Exclusive: Machine learning won't kill consulting. It will accelerate it

In an age where artificial intelligence is developing at a rapid pace, capable of deep research on specific subjects while providing detailed analysis, traditional consulting may be on its way out, but a new wave is approaching.

According to Brendan Williams, Senior Director Analyst at Gartner, research shows that AI will accelerate the sales of consulting services. Although the AI boom is a hugely disruptive technology wave, it's not unlike previous technological booms, says Williams.

The cloud computing wave pushed firms to integrate their workflows into the cloud model, which opened the opportunity to enhance data analytics capabilities and improve internal collaboration tools. Williams says it opened up doors to additional consulting streams as well.

"It also added a whole lot of work in helping clients migrate to the cloud, helping them build things in the cloud, digital business applications and such, which wouldn't have been possible if they weren't on the cloud."

Similarly, AI will remove some current job positions, but will create new jobs, and migrate workers to departments that have a higher work volume as a result of the revolution.

"If you look at the day-to-day of an average mid-level consultant, a whole lot of what they do, it's rife with inefficiencies everywhere. Even if you're talking top-tier strategy consulting firms, they still have to do a lot of things like putting together PowerPoint decks."

Biren Agnihotri, Chief Technology Officer for EY Canada, says his firm has moved from a stage of generative to agentic AI - a solution that will automate workflows like scanning documents to reduce research work and scheduling tasks, but the main priority is using clean, original data at the core.

"Agentic AI is really helping us in preparing that initial hypothesis, and then consultants are now spending time on number one: validating the content, number two: improving the content, and number three: adding the value," says Agnihotri. "70 per cent of the time people spend just gathering the data, and 30 per cent of whatever is left is for the research, cleansing and analysis. So that is where the game shift is."

EY is not done implementing AI. After piloting CNNs, RNNs, LTSM, and LLMs, the next step is towards further SLM integration. 

Williams says research shows consulting firms are moving towards domain-specific language models for a more cost-efficient approach and going further in-depth to tasks demanded by a client, compared to an LLM.  

Agnihotri says EY is acting as "Client Zero" when integrating artificial learning models into its consulting workflows. This includes making sure proprietary data is in safe hands. 

"Whether it's document generation, creation, summarisation, analysis...we tested [it] on ourselves, and figured out how to find the ideas which can be impacted with AI. We are creating methodology frameworks, and we are taking it to our clients."

But while the world of AI is changing by the day, consultants must stay ahead of the curve, or face a disadvantage as many firms adapt, and some startups try to automate entry-level job tasks. 

Last month, Fundamental Research Labs unveiled Shortcut, an AI agent for Excel. It claims to outperform some first-year hires at McKinseley and Goldman in performing tasks like building financial models. This is one of the latest AI innovations that claims to perform autonomous tasks at the same level as some positions in the workforce. However, some consulting firms are keeping ahead of the curve, trying to adapt AI into an opportunity for, yes, more efficient work but also a new area of market domination.

Williams says it's pretty early days for how autonomous this technology can be, but as it increase's there are more risks involved. 

"If we want to see really big change as really big gains, people have to change what they do, right? You need to do new things and do existing things in different ways," says Williams. "You've got good old fashioned change management, you know, integrating new AI tools into existing systems, all the cybersecurity work involved in trying to make sure that you're not exposing yourself to new risks. "

While the government and policy sector is seeing the most significant ROI on AI adoption, partly due to Canada's commitment to tech acceleration with the Ministry of AI and Digital Development, precedents set with AI regulations and policy are a significant area of analysis for EY.

"As our clients started using risk models, using models for loans, mortgages, how are we, as an auditor, going to validate the outcomes of those models is a very huge responsibility, and there's a huge transformation," says Agnihotri." So responsible AI, whether it's ethical, whether it's efficacy, whether it's the confidence in these models is going to be very important."