Your organization is misled by a vendor's AI promises. How do you navigate through this misleading situation? Here are 5 actions to take: 1. Assess the business and financial impact. Does this have any impact on your customers or suppliers or only affect your internal operation? 2. Gather evidence to prove your hypothesis. This includes investigating among those who were involved, collecting data points and getting legal counsels. 3. Engage with vendor for an honest dialogue and understand both sides of the story. 4. Take the responsibility for what you can do better and hold your vendor accountable for what they missed. 5. Come up with a mitigation plan and do it right.

• Gather evidence • Consult experts • Internal alignment • Open dialogue with the vendor • Review legal options • Negotiate remediation • Consider exit strategies • Learn and improve • Communicate transparently • Plan for continuity

In the era of deep fakes, Seeing is NOT Believing! This is the harsh reality of our times, particularly in over-hyped areas, such as AI. Every application and service seems to be AI powered these days, and each one promises the moon! The answer lies in good old wisdom, but with an augmented dose of caution and vigilance: - Try Before You Buy...don't rely on recorded demos (not even from Google!) - Clausify...convert the promise to a clause in the contract, and make eye contact while making this request (in this case, seeing is believing) - Peerless solution...ask peers in similar roles - Look under the hood...get hold of some techie, or ask them to explain in business terms, 'how' they will deliver. If they reply in jargon, you can run!

After due assessment, enter open dialogue with an open mind to metaphorically peel the onion. Consider that all AI solutions need data which vendors won’t have direct access to. Thereafter :- 1) Take care of the data availability and access 2) Be firm in asking for prototypes or working models of past implementations to understand vendors true capabilities Building trust and keeping promises is the cornerstone of long term relationship with the vendor. Of course legal recourse is available, time lost for your critical projects could be irreparable. Hence I suggest gaining full understanding before the legal path.

When AI vendor promises seem unrealistic, first form a team of in-house AI experts to evaluate the claims. This team should rigorously review product specifications, run tests, and compare actual functionality against initial promises, documenting every detail. If your team lacks sufficient expertise, bring in an external AI consultant for an objective analysis. Thorough assessment and documentation are key to discerning the truth and making informed decisions, ensuring that the technology truly meets your organization’s needs and expectations.

Your approach is no different than ANY vendor not delivering on their promises. I've seen this a few times and with various resolutions. Once it was delayed payment until the vendor was able to provide the needed functionality, which they eventually did. Another time the vendor was motivated to implement a (challenging) workaround but publicity was offered (a conference or webinar speaking spot I believe) to incentivize vendor further. Unfortunately in one instance when the vendor was stubborn and inflexible, we just worked our tails off and migrated to competitor before the renewal. They didn't expect us to and apparently the end of our recurring revenue really hurt them.

🚀 Open dialogue with your vendor is a game-changer! Lay out your findings with a mix of clarity and candor. Imagine you're uncovering a treasure map—you're not just pointing out the X marks, but also discussing the paths that got you there. Ask them to clarify any murky waters or technical jargon. This isn't just a meeting; it's a litmus test for their transparency and dedication. Are they open to addressing issues? Ready to adapt? This chat will reveal if you're steering towards smooth seas or need to course-correct. Keep the conversation dynamic, and remember, your goal is not just to fix today’s problem but to build trust for tomorrow’s AI adventures. 🌟 Sources: BotOracle, Experience, TechCommunity

The vendor boasted about their AI's ability to analyze customer sentiment with 95% accuracy. Your assessment reveals it's closer to 60%. In your conversation, say, "We appreciate your enthusiasm for the product, but our independent assessment shows the accuracy rate is significantly lower than advertised. Can you provide insight into this discrepancy?" This serves several purposes: The vendor may have misunderstood your needs, or there might be technical nuances they can explain. The vendor's response will reveal their integrity and customer focus. Are they defensive? Or do they actively seek solutions? You may need to renegotiate, adjust expectations, or explore other options. Stay focused on facts and solutions, not accusations.

It is important to start an open conversation with the vendor after the assessment has been done. This allows you to present your findings as well as professionally express your concerns hence allowing them to elaborate on any differences between what you heard about AI technology through the grapevine as opposed to how it works on the ground.

