Most companies do not fail at AI because the technology is bad. They fail because they point it at the wrong problem, underestimate the data and process work involved, or never get past the pilot. A widely cited 2025 MIT NANDA study, The GenAI Divide, estimated that about 95% of enterprise generative AI pilots were delivering no measurable impact on profit and loss. The tools are everywhere; the ability to turn them into results is not.
That gap is why the AI consultant has become one of the most influential roles in business technology. A good consultant does not sell AI. They find the specific processes where it pays, fix the data and workflow problems that AI exposes, handle the industry’s regulatory constraints, and get people to actually use what is built. How that plays out differs sharply by industry, and that is where most of the transformation is quietly happening.
What AI consultants actually do
Across industries, the work follows the same arc:
- Find high-return use cases. Rank opportunities by value, feasibility and risk, and steer away from impressive demos that do not move a business metric.
- Bridge business, data and engineering. Translate a business problem into technical requirements, and translate technical limits back into realistic expectations.
- Make it practical. Audit data quality, redesign workflows, select models and tools, and define how impact will be measured.
- Prevent waste. Kill weak ideas early, avoid duplicate tools, and keep costs visible before a pilot scales.
- Prepare people. Train staff, redesign roles and build the governance that lets adoption stick.
BCG has long argued that successful AI efforts are roughly 10% algorithms, 20% technology and data, and 70% people and processes. Consultants earn their keep mostly in that 70%.
How the work differs by industry
Healthcare
The biggest early wins are administrative: ambient documentation that drafts clinical notes from the patient conversation, prior-authorization support, coding assistance and patient messaging. Consultants spend much of their time on HIPAA-compliant architecture, business associate agreements with AI vendors, and clinician trust. A tool that saves a physician time but produces notes they must heavily rewrite will not be adopted.
Legal and professional services
Firms use AI for document review, contract analysis, research summaries and first drafts. The American Bar Association’s Formal Opinion 512, issued in July 2024, makes clear that lawyers’ duties of competence, confidentiality and reasonable fees apply to generative AI use. Consultants design workflows where AI output is always verified by a professional, client data stays in approved systems, and billing practices reflect the time actually saved.
Financial services and insurance
Use cases include fraud detection, customer service, underwriting support and compliance monitoring. Regulators expect explainability and fairness: the NAIC’s 2023 model bulletin on insurers’ use of AI, adopted by many states, requires a written AI governance program. Consultants help build model inventories, testing for bias and documentation that will stand up to an examiner.
Manufacturing
AI is applied to predictive maintenance, visual quality inspection, demand forecasting and production scheduling. The hard part is usually data: sensor feeds and machine logs spread across old systems. Consultants often spend the first phase building the data pipelines that make any model possible. Operational technology security matters too: connecting plant equipment to cloud AI services must not open a path for attackers into production systems, so network segmentation is part of the design from day one.
Field services and small businesses
For contractors, property managers and service firms, the fastest returns come from AI that answers calls, schedules jobs, processes invoices and checks job-site photos. Here the consultant’s job is to integrate with the CRM and accounting tools the business already uses, and to keep the setup simple enough that a small team can run it.
What a good AI consulting engagement looks like
A well-run engagement usually moves through clear stages:
- Discovery. Interview stakeholders, map key processes, and inventory data, systems and any AI already in use, including unsanctioned tools.
- Prioritization. Score candidate use cases on value, feasibility and risk and agree on one to three to pursue first.
- Readiness work. Fix data quality, access controls and security gaps the chosen use cases depend on.
- Pilot with metrics. Build a limited version, measure it against the current baseline, and decide to scale, adjust or stop.
- Scale and govern. Integrate into production systems, set usage policies, monitoring and ownership, and train staff.
- Hand over. Leave the internal team able to run and improve the system without permanent dependence on the consultant.
The MIT research also found that companies buying from specialized vendors and working with partners succeeded more often than those building everything internally, which matches what consultants see: the advantage is in focus and integration, not in reinventing the model.
How to choose an AI consultant
Look for industry experience, not just technical credentials. Ask for examples of projects that were stopped as well as those that succeeded; a consultant who has never recommended against an AI project is selling, not advising. Check that they address security and compliance from the start, measure results in business terms, and plan for knowledge transfer. For a deeper look at the role itself, see Why the Artificial Intelligence Consultant Is the Missing Piece in Your AI Strategy, and for keeping pilots on budget, AI Cost Overruns in 2026.
Frequently asked questions
When does a business need an AI consultant?
Typically when AI experiments are not turning into measurable results, when the organization is in a regulated industry, or when leadership needs an independent view of where AI will pay off before committing a significant budget.
Do AI consultants only work with large enterprises?
No. Small and mid-sized businesses often benefit most, because they lack in-house AI specialists and a focused engagement on one or two workflows can deliver results quickly.
How is AI consulting different from buying an AI product?
A product solves one problem in one way. A consultant helps decide which problems are worth solving, which products fit, how they connect to your systems and data, and how your people and processes need to change to get the value.
Work with an AI consultant who knows your industry
Delana Technologies provides AI consulting for businesses in healthcare, professional services, financial services and the trades, from use-case discovery to secure deployment and governance. Explore our AI consulting and agentic AI solutions and our compliance services, call 239.414.5126 or contact us.
Sources: MIT NANDA, “The GenAI Divide: State of AI in Business 2025” (August 2025) and Fortune coverage (August 18, 2025); American Bar Association Formal Opinion 512 on generative AI tools (July 2024); NAIC Model Bulletin on the Use of Artificial Intelligence Systems by Insurers (December 2023); BCG publications on AI transformation.