Ian Gotts is the founder and CEO at Elements.cloud.
Much of what a consultant delivers during the planning phases of the implementation life cycle is the documentation: specifications, business requirements, process maps, architecture diagrams, solution designs, user stories and training materials.
Virtually every step here can be massively accelerated by GPT (generative pre-trained transformers). By the way, GPT is more than ChatGPT, which is often seen as a “search on steroids.” GPT is amazing at solving puzzles if you give it all the puzzle pieces. So, if you give it a dataset or document and explicit analysis instructions, it will deliver great results. The power is in the data and the structure of the prompts.
Using GPT in this way, we are seeing massive productivity improvements. These benefits are too large to ignore. But we are still seeing consulting firms hesitant to get started, citing security and LLM hallucinations as reasons to limit or even ban GPT access for their consultants. This is a mistake.
Those consultants who lean in and operationalize GPT across their practice will quickly dominate. We are seeing a GPT divide emerging: those who have embraced AI and operationalized it inside their firms and those still looking on from the sidelines.
GPT is accelerating in terms of its capabilities, and those consultants who don’t start quickly will get left behind. Now is not the time to wait and see. Now is the time to experiment and refine.
Even before we consider GPT, consultants are gaining a competitive advantage with change intelligence. Change intelligence is a centralized platform that supports the creation and analysis of all the business analysis content during the planning phase of every implementation project. One aspect of change intelligence is its ability to provide insights into the complexity of the systems that the consultants are bidding to enhance. Being able to quickly understand the risks and complexity associated with systems means that they can then write a statement of work that is both compelling to the customer and can be delivered profitably.
We are already seeing examples where consulting firms can bid with a higher day rate and more man-days than their competitors but still win. This is because they can justify the scope of work backed up with defensible data analysis. Their competitors, armed with very little information, will give a ballpark quote aiming to undercut the other bidders. The client knows that if they choose a firm that has submitted a low cost, they will be hit with a series of change requests when the consulting firm understands the true nature of the work they’re expected to deliver. And this will only emerge after several weeks of analysis once the project has started. And then it is too late to select another firm.
Often, a client submits a description of the business process as they want to change in their RFP. GPT can now read that document and produce a process map optimized by its knowledge from the LLM. This is a first cut and can then be reviewed and updated by the consultants. But a lot of heavy lifting has already been done. There are a number of benefits associated with this approach.
• It reduces the time taken to understand the proposal. At this point, the consulting firm may decide not to bid due to the risk or the complexity of the project, or the unrealistic expectations of the client
• The process map will help define the work more accurately and can be used to have initial discussions or raise questions with the client on the scope of the project. These discussions will immediately give the client more confidence in the consulting firm’s ability to deliver and build a strong relationship. This increases the chance of being selected, even if they are not the lowest-cost bid.
• GPT could even take that process content and start writing sample user stories with acceptance criteria to demonstrate the level of understanding the consulting firm has in the client’s business.
Consulting firms cannot be generalists: “We implement Salesforce” or “We deliver business transformation.” Any client, armed with GPT, can find generic information about implementing a particular system. The top consultants need to be seen as trusted advisers with deep industry knowledge and insights.
This is another area where GPT can create a differentiator for consulting firms. The industry subject matter experts inside the consulting firm can be interviewed. These transcripts can be combined with other industry knowledge articles and blogs. They can then be given to GPT, which can build out industry accelerators as business process models. These will give the client confidence in the consulting firm’s ability to deliver results quickly.
Once the project has been won, GPT can accelerate much of the planning and analysis activities. Those business process maps generated for the bid can be refined and taken into more detail through discovery calls with the clients. The process maps can be drawn from those transcripts. GPT can even take the transcripts, identify the data model and relate it to the system configuration.
This may seem daunting, so here are some best practices to get started.
• Start using ChatGPT in some narrow internal use cases to improve your prompting and understand what is possible. For example, start by summarizing client project status reporting.
• Start operationalizing those prompts. Save them. Refine them. Track their usage.
• Look at every aspect of your delivery processes for opportunities to leverage GPT. If you haven’t agreed upon and documented processes, this is the perfect catalyst to run some mapping workshops. I highly recommend the UPN process notation.
In summary, now is a very exciting time to be a consultant, provided you are on the correct side of the GPT divide. Those who take the time now to understand how to exploit GPT and infuse it into every aspect of their firms’ market and delivery methodology will emerge as the clear winners.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?