How we can use AI in consulting delivery
AI is giving us credible use cases for how consultants can use it to reinvent their discovery process and their ability to extract deeper insights for their clients. In considerably less time. Working with qualitative data.
The Discovery phases of projects are by their nature, iterative.
As you work through each new document, workshop, interview etc., each new piece of data expands the information you need to act on. Your hypothesis and the assumptions which support it evolve. Some die on the vine, some need tweaking and others emerge.This is the good, healthy practice of real discovery work. An open sensing and exploration of what's going on. Rather than a validation exercise towards a pre-supposed outcome.
This rapid iteration can be quite simple with quantitative data. (Once you have it) a few pivot tables and charts later you can see if the data backs up a recommendation or point of view.
However, the reality is that the majority of discovery work relies on qualitative data.
As a quick level set. We view qualitative data as the non-numerical data used to understand the attitudes, beliefs and motivations of an individual or social group. E.g., How do they feel about business change? Why do they approach work this way? What if we pivoted into this market? etc.
This poses a few challenges.
First, despite the obsession with quant data and the effort to source it. Between timelines, availability of data, accessibility of data, and reliability/ quality of data... discovery tends to be 'quant data light'.
Second, Qualitative data is hard to work with... You have to make sense of it.
Making sense of it tends to rely on grouping it by similar words/ language or sentiment i.e., to find themes and/ or outliers in the data.
This becomes trickier as you add new information (especially when working in a team). The prospect of a new hypothesis to test often leaves you with two options. 1) Rely on a hunch, or 2) Trawl through data to reassess/ re-theme/ re-group.
#2 is an arduous task at the best of times. Even more so as you near the end of the project (i.e., when your late or rescheduled interviews happen!).
While none of us would like to admit it - the quality suffers.
There is also no simple salve here.
We can start with assuming the themes upfront so we can organize data around it as we go. But in my experience - this approach tends to wilt when it hits reality... It also bakes in bias (the kind of closed or narrow thinking which we're trying to dirsupt and change!).
Enter AI to redefine how we approach discovery work
AI is a little known umbrella of technology that you might not have heard of... if you don't use the internet.
While there is a lot of hype and unknowns, we are already seeing it augment several use cases. The high level steps of gather, analyse, and conclude remain. But the constraints of working with qualitative data are quickly falling away.
Take a recent Discy example:
A team needed to analyse a survey with hundreds of responses to 20+ questions.
Previously, this meant weeks of manual theming and close collaboration amongst the team. To ensure consistency and minimise the lost of context and insights as they worked.
With Discy, the team imported the survey responses and used AI to cluster similar themes. This gave them a valuable headstart. It helped them spot themes that sat across questions. Furthermore, they used natural language to query the dataset on specific insights.
“We have been able to conduct such a thorough analysis of this sort of survey, and we did it in half the time.”
So what?
The quality of a consultant is in their precision, specificity and ability to tackle the hard complex problems. Those often live deep in qualitative data.
These challenges deserve solutions specifically crafted for them... Rather than frankensteining a solution from general purpose tools.
Our goal with Discy is to help consultant's harness natural language AI to create deeper insight in less time... and we've only just started.