- What Does AI Mean For Consulting?
What Does AI Mean For Consulting?
This new entry in Kimble’s PS Insights series focuses on what AI will mean for the delivery of consulting projects. Reducing the human effort needed to deliver better results will create new opportunities to innovate, argues consulting entrepreneur Michael Taylor.
A co-founder of SchellingPoint, a provider of AI-enabled consulting software and training to internal and external consultants, Taylor explains he prefers the term “augmented intelligence” over artificial intelligence because he wants to place AI techniques firmly in a supporting role to human management consultants. He sees AI as enabling the “creativity, imagination and innovation” of human consultants and helping them to operate at their best every hour of the day.
Welcome to the PS Insights Podcast Series sponsored by Kimble Applications. Professional services organizations strive for efficiency, success and growth. This series is intended to provide key insights on how to achieve this from industry leaders.
Ian: Hello, my name is Ian Murphy and today I’m talking with Michael Taylor about the application of artificial intelligence to improve the efficiency and effectiveness of the analysis and solutions in consulting projects. Michael has worked in professional services for nearly thirty years. He first worked in AI at Anderson Consulting and was involved in their first software patent. Michael left to build a mid-sized professional services firm that was sold to Golden Gate Capital. He is now the co-founder of SchellingPoint who provides advanced management consulting skills and software to professional services and organizations.
Michael: Morning Ian, nice to be here.
Ian: Artificial intelligence, augmented intelligence. People are getting confused as to what is what. What’s the best definition that you can give us?
Michael: At SchellingPoint we don’t intend to put management consultants out of business, we’re not trying to replace the curiosity, the imagination, the innovation, the things that the human management consultant brings to the table. We define AI as augmented intelligence because as humans, despite how brilliant and smart we are, there are things that we don’t do as consistently as we’d like to. There are insights into data that we can’t draw, just following procedures and protocols in the reality of busy, hectic projects, things can get missed. For us augmented intelligence is really using technology and the AI concepts to enable a management consultant to, every minute of the day, be the best that they can be in terms of applying the best management consulting process, seeing the things that a client wants them to see.
Ian: When companies engage with you and start to talk about augmented intelligence, what does it mean to them in terms of their efficiency and the bottom line?
Michael: It’s a great question. If you think about management consulting projects, I’ve never spoken to a management consultant that says we spend six hours at the client on an eight hour day and then we go and swim in the hotel pool afterward. I’m more used to the client buys an eight hour day, my day ends up being ten hours or twelve hours, that’s part of the romance and reality of management consulting. We can do things using technology to bring the day back to a more workable day where the consultant can focus their capabilities and competence on the right things in that eight hours and so at a very tangible level, one of the things that we’re doing with AI is the automation of analysis. Think about the start of a project, quantitative data analysis, qualitative data analysis and we’re saying to the client this is what we have learned and this is where we should now go in the solutioning of your needs. What’s happened is that we can reduce, for example, twenty hours down to one hour. There are two benefits, one is the consultant was able to save or avoid many hours of work to come up with more accurate conclusions and the other side of it is what the client’s getting, the best thinking applied and a more accurate result out of the back end.
Ian: When you talk about that analysis of the information coming in, from a consulting perspective, what is augmented intelligence bringing here?
Michael: Now let’s say we’re dealing with a merger. We’re now in month one in the implementation of that merger and we have a post merger integration plan that we’re following, but we’ve just mushroomed the number of stakeholders involved in this post merger integration from those involved in the clean room and before the deal was signed to after the deal is announced we’re dealing with possibly hundreds of thousands of people that need to make it successful. One of the things that you’re doing as a management consultant in those first and very sensitive one, two, three months is trying to both deal with the hard tangible actions that you came up with, but you’re also trying to deal with all these stakeholders. There are two sides to your thinking; one is what we call the topic aspect. On average there are sixteen components, we call them themes that you’re trying to deal with on the actual topic, in other words we’re dealing with IT integration, we’re dealing with what products and services we take to market, we’re dealing with people’s roles and responsibilities. On the pure work side of it you’ve got all these different dimensions, then you have the people involved in those. Some are thinking we should have done this and others are thinking no we shouldn’t, there’s others that are thinking this is possible and others are saying no it’s not. As a human it’s beyond my limitations to be able to understand how do fifty senior stakeholders think about sixteen different themes of what’s going on. Now at a touch of a button we can now tell the consultant on those two dimensions, here’s where these groups do support this activity, here’s where these groups don’t agree with this activity. There’s a lot of intelligence now that we can provide the management consultant to enable them to effectively and efficiently execute the tasks that have to be conducted in a merger.
