Slalom

slalom.com

Slalom combines business advisory with technology strategy, implementation, engineering, and selected managed services. Buyers should distinguish these engagement types: an AI readiness assessment or technology roadmap is not the same purchase as a production application or ongoing operations. Its published delivery examples establish work beyond advisory, while its general business and customer strategy services remain distinct from technology consulting. ^[1] ^[2] ^[3] ^[4]

Services and delivery

AI readiness, strategy, and governance advisory

Advisory. Assesses AI readiness, prioritizes use cases, and defines strategy and governance for generative and agentic AI. This is an advisory engagement rather than a commitment to build or operate the resulting solutions. ^[3]

Generative and agentic AI implementation

Implementation. Designs and builds assistants, copilots, and agent-driven workflows, integrating AI into business processes. The offering includes implementation rather than research or software licensing alone. ^[3]

Agentic workflow managed services

Managed services. Operates, monitors, and improves agentic workflows in production. Buyers should contract separately for operating responsibilities, human oversight, performance objectives, and escalation arrangements. ^[3]

Cloud migration and application modernization

Implementation. Moves workloads to the cloud and updates applications, with infrastructure design and automation supporting the transition. The service evidence does not establish a standard ongoing support commitment. ^[5]

Application development and systems integration

Implementation. Designs and builds customer-facing applications and business applications for employees and partners. Integrates applications, data, and other systems rather than providing software resale alone. ^[6]

Digital product and cloud platform engineering

Implementation. Engineers new products and develops mature products at scale, including cloud platforms and machine-learning-enabled product features. The service also describes upskilling customer teams. ^[7]

Data engineering and embedded analytics

Implementation. Develops data architectures and engineering solutions using business-specific design principles, including embedded analytics. This entry does not treat ordinary data engineering as an AI or machine-learning service. ^[8]

Security and privacy advisory

Advisory. Provides security gap assessment, privacy governance, and risk and compliance alignment through its security adoption framework. This advisory scope should be distinguished from implementation support. ^[9]

Security implementation support

Implementation. Supports new security implementations. The documented scope establishes implementation assistance, but not specific penetration-testing or security operations center services. ^[9]

Technology and data strategy advisory

Advisory. Develops technology and data strategies with executable roadmaps. Implementation must be scoped separately rather than assumed to be included in a roadmap engagement. ^[2]

Business and customer strategy

Non-technical service. Provides general business and customer strategy advisory alongside its technology strategy practice. These offerings are not classified here as technology implementation or engineering. ^[2]

Company details

Technology and platforms

Documented AI delivery includes Amazon Bedrock, Anthropic Claude, retrieval-augmented generation, REST API integration, and Terraform in the Jaja Finance engagement. United Airlines’ AWS-based environment exposed models including Amazon Titan, Anthropic Claude, and Meta Llama. These are project-specific examples, not evidence that every engagement uses the same stack. ^[10] ^[4]

Slalom describes Traverse as the operating platform supporting Agentic Managed Services, connecting agents, human experts, and customer systems. The announcement says it can run in Slalom’s cloud or the customer’s environment; it does not establish standalone licensing rights. ^[11]

Track record

For Jaja Finance, Slalom describes progressing from an LLM and semantic-search feasibility assessment to the production customer-service assistant Airi. Deliverables included API integration, testing tools, and reusable deployment scripts. Slalom reports 90% lower customer response times and 30–40% more time for service teams to handle complex queries. These are firm-reported results for that engagement. ^[10]

For United Airlines, Slalom describes building an MLOps platform for creating, deploying, and managing models, then extending it with a secure generative AI experimentation environment. Released applications included flight-status messaging assistance and a customer-care email copilot. Human reviewers retained responsibility for AI-generated flight messages. ^[4]

In Sionic’s fraud-detection collaboration, Slalom documented use-case prioritization, synthetic-data requirements, model guardrails, and a rules-engine contribution. The story assigns technical application and synthetic transaction generation to Google Cloud engineers. It therefore does not establish that Slalom independently built or operated the entire fraud platform. ^[12]

Engagement and delivery

Slalom lists AI services under a Texas DIR contract covering areas including machine learning, natural language processing, computer vision, robotic process automation, and digital assistants. The contract page explicitly describes services only, excluding hardware and software product sales through that vehicle. Buyers must verify eligibility and applicable procurement terms. ^[13]

Scope and limitations

The supplied research establishes commercial scope, not independent validation of delivery quality. Case-study results are Slalom’s own reports.
Pricing, minimum engagement size, staffing commitments, acceptance criteria, and customer-specific support obligations were not established.
The reviewed security evidence does not establish specific penetration-testing or security operations center offerings; those scopes remain unverified.
Traverse is documented as supporting a managed service. Standalone pricing, licensing, and independent deployment rights were not established.

Questions to ask

Which deliverables will Slalom own, and which remain with platform partners or our internal team?
Does the proposed AI scope cover advisory, prototyping, production implementation, ongoing operations, or separate phases?
Which model evaluations, human-review controls, acceptance tests, and monitoring are included?
Who owns application code, deployment scripts, prompts, and operating documentation?
For managed services, what service objectives, escalation terms, exit rights, and platform dependencies apply?
Can Slalom provide references matching our industry, technology stack, and proposed delivery responsibilities?

Overview