Finding relevant AI tools for practical use in management consulting

After 15 years in consulting, Joel Osborne and Tim Hawke both stepped into the product and SaaS world two years ago – drawing on their experience in both large and small firms to build more relevant tools. Consultancy UK asked the founders of Discy.ai to share their view on the reality of AI adoption in the industry.

This article will help address three core themes. The industry’s challenge with the AI tools available to it; a reframe on how to match consulting's requirements to what's available; and a plausible future for consulting-relevant tools.

We looked at 'relevant tools' from the lens of recurring 'jobs' consultants do in project delivery. Considering how and when you might delegate work to AI tools. We categorise tools as generic, task, domain or custom. Today, the majority are generic and only support parts of a step across several jobs. Others are task oriented and can support a whole job step. Both are immensely helpful - but none yet transform an entire job.

We believe this requires a tool specialised for consulting, or one built yourself. Our solution aims to be this for qualitative research and analysis. Freeing consultants from important (but tedious, manual work), to focus their creative energies into building deeper insights.

Consulting's tool dilemma

Management consultants know the value of a powerful tool to help with their work. From Lotus 1 2 3, to MS Office, to Google's innovation of making office apps easy to work with online.

So, when ChatGPT launched last November, consultants held their breath. Not since digital whiteboards, had there been a worthy addition to their 'tool-stack'. Here was something seismic, and with it, a wave of Gen-AI tools promising the world.

Except, when asked to do the kind of qualitative reasoning that we as consultants do, they couldn’t. As a result, the entire market saw a trend of people trying it, and giving up soon after – because consultants had a more nuanced challenge. There were only a few tools built for the reality of how consultants work. This hadn't changed.

This is summed up by the experiences of Matt Cheung, CEO Clarasys Business Consultancy, who explained, “We’ve run a few AI experiments this year, looking at improving knowledge base retrieval, drawing out insights from project data and creating a chat interface for our internal policies. Most have reached the same conclusion. We can reduce admin. But AI has to be trained to understand the consulting context if it is going to drive up the quality of delivery.”

Consultants do lots of small tasks across a lot of 'jobs', in a non-linear way. Jobs such as; manage the project, research & analyse, communicate & present, create outputs. Their output is often dictated by the client, as can be the tools they use (given security/ data controls). Given any one tool may only support some of the work, the effort to onboard/ set-up a new tool often doesn't stack up.

If you ignore the word ‘AI’ for now. This is a story of how an old profession adopts new tools. You have motivation (hype, excitement and fear of missing out/ losing relevance). But you also have confusion with the options and fear of getting it wrong.

Reframing how you look at AI tools

Over the last 12 months we've worked with several consultancies to apply AI on live projects. In deciding relevant AI solutions, we find it helps to frame AI tools as generic, task, specific and custom. Illustrated with a reference to lunch!

The supermarket
You’ve got five minutes for lunch, you aren’t fussy, it just needs to get you through. You stop at a supermarket. This is like Generic AI. You can get a bit of everything, but it won’t rock your world.

The sandwich shop
You’ve got a bit longer, so you want something more to your needs. Sandwich shops are designed for lunch, and are a bit nicer – but you’re still limited in options. This is Task AI. Designed for a specific task, but not tailored to your specific situational needs.

The restaurant
For something truly delicious, you might go to a restaurant for the cuisine you like. They specialise in the food and keep perfecting it. This is Consulting AI: built and trained for consulting. But the market is still relatively empty, so your options have more in common with the fine-dining scene in Hull than in central London.

The DIY option
Now, if you want your signature dish, your famous Spaghetti Bolognese, nobody else can do that. It is specific to you, and so you have to make it yourself, from the ingredients available to you. This is Custom AI. If you want AI to do what you do, based on your 15 years (or more) of experience in your specialism, you will probably need to train it yourself.

Being consultants, we put this in a 2x2 chart, to share our experience of using AI tools across the spectrum.

Like most, we find chatbots like ChatGPT or Claude helpful to research a topic/idea. To clarify our thoughts, or give ideas for rewording text. Other tools we use include meeting transcription, slide editing and image generation. In order, these help us ensure we don't miss key points, can speed up proposal design, or find images to convey an idea.

We used a task-specific AI to create a learning course with three levels. An output to support a profound process change. Under real conditions of time pressure, cognitive switching and a need to discuss/ present. Following good practice in prompt engineering produced a good level 1 course. It extracted the right data in a way that saved a day. Level 2 + 3 were a bust. It couldn't grasp the domain to generate a usable output and a feedback loop was absent.

Ultimately, choosing AI tools is like delegating work. Ask yourself…

• Do they have the expertise to do this task?
• Will their output be accurate/ quality enough?

While not ‘transformative’, both generic and task-oriented tools have useful potential across your ‘jobs’.

A plausible future, for consulting relevant tools

Where AI tools become transformative is when they support entire consulting jobs & workflows.

As Umbar Shakir, Digital, AI & Experiences Lead Partner, Gate One, noted, “For me it’s about redressing the balance of graft and craft and where value truly lies in consultancy. We have to graft (research, analyse, collate, summarise) in order to drive insights. To then craft solutions that are relevant, differentiated and creative... I know where I’d rather spend my brain power and time! So, tools like GPTs, and Discy.ai help to take on the graft so I can spend more time, energy and creativity on my craft…"

For large firms, transformative has become 'build your own ChatGPT'. Makes sense, their global workforce adds reams of new data to decades of past data - daily. Giving teams a quick way to extract insights from the firm’s vast moat of knowledge feels transformative.

But that doesn’t mean it's transformative for all consultancies. This will depend on your USP.

At Discy, we see 'transformative' as giving consultants a better way to draw more insights out of their qualitative data. Every project involves some discovery and diagnosis. Tagging what’s relevant across the team’s interviews, surveys, ethnography and research. This is important but tedious manual work with the risk of late nights or missed insights.

Discy uses consulting trained AI to streamline this. Giving teams more space and possibilities to explore the data. A foundation to build ‘so-what’ insights, and spend more time advising clients. Discy fixes workflows by matching the right type of AI use case to the process.

We see a trend of more workflow specific products. Lowering the adoption barrier to 'AI-augmented consulting'.

As Charles Vivian, Consultancy Advisor, Former CEO North Highland UK, put it, “AI tools have the potential to speed up and improve the quality and accuracy of the work that consultants do. Those who embrace tools like Discy and integrate them into their approaches have the opportunity to gain competitive advantage.”           

Discy.ai is a secure qualitative insights solution, built for management consultants. The company looks to reinvent the way consultants, and other business change agents build strategic insights; providing an AI-powered SaaS platform to efficiently analyse qualitative data. This gives teams of transformation experts a better solution to harness and apply their collective wisdom, enabling them to deliver greater client impact, faster.