Microsoft's QnA Maker service enables you to generate a chatbot from an existing set of questions and answers and deploy it to the Microsoft Azure cloud rapidly, inexpensively and with no coding required. Organisations looking to offer self-service access to an existing knowledge base could find QnA Maker a quick and cost-effective way of doing so.
The questions an answers can come from a set of web based FAQs, or a document (Word, Excel, PDF and others are supported).
You can access the QnA Maker service here: https://qnamaker.ai and this tutorial will get you started if you'd like to try it out for yourself.
I generated a bot using a few questions and answers about my employer (Objectivity's) bot development services, you can interact with the bot below (type "help" to see a list of some of the questions the bot has been trained to answer.
** Update - the bot is currently offline, it'll be back soon! **
Monday, 27 November 2017
Wednesday, 22 February 2017
The contextLike many people in business I'm exercised by the innovation agenda, that desire to improve the way we work so that we can deliver value and make our lives better.
When you're working for a technology company (in my case Objectivity Limited) innovation takes on another dimension. It's not just something tech companies do as a value add, it's at the heart of our very existence - helping customers to exploit new opportunities and maintain competitive advantage through the appropriate application of ever evolving technologies.
In order to serve our customers changing needs we also need to innovate internally, seeking out new methods, technologies and tools, so we can also maintain a competitive advantage among our peers.
The technology landscape is changing ever more rapidly, and this is driving a change in customers' expectations. Factors including the widespread deployment of inexpensive cloud based services, the ubiquity of mobile platforms, and an explosion in the volume of tools and libraries available to developers mean it's possible to build and deploy software systems in record time.
In addition, technology vendors are increasingly marketing their products directly to business users, creating a demand that the IT department is then under pressure to satisfy.
These pressures push a specific constraint increasingly to the fore - that any new development should deliver rapid time to value.
|Applied Innovation Model - Justin O'Dwyer (Creative Commons Attribution license)|
Spurred on by insights gleaned from Eric Ries' seminal book The Lean Startup, I produced a model to help me visualise how and where innovation should be conducted if we're to deliver against this rapid time to value constraint.
The model helped me to formulate a practical approach innovation, and to understand the difference between innovation and research.
It helps me further differentiate if I refer to the process of conducting research with a demonstrable short term payoff (i.e. rapid time to value) as "Applied Innovation".
As the model evolved, this focus on value driven innovation became the core concept, which is why I now refer to it as the Applied Innovation Model. I'm sharing the model in the hope that it's somehow useful to you too.
As an aside, I recently came across this excellent, scholarly article by Chahab Nastar, which contains a particularly elegant quote: “Research is using money to create ideas. Innovation is using ideas to create money”. Couldn't have put it better myself - I really wish I'd written that!
The model features 4 roles and more than one of them (or potentially all four) can be performed by a single individual. The roles are intended to capture specific behaviours intrinsic to the innovation process, and they break down as follows:
- Scientist - has the initial idea for an innovation, either through discovery as the result of research activity or through a spark of inspiration.
- Engineer - brings the idea to life, building products (whether they be algorithms, physical devices, or something else of value) for use further down the value chain.
- Technologist - integrates existing products to address specific challenges, assembling new systems from existing components that may have originally had some other purpose.
- End User - operates the systems they're provided with to execute a specific function that has value to the business.
You'll see that the model assumes a relationship between innovation type and time to value, in that theoretical innovation (i.e. research) has a greater time to value than applied innovation (i.e. building things in conjunction with short customer feedback loops). There may be domains where this does not hold true, and I'd be especially interested in hearing thoughts on this.
The final noteworthy features are the "footprints" shown on the model. Thinking about the specific activity you're conducting, the reason for performing the activity, what the outputs will be and who is involved in the activity, yields a footprint. The taller the footprint, the greater the time to value.
The model shows a definite "sweet spot". Innovation performed close to the customer, where new products are provisioned and used directly (think SaaS offerings such as Salesforce), or combined with others to provide enhanced capabilities (mashups for example) provide shortest time to value. Proof of concept work, such as hackathons and spike solutions can also fall into this footprint.
Note that we're talking about time to value here. Magnitude of value still depends on the relative robustness of the business case!
I hope you've found this article of interest, the model is shared under the Creative Commons Attribution license, so please feel free to reuse and adapt it if it's of value. Please also feel free to share your innovation insights - I'd love to hear your views on this fascinating subject.