Lean tender offices, larger competitors: how to adopt AI

A small tender office may find itself competing for the same public contract as multinationals with dedicated functions for every stage: monitoring, analysis, pricing and contracts. This is common in the medical device sector, where SMEs account for around 90% of companies.

This gap has become more significant in recent years. Evaluating a tender now means navigating sustainability requirements that vary from one procedure to another, assessing quality criteria that affect scores and considering margins in light of Italy's payback mechanism. More analysis is needed before a company even decides whether to bid.

AI can expand a tender office's analytical capacity by reviewing documents, comparing requirements with the product portfolio and preparing an initial assessment. But there is no single way to adopt it. The right approach depends on the strengths an organisation already has. Here are three possible paths, each with a different starting point.

Even the smallest tender office has something no competitor can replicate: its knowledge of the company's product portfolio and past tenders. This is a natural starting point for using AI to answer a practical question: which tenders are worth pursuing?

Consider a tender specification with hundreds of technical requirements. AI reviews them, compares them with product data sheets and produces an initial compatibility assessment, lot by lot. Instead of working through the document from scratch, the office starts with an analysis it can check and refine.

The decision, however, remains with the company and also depends on factors the tender documents do not cover:

• Supply capacity: The company must be able to guarantee supply throughout the contract.
• Margins: In the medical device sector, these must also account for payback. If an Italian Region exceeds its spending cap, half of the excess is borne by suppliers, sometimes years later.
• Strategy: Defending an existing account or entering a new market may justify pursuing a less profitable tender.

Key principles

• Build on the knowledge the company already has
• Apply AI to tender selection
• Leave decisions to those who understand supply capacity, margins and strategy

In many organisations, a tender manager who oversees the process day to day takes the lead in adopting AI. They train the team using real cases, adapt the approach to different product lines and keep AI grounded in everyday work.

With each tender, AI reviews the documentation, distinguishes mandatory requirements from scored criteria and flags issues that need clarification from the contracting authority. This allows the tender manager to focus on checks and decisions that require experience.

The company remains responsible for the statements made in its bid and the decisions behind them:

• An eligibility requirement: The tender rules require a specific level of turnover in the relevant sector. Determining which revenues qualify is a decision for those who know the company.
• A technical specification: The product meets the requirement in a similar, but not identical, way. The company must decide whether to declare compliance, propose equivalence or not bid for the lot.
• A technical proposal evaluated on quality: Every claim about the product can affect the score and must be supported by evidence the company can provide.

Under Italy's Public Contracts Code (Article 98, Legislative Decree 36/2023), providing false or misleading information that could influence the contracting authority's decisions may constitute serious professional misconduct, even when it results from negligence alone. This responsibility cannot be delegated.

Key principles

• Give ownership of adoption to someone who knows tenders from the inside
• Train the team using real cases
• Keep responsibility for bid statements with the company

In Life Sciences, tender decisions are not made by a single department. Sales manages customer relationships, Pricing develops the financial offer, Regulatory verifies technical compliance and Legal reviews the contract. Some companies have already brought tender management and pricing together in one team, with a single manager overseeing the process through to the final pricing decision.

Where this coordination is already in place, AI can give all functions a common basis: an initial assessment of requirements, quality criteria and critical issues. The meeting to decide whether to bid can then work from a shared analysis, rather than separate reviews of the same documents.

Roles and responsibilities remain unchanged: go/no-go decisions and pricing often sit with Sales, while technical compliance stays with Regulatory. What changes is the quality of the discussion. Teams can focus on the decisions to be made, rather than spending time piecing information together.

Key principles

• Build on an established shared process
• Give everyone the same analytical foundation
• Keep each decision with the function accountable for it

The three paths build on different strengths, but have three elements in common: applying AI to live tenders from the outset, assigning clear ownership of adoption and turning early results into a repeatable approach for the entire team.

For a lean tender office, the question is not how to replicate the structure of larger competitors, but which existing strength to build on. The goal is to expand analytical capacity and focus human judgement where it matters: tender selection, statements made in bids and shared decisions.

FBK has worked with Life Sciences companies participating in public tenders for thirty years. Learn more