Ian: Companies face a challenge in identifying and delivering that gap between what we know and how we fix it. What is augmented intelligence providing when they start to evidence the final solution?
Michael: We’re dealing with a human, only has so much capacity to cope with information and data. When you get to the end of the project and look at the strategy, we should take this approach, we should go here, we should enter this market, we should leave this market, when you look at the action road map that says we should do these three things to get here, they didn’t just appear out of magic. You can trace them right back to the client stakeholders and the consultants’ opinions very early on in the project. The issue that we have in traditionally very human driven management consulting is that we can be very deep and work at a very detailed level in the analysis stage, we then draw conclusions and say well these are the four areas that we need to focus on and we then take those through into the solutioning phase. The problem with that is that we then lose the audit capability, we lose the integrity of those solutions that we come up with in terms of the ability to track them right back to the source, the customer wants more of this, the “we’re struggling with this, the competition’s getting better at this”. Mathematically we found that during those interviews and workshops at the beginning of a consulting project there’ll be fifty to ninety unique different opinions that are processed. Now let’s just take the number eighty, imagine there’s eighty that we’re working with. What we’ll do is we’ll draw conclusions from those eighty and then we’ll aggregate that up into, well there are four things we should be focusing on. What we’ve found is that in actual fact there’s not just eight, there’s twice as many. On average a group of people, consultants and clients will have about a hundred and sixty unique relevant opinions about a project. Now we’ve done that because we’re now using technology so we can cope with this larger volume. The difference now is rather than aggregating up and working at the higher level, we can take those a hundred and sixty and process those all the way through to the end strategy, the end road map, the set of actions. We can along the way synthesize up and say this is what they mean, so we can make that level of information meaningful and consumable to people. It means that when we get to the end strategy, the end road maps we’ve got a full audit trail all the way back to the source and the key thing is, it means that they don’t just look right, they are right. In an academic sense the term is technical integrity. The client might not see a difference in terms of what’s presented to them; it’s just in the traditional consulting world they trust that we’ve come out with the right answer. We can prove through the reasoning and data flow that we have come out with the right answer. When we go into implementation we’re going to get more of the results that were expected, we’re going to actually get the outcomes that were required.
Ian: If we remove those artificial assumptions that have therefore occurred, how does that enable you to do deeper analysis and more accurate comparative analysis between different portions of the project?
Michael: The main impact of applying AI and augmented intelligence to the consulting process is that the client is relying upon the management consultant to get them the most accurate prescription that is viable and endorsed by everybody. That’s what a client wants. Give me something that’s right, that is feasible and that people support. The reality is is that when Michael or Ian gets assigned to the project or Susan gets assigned to the project my reasoning processes, my experience on prior projects, my world view of things, that all influences the analysis that I conduct. When I take, then look at the data and make observations and draw conclusions and suggest actions to the client, they’re Michael’s interpretation and they will be different to Ian’s and Susan’s. Now the client doesn’t know the difference. The client has Michael and thinks Michael’s good and hopefully trusts what Michael comes up with. The reality behind the scenes though is what I would come up with is different to what Ian and Susan would come up with and that’s because Ian would take twenty hours to go through the data, I’m busy and I’ve got two other clients and I’d take two hours. Ian starts at the beginning of the data and passes through logically, I jump into the bit that I think is most interesting in the middle and run around in circles and then come out with my answers. Basically there are many things that go on in that very important stage called analyzing the data, drawing a conclusion and suggesting actions. One of the greatest impacts AI provides, at a touch of a button, complete coverage. It starts at the right point and takes the correct pathing logic. It’s using identified reasoning that says this data shows this here, this data shows this here because of those two observations we draw this conclusion and from this conclusion and that one we suggest this action. It’s now consistent every time and more accurate which is what the client’s paying for at the end of the day.
Ian: Is the model changing over time and how is that affected by industry or other factors?