This really depends how deep you are in. And if you are in a situation where large portions of the organization is now dependent on this, you have to sit down with the vendor and have a straight conversation about need and solution. And really be open to the idea that it may not be the tech, but rather how you are using the tech. AI is not a miracle answer and it has severe limitations. So consider really talking to the vendor about training and making sure you understand how to interpret, train and use it properly.

If your organization is misled by a vendor's AI promises, navigate this by resetting expectations. Communicate openly about the discrepancies and re-evaluate the vendor’s offerings. Practically, gather evidence of the mismatches, hold a meeting with stakeholders to realign goals, and consider alternative solutions. This proactive stance helps in managing the situation and maintaining trust.

The word optimization may sound simpler but it has larger meaning in the world of tech. It is quite common to be sold out to a vendor with price tag and at the end of the day it falls way below expectations. In such cases it is wise to take some time for internal meetings within the team and stakeholders, consolidate a new plan and lower overall expectations to make sure the overall project is not hampered👍

Navigate AI Vendor Overpromises: Resetting Expectations 1. Clear communication: If the AI hasn't lived up to the hype, it's crucial to communicate this new reality to key stakeholders. Honestly prevents further misunderstandings and damage. 2. Adjust goals: Reassess what the AI can realistically achieve. This may involve altering your project scope to align with the true capabilities of the technology. 3. Modify resources: Revise timelines, budgets, and resource allocations to fit the actual potential of the AI solution. By resetting expectations, you can manage the situation effectively and steer your project back on a course that remains true to your organization's goals.

I'll be honest that from my experience with various types of technology it is rarely an oversell that is the issue. Rather, the problem is the buyer's ability to effectively communicate what they need to accomplish, how the reviewed solution will address this and their ability to use and support it. Big decision without effective validation and testing. So resetting of expectations must go both directions. Start again with the business case, why are you considering this? What are the must have key elements? What must potentially be compromised with this solution? Then work from there.

Clearly communicate to the vendor the discrepancies found and the realistic outcomes you expect based on your assessment. Align internal stakeholders on these revised expectations, ensuring everyone understands the adjusted goals and timelines. Document all communications and agreements to hold the vendor accountable. By resetting expectations, you create a transparent framework that mitigates future misunderstandings and ensures that the AI solutions delivered meet your organization’s actual needs and capabilities.

It's crucial to assess and mitigate any potential damage to your organization's reputation or operations. If the misleading AI solution has affected your products or services, develop a transparent communication strategy for stakeholders, including clients and partners. Be honest about the situation and outline your plan to rectify it. Transparency is crucial. Internally, a thorough after-action analysis should be conducted to identify how the misleading promises were not caught earlier. Use this as an opportunity to strengthen your vendor vetting process and AI literacy across the organization. Consider establishing an AI ethics committee to oversee future AI implementations.

1. Look for specific clauses about the AI technology's performance, accuracy, or capabilities. Were there any guarantees provided by the vendor? 2. Bring your legal team into the loop. They can interpret the contract's fine print, assess the strength of your case, and advise you on the best course of action. 3. Keep meticulous records of all communication with the vendor, including emails, meeting notes, and any documentation of the AI's performance. Legal action is often a last resort, and your legal team can help you explore options like: You can sue if the product doesn't match the contract. You can seek damages if the vendor lied about the AI's abilities. You can try to reach a settlement with the vendor outside of court.

You can consult legal services if a seller scams you significantly. Discuss work conditions with a lawyer to manage the regulations and pledges. While going to court could be an extreme measure, any entity needs to know the laws that apply to safeguard its interests.

Consider this the "legal consultation" phase. It's like calling in a lawyer to review a contract before signing on the dotted line, except this time it's about a contract that's already been signed and might not hold up under scrutiny. Gather your legal eagles and have them pore over the fine print, looking for any clauses about warranties or guarantees that the vendor may have violated. Think of it as a game of legal chess, where you're strategizing your next move based on the evidence you've gathered and the potential consequences of each decision. It might not be the ideal situation, but knowing your legal rights and options is crucial when dealing with a vendor who may have misrepresented their AI technology.

