If your AI consultant ghosted you, stabilize the project first—then choose a partner built for continuity

Being ghosted creates urgency, but urgency is exactly what can lead a business into a second bad decision. Before interviewing another consultant, separate project recovery from partner selection. Recovery establishes what you own, what still works and what is exposed. Partner selection determines who should take responsibility for the next stage. Treating those as distinct decisions gives you leverage and prevents a new provider from defining the problem entirely on their own terms.

If the project contains half-finished code, use a recovery process designed for taking control of an abandoned AI build. If your main concern is finding an organization that will remain accountable through completion, examine how to choose a firm that will not disappear before signing another open-ended engagement.

A vanished consultant does not automatically mean the project should be discarded. Equally, money already spent is not a reason to preserve weak work. The right decision depends on evidence: code quality, security, data provenance, architecture, documentation, operating cost and alignment with the original business need. Ask the assessor to distinguish recoverable assets from sunk costs and unsupported assumptions.

Require the assessor to explain the trade-offs behind the recommendation. A useful assessment identifies dependencies, unresolved questions and the consequences of each option. It should also separate defects caused by unfinished work from flaws in the underlying idea. When an existing model has stopped functioning and its creator is unreachable, focus specifically on finding long-term support for a broken AI model rather than assuming that a one-time repair will create operational stability.

Trustworthiness is not established by a polished proposal or the phrase “AI expert.” It appears in how a provider handles uncertainty. A credible candidate asks about your business outcome, users, data, constraints and existing assets before prescribing a stack or promising a deadline. They explain what they need to inspect and what cannot be known until that inspection occurs.

Documentation deserves special attention because it is the bridge between today’s developer and tomorrow’s operator. If the abandoned project has scattered notes and undocumented decisions, use a selection process centered on robust project management and clear communication. For marketplace-based hiring, distinguish profile popularity from evidence of dependable delivery; this guide to finding vetted AI professionals after a bad experience addresses that decision directly.

Run the selection process as a small risk-management exercise. Give each candidate the same concise project brief and ask for the same outputs. That makes comparisons meaningful. The brief should state the business objective, current condition, known assets, access limitations and the immediate decision you need help making. Avoid disclosing sensitive credentials or unnecessary data during initial conversations.

Interview the delivery team whenever a firm proposes a team-based engagement. A recognizable leader or persuasive salesperson is not a substitute for meeting the people responsible for architecture, implementation and day-to-day communication. If you previously hired on the strength of an online profile, compare that experience with a process for choosing a partner who completes the work. If payment has already been made against incomplete code, use milestone controls for an abandoned paid project to structure the replacement engagement.

A good contract cannot rescue a poor working relationship, but a vague engagement makes failure harder to detect and more expensive to correct. Structure the work so that knowledge and value accumulate in your control. Tie payments to observable deliverables rather than elapsed time alone, and make each milestone small enough that your team can evaluate it before the project moves substantially further.

Data projects require an additional safeguard: distinguish raw data preparation from the production of reliable business answers. If a data scientist left before delivering usable analysis, seek a provider with a clear method for turning abandoned raw data into decision-ready insights. Ask how the team will establish data meaning, quality and provenance before presenting conclusions.

Rohan Hall has worked in technology since 1984, with professional experience across the United States, Europe and Asia. His background includes software development, systems and databases, enterprise implementation work, AI and blockchain systems, technology strategy, architecture and leadership of a distributed global engineering team. Explore Rohan Hall’s work, ventures and publications for the broader context behind that experience.

Rohan founded OceSha Ventures and its AI-first solutions. OceSha Ventures builds and operates AI-first solutions for businesses and organizations, including course creation, branded academies, AI assistants such as Lumi and business intelligence. Lumi supports conversational websites, customer inquiry handling and customer intelligence using business information that the organization has reviewed and approved.

This background is relevant because selecting a durable AI resource is not only a coding decision. It requires architecture, operating judgment, governance and a clear view of how emerging technologies fit the business. Rohan’s work on responsible and trustworthy technology emphasizes evidence, transparency, governance, assurance, standards and trust in AI and blockchain adoption.

For a deeper treatment of AI alongside blockchain, neuromorphic technologies and other major developments, Rohan covers the subject in The Convergence of AI and the Top 10 Emerging Technologies. The book is the appropriate next resource for readers examining how these technologies converge beyond the immediate challenge of recovering one project.