Using new techs: let’s focus on the needs first!


Companies from all sectors are challenged by new technologies (artificial intelligence, machine learning, cognitive agents, chatbots, natural language processing, etc.): how can they use them to generate more value? However, the starting point before implementing, is not the technology itself, but the operational and tactical need. Key findings from a chatbot experience.

Vision from the sky: the golden age of innovation

Creativity is not the finding of a thing, but the making something out of it after it is found. James Russell Lowell

Today for all companies from all sectors, from the new or the old economy, the challenge is: what can we do with all these new “things”? How can we generate more value using machine learning, cognitive agents, chatbots, natural language processing…? Considering the large set of technological choices available today, or soon to be available (or maybe never available), the selection of the right technical solution to solve a problem is not simple. However, seen positively, we are luckier today than Henry Ford’s customers in 1920s, who could have their car painted any colour they wanted as long as it was black. Nevertheless, in our experience of delivering successful innovative projects, the starting point is not the technology but the problem to solve.

Back to earth

When we met the head of the legal department of one of our clients, he questioned us as to whether a chatbot would be an appropriate solution to address his need, which was to provide his business teams with a legal diagnostic of their operations:

  • “Why do you think of chatbot?” I asked
  • “Because it’s so simple, people will just ask questions to the machine and get the right answer immediately”

To address this need, we met the business team that was the potential user of the solution. We asked them how they would like to interact with the machine, and the conclusion of the meeting was:

  • “We do not want to dialogue with a machine and answer 15 questions before getting the answer. With a smart interface we could easily select the characteristics of an operation then press a button and get the relevant legal diagnostic. Simple and efficient.”

We took into account the feedback from the business team. In only a few months, we built a solution with a smart interface designed with the users that was linked to a knowledge base combining a natural language generation engine. Today, this innovative solution addresses the business teams’ needs and gives the legal department the capacity to maintain the knowledge base by itself.

The chatbot can be a very useful technology in many cases, but in that case it was not adequate, simply because the user experience was not appropriate for these users. However, we could have built a great chatbot from a functional point of view that would have had a one-way ticket to the cemetery of unsuccessful innovations.

This feedback reminds us that the starting point when solving a problem is a comprehensive  understanding of that problem. The second step is the selection of the ways the problem can be solved.

Operational feedback #1: focus on the pain point

Operational and urgent

For any project, be it innovative or not, the focus has to be on the need, not on the technological means. The starting point is: what is my pain point today in terms of productivity, cost, revenue or compliance? A successful innovative project will begin with the qualification of a real need of doers, at operational and tactical level.

For instance, in a French depositary bank, in 2017, 20 young graduates used to crunch data every day in order to write monthly regulatory risk reports for their clients. The problem was that data crunching was not a satisfactory job for the young graduates. In addition, this situation generated a high and costly staff turnover. The bank needed to shift these highly educated and qualified people to the middle and front offices to generate more value for its clients. This problem had to be solved immediately: in a few months we built, tested and deployed with the client a solution. Based upon artificial intelligence technologies, the solution applies the analysts’ reasoning, and drafts risk analyses in several languages. Most of the team moved up to the middle and front offices.

Measurable with a relevant scope

We expect a solution to deliver measurable results. Therefore, needs must be measurable too (revenue, satisfaction rate, cost, efficiency…). To ensure that the innovative project will produce measurable results, it is advisable to begin with a first version of a business case. Due to the sophistication of matrix organisations in big companies, the scope of operational needs is not always limited to one single operational level (business unit or department).  When pulling at a string, one sometimes discovers that it is the tail of an elephant. In that case, launching an innovative project to solve the problem could be a risky enterprise. To be successful, our recommendation is to limit the initial scope of an innovative project: do not bite off more than you can chew. When a first step is achieved successfully and has produced measurable results within a relevant scope, then a next step can add more complexity.

Operational feedback #2: qualify technologies

Once a business need is qualified, it’s time to focus on the solution and select the right technology. It’s a tricky challenge, because the right technology has to be identified among an enormous quantity of marketing messages, white papers, gurus’ points of view, scientific popularisation articles…

Qualifying an innovative technology should be based upon the same criteria as for any other one:

  • Alignment on the business needs
  • Alignment on the existing IT architecture
  • Availability of the functionalities
  • Compliance with the standards
  • Identified constraints and prerequisites
  • Cost of acquisition consistent with the business case

Operational feedback #3: action!

Once the business need is qualified and the relevant technology is selected, it’s time for action! The project itself has to demonstrate the major benefits of innovation: time and money savings. Innovative projects must be restricted in terms of schedule and budget. A small team, combining complementary areas of expertise, using agile methodology, under the supervision of a committed and empowered sponsor, is the best way to deliver rapidly visible results: there is nothing new under the sun.


An innovative project is not a research and development adventure. The needs must be well qualified before the projects starts, and not discovered during the project. Technologies are just great means that have to be qualified not for themselves, but for their ability to resolve problems at lower costs.

We believe this pragmatic and operational approach, in the field, may feed and enhance a digital strategy as Guru visions, marketing messages or newspaper articles do in a more conceptual framework.


Serge Baudin, CEO and co-founder of Addventa. Serge started his career in large software vendors, system integrators and consulting firms (IBM, Oracle, Capgemini and PwC), where he held business development and advisory positions notably in the technological innovation area. In 2014, he co-founded Addventa, dedicated to the development of solutions using the power of artificial intelligence for large companies and administrations. Addventa was awarded the Pôle Finance Innovation label in 2017. Its clients include leading European global banking or insurance groups and the French Tax authorities.

 Press release 17 October 2018: Societe Generale Securities Services and the fintech Addventa have signed an exclusive partnership

Did you like this article ?