year for voice to text capabilities – this is something close to my heart, given my thick Indian accent often confuses existing APIs.
4. AI will start democratizing visualization of data
I grew up in BI. From the time I started as a young BI consultant, I have believed that the best BI experts are more artists than engineers. It took me a long time to become a decent visualization guy. And having been in the field for a long time, I know I am in good company. We have more people who are experts in back end engineering than we have people who can make high impact visualizations. I don’t think the core principles like making data actionable, making sure it is context sensitive etc will ever change . Now the tooling has improved significantly and that is absolutely a good thing. Unfortunately, the complexity of the data (types of data, their interconnection, the speed of change of data etc) has also increased a lot and the challenge of visualizing has also increased a lot. I think this year we will start seeing the world of visualization start to rely more on cognitive technology and try to democratize data visualization for lesser mortals like me. I am not sure if this is a prediction or really a cry for help 🙂
5. Open source starts looking more like proprietary
Everything new eventually starts to look like its predecessors in our industry. It usually takes 15-20 years or more. I think Opensource software is now at a stage where no one has any sustained advantage, because there is hardly any barrier to entry for someone else. Also, every popular category – like databases for example – is way too fragmented. By becoming extremely developer focused, many new companies ignored ops tooling which adds to the customer head ache. At some point it becomes an untenable management overhead for customers to run a different software for every unique workload. I think this year we will see a change to this – OSS companies will probably start keeping more of their wares on commercial licenses , some larger companies will buy out a bunch of smaller public and startup companies and so on. I could be wrong on timing – maybe status quo will prevail another year or two, but I definitely think this will happen very soon.
6. World starts to come together for better/simpler debugging and monitoring
2016 and 2017 have made sure that containers and micro services are here to stay. Most new development will be cloud native in nature. While my purist friends still are waiting for one public cloud to rule them all, I am still on the commoner band wagon of hybrid cloud as the only pragmatic option. With every passing day, we will also create more and more sophisticated abstractions. All good things for the “happy flow”. But life in enterprise computing is rarely about happy flows – the effort to debug and monitor across all these layers has also become tedious. With all my previously stated misgivings on efforts to standardize – I do think we need thoughtful and simple open standards for debugging and monitoring in the increasingly distributed computing landscape. Given the momentum we are seeing, I am betting on this year forcing the community (perhaps led by APM gang) to come together and start putting the building blocks in place.
7. Lot of tech M&A in store, probably more startups will exit/ IPOs too.
Companies are sitting on a lot of cash already. On top of that, the GOP plan has a tax holiday for bringing money from abroad. After giving $1000 bonuses and increasing dividends, there will still be plenty of cash sitting around in big company bank accounts. Estimates of $1 to $3 trillion have made rounds on how much cash is stashed abroad by American companies. The sensible move is to use a good amount of this cash to start massive consolidation in the industry. This should happen across all segments – HW, SW, Cloud,…
A side effect of this is that startups should get a lift – either via IPO and/or by selling out to someone with deep pockets.
8. Devices/Things will become smarter and more secure
IOT , despite the hype, is a thing already. The fear of man made calamities like DDoS is also very real. And it is clear that leaving all decision making logic to cloud is not viable for more interesting use cases (say like a self driving car). I expect to see a lot more logic being executed inside the device itself , and a lot of hardware level security features added that cannot be changed via a software hack. None of this is new – just that I think 2018 will be the tipping point for this to become more mainstream. A good starting point in my opinion would be the routers on home networks – designed ground up with the idea of securing connected home devices on the network it controls.
9. More block chain branded companies this year
Last year we saw big data companies make the pivot to be machine learning companies. They did not want to be known as hadoop or ETL or NoSQL or anything remotely related to data, but over night change to Machine learning companies . Those that missed that round won’t probably bother with AI/ML anymore – I expect them to find a way to brand (hopefully also engineer something real – at least some of them) themselves as blockchain companies. Nothing wrong with this per se – no one is really fooled in this industry anymore with branding changes. There will be a temporary head ache for real blockchain companies to demystify stuff for their customers, on top of the topic of the day which is crypto currencies themselves.
10. ERP companies will yet again start to design their next generation products
ERP has evolved a LOT over the years, and mostly for the better. Just a few years ago, I thought their hardest challenges will be the move to cloud and improving usability (better UI, speed, simplicity etc). Those challenges have been addressed – admirably in general, compared to where they started. But I think even bigger challenges have now come up for this category.
ERP was fundamentally designed for efficiency and for human users. Now with AI allowing machines to learn and improve, the static nature of ERP is fast becoming a thing of the past. Small AI innovations have been started by pretty much every ERP vendor – but that is not even minimally indicative of much their world is going to get disrupted. The next generation needs AI at its core – it should be the center of continuous learning for every organization . It means not just efficiency is key, effectiveness becomes the new normal. On top of that – human users won’t be keying in much of the data any more. That work will be taken over more and more by machines . A lot of logic associated with screen flows in ERP today will be useless in that world. Even the current sophisticated interfaces built on ERP will be less efficient when it is always a machine that is going to talk to it in binary, or if its a human using voice or text strings . To some degree, I know the internal architecture of the main ERP systems in use today. Barring maybe one exception (not naming anyone given everyone is a friend) – I think rewriting most of their software from the ground up is probably the only way these existing systems will move into the future. If they don’t do it – I am reasonably sure that someone else will disrupt them from outside .