Review the contractual terms and conditions to identify any breach of agreement. Consult with legal experts to understand your rights and the possible remedies available, such as seeking compensation or terminating the contract. Ensure all communications with the vendor are documented to support any legal actions if necessary. By addressing legal considerations, you protect your organization’s interests and establish a framework for holding the vendor accountable for their promises and deliverables.

Research other vendors. Look for those with a proven track record of successful AI implementations. Seek recommendations from industry peers or consult analyst reports.

AI startups are mushrooming everywhere with all kinds of promises. At the same time, few companies have had sufficient time to really build anything beyond an over-promised demo. So, there is a real build vs buy question if your organization is trying to be on the leading edge. Call a few product leaders in the space and ask what would it take to build, and at the minimum, you could be more informed with the next vendor.

When a seller’s AI solution does not meet all needs, there is a need to carry out an exhaustive market assessment to seek other options. Few things can be done to find these alternative solutions; first make sure you look for vendors who have previously demonstrated how their products work, have customers that you can contact and have technical documents. Learn from your previous experiences and screen subsequent alternatives more harshly to ensure that they are effective.

Start by evaluating other vendors who offer similar AI solutions, ensuring they have a proven track record and reliable references. Conduct thorough due diligence to compare their offerings, pricing, and customer support. Additionally, consider in-house development or partnerships with academic institutions or AI-focused startups. By exploring and assessing alternative options, you can find a more trustworthy solution that aligns with your organization's needs and objectives, ensuring the project's success and minimizing disruptions.

If the AI honeymoon phase with your vendor has fizzled out, it's time to start swiping right on other options. Think of it as a matchmaking process, but instead of finding love, you're seeking the perfect AI companion for your business. Scour the market, read reviews (aka case studies), and have those "getting to know you" conversations (demos and technical discussions) with potential new vendors. It's like speed dating with AI companies, where you're asking the tough questions to ensure they're a good fit before committing to a long-term relationship. Learn from your past heartbreak and don't settle for anything less than an AI partner that can truly deliver on its promises.

🔍 Leverage this vendor mishap as a catalyst for future success. Bolster your AI expertise by training your team or onboarding new talent specialized in AI. For instance, at BotOracle, we introduced specialized AI training modules to enhance our team's capability in evaluating and integrating new AI technologies. This proactive approach ensures that your team can scrutinize future vendor offerings with a critical eye, preventing similar issues. Additionally, develop a robust vetting process for new technology acquisitions, learning from this experience to establish a more thorough evaluation protocol. By embedding these practices, you'll strengthen your organization's AI strategy. Sources: BotOracle, Gartner, McKinsey

This can absolutely serve as a learning opportunity for the future. Moving forward, I would consider: (1) establishing a set of AI due diligence questions for vendors and documenting answers to those questions; (2) investing the time to investigate any claims or promises being made; and (3) ensuring that you've established a strong dynamic for ongoing collaboration and communication with the vendor.

This is really important. As a general rule right now, most claims made by AI vendors should be taken with huge bags of salt. Operationally useful AI technology is not the same as what is seen with AI research - it is worlds apart. So many instances where claims are made of AI, where AI is all but non-existent. Considering an AI solution means you MUST first fully understand what AI actually is capable of and where it can be best deployed. THEN you have to understand what problem you want to solve. If you have the capability, treat AI as a research project to understand how it will impact your business in the future and carefully test various applications. But don't bet your business on it.

Practical Example: Imagine you were sold on an AI-powered customer service chatbot that would handle 90% of your inquiries, freeing your human agents to focus on more complex tasks. But in reality, the chatbot can only handle 50% of inquiries, and it's not very good at understanding nuanced language. In this scenario, you might need to: Revise your customer service goals: Instead of aiming for 90% automation, you might need to settle for 60% or 70%. Extend your timeline: Training the chatbot to be more effective might take longer, so you need to push back your launch date. Reallocate resources: You might need to hire more human agents to handle the inquiries that the chatbot can't.

Doing due diligence is the no. 1 most valuable & effecting method to ensure you get the most out of your vendor. It is also advisable to get your CTO to evaluate the vendors development capacity and their products.