Michael: Great question. One of the things that’s exciting about this work is that management consulting projects, they’re documented in the form of output, here’s the client’s strategy. There’s lots of notes taken during them, but we now have over two hundred and fifty management consulting projects in a database. We know the stakeholders that are involved, we know their titles, we have their roles, we have the opinions, we have the consultants that were involved, we have the consultants’ opinions, we have all those opinions across those sixteen various themes, that we have data in terms of the way people were talking about industry drivers, customer needs, barriers to success all the way through the process through to the road maps, through to the what was achieved on the road maps. By having everything in this data format we’re able to compare projects in strategy versus mergers versus outsourcing versus strategic alliances and joint ventures versus transitions to Cloud, going digital. We’re now able to look at different parts of the consulting process and the project and make comparisons. We’re also able to start to use this for predictive capabilities. What’s interesting is when we look at all of the projects in the strategy arena; one of the areas that we have found where a client team is going to struggle the most is in decision making and decision making roles and responsibilities. What we can do is proactively bring this up as a subject in the strategy project to say what’s going to be the decision making protocol and anchor that at the beginning of a project rather than deal with it later on when the client’s disagreeing over things.
Ian: If you’re able to do this prediction you have access to a much wider body of knowledge. How much is that reducing costs and improving efficiency for customers?
Michael: In terms of taking a strategy project, we worked out that we can use this advanced management consulting methodology using the AI and the automation and conduct a project in about one quarter of the hours. Rather than taking a team, I can now, in one quarter of the hours, provide a client with a more accurate, more viable and genuinely endorsed strategy and set of action plans. For the client we’ve found that the number of hours that they need to contribute is anything from a half to a third of typically what would be needed because the reality in projects is the clients are all excited when they sign up for them, but as soon as they sign up for them you can never find them half the time because they’re busy people. They understand the importance of this thing, but they’ve also got things going on. Those are the numbers we’re talking about, a firm that’s using advanced management consulting can now conduct a project in about a quarter of the hours requiring half to a third of the time of the client, massive efficiency gain.
Ian: Companies demand things faster and cheaper in a much more competitive business environment. This has to give them a significant commercial edge doesn’t it?
Michael: If you think about what’s gone on in recent years, the traditional six month strategy project that’s at the point these days or what I’m hearing is the consulting firms are being told give me something in six weeks. What that means is I have to work out how do I staff four, six, eight, ten people onto a project and start a project and deliver something that is accurate, viable and endorsed six weeks later. I don’t believe you can do that genuinely without automation and technology assistance. Basically, that’s throwing bodies at it and our experience in those projects, observing them and watching them is there are many assumptions that get made, there are many times when the consultant is saying well because of this and this we’ll assume this, we will recommend this. So basically what you’re doing you’re doing lots of short cutting. We can take a set of clients and now take a client through an efficient process where they have produced the result. For a consulting firm focused on billable hours and staffing people that can be terrifying. If your value proposition is about, well I’m selling you eight people at X dollars an hour, the notion of being able to do it in a quarter of the time, well that doesn’t sound good to me, my revenue just dropped. The implementation of AI into management consulting is, really it begs the question what’s the value price, how do I price the work based upon value. One consulting firm said well actually if I can get this job done in four weeks and I can do it with fewer people and I can take a half to a third of the client’s time maybe I should be charging more, but the point now is it changed the economics. For every project where the client uses a professional services organisation, a consulting firm, there’s twenty projects where they don’t and the reason they don’t is because it takes time to acquire the consultant, to get them in and running and there’s the cost that goes with it. Now I’ve been able to conduct these projects now in a matter of four weeks at a lower price. The client now can buy more; they can actually get the value of good management consulting, process and thinking applied to more of their projects which is better for them in the long run.
Ian: Thinking about what you said earlier, that means the evidence for those projects can be applied to a much wider group of projects. As a result, when a company does an analysis on what it’s working on it’s able to compare apples with apples, it’s not analysing two very different completely random sets of data. How is that helping them improve what they do?
Michael: Very good point Ian, so let me give you an example. There was a company where five projects were conducted in various parts of the business all using this methodology. In addition to bringing the value to those individual projects of using this approach, what happened then was because it’s all in a database and we have these notions of themes, we know roles is essentially a big data view now across these five projects. The consultant was able to sit down with the client, with executives and say look we’ve noticed over these five projects one of the things you’re struggling with consistently is authority levels, it’s a global business and one of the tensions that we see creeping in is a lack of clarity over roles and responsibilities, a corporate versus a regional level and that pattern has appeared in all five of these projects. That was one of three areas that could have synthesized out. There’s a USCPG company and they just completed twelve projects and what that’s enabled them to do is say this leadership behaviour, this is where it’s helping things and this leadership behaviour is where it is not helping things. They were able to take this very data evidenced based work, sit down with leaders and say could you change how you work in these areas because this is where it’s helping, this is where it’s hurting business relationships and initiatives.
