The finance director is asked to approve the automation of accounts payable, the introduction of a customer chatbot and the distribution of an AI assistant to each employee. Each of these projects promises cost savings, but no one can tell which should be backed up first. The role of AI strategy consulting services should be to create a connection between business priorities and evidence on the feasibility, costs and readiness of the organization, and to create the real conditions for investment.
Begin with the decision the company cannot make
An assignment must have a question that can be researched. ‘How should we use AI?’ suggests expansive solutions. ‘What administrative process could be improved with existing systems and budget?’ will produce a more practical task.
Consultants should establish who is to act on their research and what information they need. A department head, for example, might need proof that employees will be able to make better use of their time. A finance chief would require the operating costs, implementation assumptions, and a credible prediction of the timing of the expected benefits.
This agreement gives direction to the research and prevents good ideas from being consumed by ideas which are irrelevant to it.
Use frameworks to expose assumptions
An operational process map describes a process from start to finish, showing the information required, the transitions, the decisions, and the exceptions. It shows how in the case of the invoice audit the delays are due more to the lack of references to purchase orders than to a reading of the document.
Opportunities are compared in a system with consistent criteria. Useful dimensions include workload, avoidable effort, availability of data, the degree of integration, and the potential for error. Every mark should be explained; otherwise the table is only concealing opinions in numbers.
A dependency map shows what needs to be completed before a project can begin; a pilot for example would be connected to the approval of the documents, the approval for access, and the assignment of the operational manager. This is how we avoid an imposing deadline concealing unfulfilled prerequisites.
Expect deliverables that support action
The last parcel was to be enough to start a working party without any further discovery.
An opportunity report explains the chosen problem, the people affected, the current situation, the expected improvement and any alternatives considered and the reasons for rejecting them.
The readiness report lists all the obstacles and appoints who will be responsible for overcoming them. It is not enough to say that the data needs improvement. It has to specify which records are affected, why they matter and by whom they can be corrected.
A Plan of Trials defines the intended experiment, its representative inputs, its means of operation, the manner of evaluating the result, the expenditure limits, and the termination decision.
A financial model separates the figures that have been taken as true from those that are only estimated. It includes the costs of implementation, running costs, human resource expenditure, and maintenance. The work program is a compilation of dependencies and budgetary decisions.
Look for investigation between workshops
Meetings can bring about views, but recommendations require facts. They require the investigation of the samples of records, the observation of the work of the employees, the verification of the statements on the use of time.
Management might estimate that the reviewing of invoices takes ten minutes. Observation may show that the reading takes two minutes and resolving the errors another eight, which has a significant effect on both the proposed solution and the estimated savings.
A strong group is one that revises its position as evidence alters, and one that makes it evident when there is any disagreement, so that the leader can take the initiative in settling the argument.
Require an honest stopping point
An undertaking should explain when the company ought to stop. The lack of ownership, the inaccessibility of the records, the negligible extent of the expected savings, could all provide grounds for postponement.
The final discussion must determine the next decision, who will make it and what evidence is necessary to make it. Some organizations proceed directly to a test flight, others to modifying a prerequisite.
It is the test of the process if the company can allocate resources to a new set of goals with a clearer picture in mind. Management must understand the value at stake, employees must understand which work steps are involved, and the project teams must know what they must demonstrate to win more money.
This content is provided for informational purposes only and is not a substitute for professional advice. AFP editorial staff were not involved in the creation of this content.
Share
