Is adequate evidence available to support the claims of efficacy of Ayurvedic medicines and treatments? The answer is tricky and a far cry from a simple yes or no. When we talk about evidence, we mean data that confirm to the parameters and interpretations developed by modern science, which may not make much sense in the case of health care systems like Ayurveda that are built on a different epistemology.
We have to qualify therefore and fine-tune our question. For convenience, we have to talk about two kinds of evidence. Let us, say, reframe the bigger question into smaller ones: (i) Is there evidence that Ayurveda works? This is very different from the question (ii) is there evidence regarding how Ayurveda works?
It seems feasible to attempt to answer the first question - Does Ayurveda work? But it is quite a challenge to find an answer to the second question - How Ayurveda works? More than often, the second question gets juxtaposed with the first question complicating the efforts to find an answer by force fitting methodologies and frameworks that are perhaps ill-suited to evaluate Ayurveda.
The Ayurvedic community is often accused of citing epistemological differences conveniently to evade the need for generating evidence to demonstrate the efficacy of Ayurvedic treatments. But, is there a need to evade the simple question whether Ayurveda works? It is an altogether different ballgame to prove how Ayurveda works.
At a very early stage in its evolutionary history, we can find debates in the classical literature of Ayurveda that tries to address the skepticism regarding the efficacy of Ayurvedic treatments. In the Caraka Samhitā, there is a chapter that narrates a debate that ensued between Maitreya, a skeptic and Ātreya, the teacher.  Maitreya bluntly states that there is no evidence that Ayurveda works because some people who are subjected to Ayurvedic treatment are seen to succumb to disease whereas some of those who do not undergo any treatment are seen to survive. Ātreya responds by making a distinction between self-limiting and intractable diseases, pointing out that patients get cured of the former without any treatment whereas the latter do not respond to any treatment. On the other hand, diseases that are tractable always respond to proper treatment.
There is further evidence that can be culled out from the same text supporting the view that the tradition of Ayurveda was open to the demand for evidence, especially the evidence as to whether the medicines and treatments work. There is a passage in the Caraka Samhitā, which states that the outcome of a treatment that cannot be substantiated with proper reasoning is to be dismissed as chance/serendipitous success (yadrchhāsiddhi).  The commentator of this text goes even further by differentiating between a real effect (prātiniyāmikīsiddhi)  as opposed to a chance effect of the treatment.
In other words, the question "Does Ayurveda work?" seems to be legitimate and in tune with the continued and sustained attempts of the tradition to generate answers which can keep the system robust and prevent its decadence.
On the other hand, it is not easy to answer the question, "How does Ayurveda work?." This is because the answer is expected to be framed within the theoretical and conceptual constructs of modern science that constitute the basis of modern medicine.
The Ayurvedic community must engage in activities that can generate answers to the primary and most fundamental question whether Ayurveda really works or not? We should think about how we should go about with implementing this task.
The gold standard for clinical evidence is the randomized controlled trial (RCT) and meta-analysis of a set of RCTs. This model of evidence ill fits Ayurveda's personalized approach to treatment, which comes as a package of multiple formulations, diet and behavioral regimen that cannot be tested as one item against a given disease.
There is still much ground to be made when we think of what could be the most suitable method to generate clinical evidence of the efficacy of the complex Ayurvedic interventions. In the meantime, reporting the outcomes of clinical practice at the grassroots level assumes great significance.
When we talk about evidence in the context of medical systems like Ayurveda, we are talking about something that is already in practice. The question is whether what we are already using is working or not. On the other hand, in modern medicine, the question regarding evidence is about something that has not yet been introduced into practice. In other words, the question is whether a new medicine or treatment should be introduced in practice, what are its long-term effects on the individual, whether it will adversely interact with other medicines, etc. The answer is that it should be introduced if it works and it doesn't do more harm than when it wasn't used (tadevayuktambhaisajyam yadārogyāyakalpate,  that is, that is a good medicine, which conduces to the health of the individual). For this reason, the approaches and methods to generate evidence in modern medicine differ from that of Ayurveda and the medicines have to be administered first to animals and then to humans who have never been exposed before. In the case of Ayurveda, we are talking about the efficacy of medicines that are already being consumed by large numbers of people who attest to the efficacy of the medicines. For this reason, it has been pointed out by scientists like Ashok Vaidya that the reverse pharmacology approach of starting with clinical studies that eventually culminates in laboratory studies that aim to unravel biological mechanism of the action of medicines is more suitable for Ayurveda. 
From this viewpoint, our task would be to find out if the medicines and treatments that are being used by large numbers of people really work in the manner they are expected to. Rigorous case studies constitute a good starting point to find the answer to the question whether Ayurveda works. Establishing a practice-based outcomes research network seems to be a much-needed step in this direction. The CARE guidelines (http://www.care-statement.org) from case reports could be adapted to the context of Ayurvedic clinical practice. 
It is often thought that building evidence to substantiate traditional medical systems like Ayurveda is an entirely new exercise. The fact that Ayurveda has accumulated and codified evidence of safety and efficacy of its treatments and medicines in textbooks that were composed and revised without a break for thousands of years is hardly recognized. We can confidently say that this wealth of materials constitutes the primary evidence that the tradition itself has to offer. However, this information needs to be deciphered and structured in a manner that would make sense in our present world.
Another kind of evidence that can be potentially salvaged from the tradition is the prevalent practice and applications of this knowledge by its practitioners in contemporary society. We have to create a mechanism that will ensure the reporting of such practices and their outcomes, which reinforce the evidence that is already codified in the texts.
New initiatives to generate evidence for Ayurveda should be erected upon the foundation of the preliminary evidence that is already available in the tradition. Without this exercise, we cannot conceptualize appropriate research designs to evaluate Ayurveda.
The problem is not the absence of controlled trials proving efficacy of Ayurveda, the problem seems to be the lack of appropriate research designs to address the complexity of Ayurvedic treatments. If we look at the research scenario in the field of Ayurveda and the number of publications, we can see that a large number of clinical studies have been conducted. On a closer look, deficiencies and lack of rigor become evident, but the most glaring shortcoming is the application of inappropriate research designs for conducting the trial. RCT designs may have to be suitably adapted to address the complexity of Ayurvedic treatments, or alternative methodological approaches may have to be envisaged to generate evidence of safety and efficacy of Ayurvedic treatments.
There is much hue and cry about the need for toxicology studies on herbo-mineral preparations.  The knowledge gap that needs to be addressed is epistemological. According to Ayurveda, it is possible to render toxic metals like mercury, arsenic and the like, nontoxic through a process of purification. This is unacceptable to modern science. For instance, there has never been documented in western science, a method to detoxify lead but Ayurveda has elaborate methods to detoxify and use lead as medicine. One-way to deal with this problem is to demonstrate through laboratory studies, what exactly happens when these toxic substances are pharmaceutically processed through the methods described in Ayurvedic texts and then testing on animals for safety as well as measure bioavailability, tissue distribution, histopathology, behavior, and epigenetics. The other approach would be pharmacoepidemiology.  A large number of patients in India are already exposed to the herbo-mineral preparations prescribed by Ayurvedic physicians across the length and breadth of the country. Safety data could be salvaged from this large pool of patients. If the Ayurvedic herbo-mineral preparations are found to be safe, it would be a major advancement in science.
Organizing the primary evidence for Ayurveda from the classical texts and the extant practices has to be taken up as a high priority activity. This could be the strong foundation upon which we can initiate activities to generate the kind of evidence that would eventually help Ayurveda to find for itself a commanding position in the modern world on the strength of its fundamentals.