Ian: You were talking about the aggregation of large numbers of points into just two or three actions. What we’ve now got is unexpected results because we’re seeing many more individual trails and sets of data than we started with, we’re not losing that critical data. We wouldn’t have seen these insights if we’d of stayed with the traditional approach.
Michael: That’s right because it’s all in a database and we can go and look at all the mergers in the database and ask questions. We can say in mergers what are we finding about this aspect. We can test hypotheses. We can now say well I have a hypothesis that this is what goes on in mergers and we can now go to the database and see to what extent that hypothesis is valid. The other thing is you always get these unusual things that appear out of it. I’ll give you one interesting example. When it comes to the tenure at companies, one of the things we can do is we can say these people have been at the company zero to one year, these have been one to three year, consistently in the projects the groups of the greatest tenure are the least aligned individuals. People find that counterintuitive, they think well the people that have been around fifteen, twenty years, they probably all think the same way or they probably all think that this is what the company’s good at, that this is what the company’s strongest at, this is the company’s needs. Very counterintuitively we’ve found the greatest misalignment around views like that is always amongst those with the greatest tenure which of course means that rather than thinking well we’ll look on the steering committee or on the project core team, as long as we have one of the people on it from the heavily tenured group we can assume that we’re getting the view of those people, this data has now shown that no that is not the case, you need to make sure you get different representation from the most tenured people because they’ll have highly divergent views of what the company’s good at, struggling at and needs.
Ian: If we think about most people’s perception of artificial or augmented intelligence, whichever phrase people chose to use it’s about reducing the number of people they employ, it’s about cost cutting, removing those elements that are the most expensive to the business, but is there not a risk here that with that turnover comes a lot of knowledge?
Michael: I’m going to go to a very, very tactical level that I’m a management consultant, I’m on a project, it’s a global organisation with a corporate function and regions and we’ve seen this happen where the management consultant would say people at corporate think this way and the people in the regions think this way and because people in corporate think this way and the regions think this way we’re going to suggest to the client this. We’ve now got the data that says here’s where corporate and the regions do differ, but here’s where the corporate and regions don’t. Again these have been some very counterintuitive situations where the consultant has gone I would never have thought that was the case. One is the data now sees that and then the AI, by the way our expert system is Joe, basically Joe says to the consultant this is where these two groups do agree with each other, when we now go to suggesting to the client here’s what you should talk about and here’s what you should now work on in the project that is now accommodated. It’s a very detailed level of insight, but it has a direct impact upon the eventual outcomes for the client.
Ian: What does this mean for companies going through an acquisition? If you’re able to capture all of this data does it change how acquisitions will be managed as we go forward?
Michael: It’s a very good question because my observation over the years and working on these many projects and surfacing out this work is that in general terms management consultants are best at is the content. We’re dealing with two banks and we’re going to merge two banks because the banks say we should do this with the payroll system, do this with the payment systems, this is what we should do with the loan origination process, so we’re very good on the content. What the client wants is the most accurate prescription that is viable and is endorsed by everybody because plans have no value, they’re just potential value, it all gets paid off in the implementation. The people in organisations that tend to understand the people side of things, these are the change managers, these are the OD experts. The nice thing now is we’re able to literally bring the two together. What AI is doing for a management consultant is enabling them to work on the domain space, okay here’s all the thoughts that need to happen around changing the load origination process and the merger, etcetera, etcetera, but at that same time integrated into that is okay now what I can see is that globally everybody supports doing that, but the IT group has strong concerns about this, I’ve now integrated into the domain space, I’ve now got the human dynamics which means I can now manage both together. That’s extremely powerful because every management consultant knows that what you’re doing all the time is trying to manage the client and work with the client to get them to own it and do it and it’s not just me walking in going you should do this.
Podcast: Thank you. You have been listening to one of a series of podcasts dedicated to sharing best practices for professional services organisations. These can be found on wwwkimbleapps.com