There’s a good chance the answer that an employee needs to do their job effectively is available within the company, but it can often be difficult to surface.
Research suggests that around three hours per day are spent searching for information. Even more frustratingly, around 44 percent of all searches end in failure. In total, around $1 million per month is lost for every 1,000 employees in a business.
Starmind believes that it doesn’t have to be this way and is using AI to power a customisable real-time knowledge network that gives teams on-demand access to the answers they need.
AI News caught up with Ronan Kirby, Chief Customer Officer at Starmind, to learn more about the solution.
AI News: Around 74 percent of decision-makers believe that data is stuck in organisational silos. How is Starmind helping to surface inaccessible knowledge?
Ronan Kirby: Starmind’s superpower is breaking down both traditional silos and knowledge silos and it does that using an AI engine that is able to learn ubiquitously across a corporate environment who is an expert in what.
It doesn’t rely on human networks of knowing someone who knows something to make a suggestion of a person who may be a knowledge-holder; it’s indiscriminate in that regard, and that’s sort of its superpower.
Regardless of what use case, or what application or problem you would use it to solve, it knows who knows what in a company and it continues to constantly refresh so that it’s always learning on the go.
It unlearns as well. You could be the expert on video editing today within your organisation, but that could be that may no longer be true in a few weeks’ time. Part of the key is making sure that you’re presenting current and accurate information to people to break down those silos.
Regardless of geography, regardless of team, regardless of language, regardless of business unit … it’s able to see, and learn, and therefore make suggestions and share knowledge regardless of what silo somebody may exist in.
AN: How do you ensure that potentially sensitive company data that not all team members should maybe have access to is kept safe?
RK: The AI engine has to be told where it can learn from and that is set up very intentionally and deliberately by the customer.
You might say you can learn from your Jira or your Salesforce , or your team’s environment or your Slack environment. You give it very strict parameters as to where it is allowed to learn from. That way, you don’t have it potentially learning and offering up things that it shouldn’t.
From the expert identification side, it’s not storing what people are saying or anything like that – all it’s doing is learning and building a profile about the people. A company’s data is never held, it’s just simply saying, “Ryan has a large association with these particular things”. It’s building a knowledge profile on you as opposed to capturing the information that it used to build that knowledge profile.
Editor’s note: Starmind’s solution is fully GDPR-compliant and, as part of the onboarding process, the knowledge platform is tailored for the client to ensure only the appropriate access to information is granted per employee or team.
AN: After a reasonable time of integrating Starmind, what is the success rate of receiving an answer that someone needs?
RK: In a help desk environment, where maybe people are using it to resolve things and self-service rather than open a ticket with a help desk, 98 percent of questions are typically answered within 60 minutes.
The answers are then stored and every time one of those is reused, that’s a human that didn’t have to be bothered with a question.
That’s where the scalability comes in. For your first question, we spoke about the identification of the expert knowledge-holder. This is how we then make it scalable and repeatable by storing an answer that was given and allowing other people to search and reference that.
PepsiCo uses [Starmind] in their R&D environments. They’ve got 23 R&D teams across the world for 23 multi-billion dollar brands that they own. In that scenario, someone might be researching how to do a sugar-free drink or treat and wants to find out how that impacts the makeup of that particular snack when it’s exposed to an enzyme. Somebody in another R&D team somewhere has had that experience and is able to shortcut that for them. That’s incredibly impactful.
Within PepsiCo, 96 percent of questions have received a successful answer.
AN: Do you think AIs should always be designed to complement humans rather than replace them?
RK: We talk about our AI as being “human-centric” and that’s the key to it. We’re using AI to help do something about people that people themselves couldn’t feasibly do but ultimately we’re still bringing you back to that personal knowledge as opposed to trying to replace that knowledge with an AI.
AI is nowhere near that level of capability anyway, but where we can use it to bring a scale and efficiency that humans can’t do – to unlock what only humans can do – is incredibly important. So that’s our human-centric approach to AI.
AN: Amidst global economic uncertainties, have you noticed an uptick in enterprises seeking ways to maximise the productivity of their workforce by improving these operational efficiencies?
RK: Absolutely. We recently refreshed what we call our Future of Work survey, so we spoke to 1,400 executives across the US, UK, and Germany from companies with over 10,000 knowledge workers. 68 percent of the executives that we spoke to rated the impact on their operational efficiency as being the single largest concern for them at the moment. What they are looking for us to do is to help drive operational efficiency.
If you think about go-to-market teams, how do you improve the conversion rates of your team? Well, one of the ways to do that is by exposing the best practices and experiences of other parts of your sales organisation to other parts. Breaking down the knowledge silos helps them to do that.
Going back to the example of anything with ticketing – typically, every time a human engages with a ticket there’s a $15 cost for an enterprise. If you can help an enterprise avoid the opening of that ticket, there’s a massive efficiency gain there
Even a relatively small reduction of six percent in ticket volume has a massive financial impact on a business in terms of tickets avoided, but it also has an impact on the time that would have otherwise been wasted by an employee searching for something, opening a ticket, and another human having to look at it. As you and I know, inevitably the ticket is answered incorrectly and they’ve told you something that you weren’t asking in the first place and you’ve got to reopen your ticket.
Those kind of operational efficiency gains as you stare down the barrel of a global recession is really important. That’s front-and-centre of a lot of the conversations we’re happening now as we look at the macroeconomic environment.
AN: Along similar lines to the global economic uncertainties question, has more remote working post-pandemic increased the need for knowledge-sharing solutions like Starmind?
RK: People networks have traditionally been so important within an organisation – people knowing, “Oh, it’s Andra sitting over there that I go to for that”.
What we see from our data is that people rarely know who knows what outside of their direct team. We can actually map this within companies because we can see the connections between people with the AI engine and we see that evolve over time.
Once you’ve offered up an expert to someone, once somebody has identified – oh, it’s Ryan who knows that thing because Starmind helped me identify Ryan previously – now we have that connection and that builds, and builds, and builds. It’s phenomenal to see that mapped over time.
So yeah, remote working has increased the need for real-time information exchange.
AN: How important is it for business leaders to look into knowledge management solutions?
RK: 68 percent of C-level executives identified knowledge silos as the most important or impactful thing within their businesses. That’s compared to 61 percent who were worried about an economic downturn. The highest impact, or the highest risk that they perceived, was knowledge silos.
I think that underlines it right there and that’s just looking at knowledge workers. Amongst the executives themselves, when we look through our data, the cost of that to businesses is even higher.
We look at an annual bill of inefficient knowledge-sharing of $71 million for the typical company with 10,000 knowledge workers or more. Even if you’re – let’s pick a big company that isn’t our customer – you’re a Johnson & Johnson with a mega turnover, $71 million is still not a rounding error. That number scales up as your employees scale up.
I think executives know that and they’re working very hard to mitigate that in a way that’s effective. Part of the challenge is that traditional knowledge management solutions are not effective, they’re about documenting information.
Once stuff is documented, it’s basically rotten straight away. You can do it with things that are static and don’t evolve over time, but you can’t exhaustively capture what an information holder knows. There lies the gap that’s impacting businesses that we’re bridging for people.
Ronan Kirby and the Starmind team will be sharing their invaluable insights at this year’s Digital Transformation Week, which is co-located with the AI & Big Data Expo. You can find out more about Ronan’s session here. Swing by Starmind’s booth at stand #466.